# Tracy AI Interviewer > Tracy AI Interviewer runs two-way video interviews, follows up on vague answers in the person’s own field, and gives you evidence behind every result. # Best AI Interview Tools for RPO Firms (Nov 2026) > RPO interview software compared: Tracy, HeyMilo, Alex, HireVue, Paradox and Sapia on depth, volume, ATS stack, evidence and price for RPO teams. - URL: https://tracyinterview.ai/blog/ai-interview-tools-for-rpo - Published: 2026-11-13 - Author: Tracy team > **In short** Speed to submit decides RPO requisitions, and a thin submit gets sent back. Tracy AI Interviewer is RPO interview software that interviews every candidate when they click the link, at any hour, with follow-ups in their own field. > > Each candidate gets an assessment per question with the reasoning, and the client gets a package: resume match, assessment, strengths, gaps and recording. A recruiter decides each submit. > > Choose HeyMilo or Alex when Bullhorn write-back decides the purchase. Next step: run one requisition in Tracy. Tracy AI Interviewer is the better pick for an RPO firm that wants every candidate interviewed the same day, with an evidence package for the client, and HeyMilo or Alex are the better pick when writing results back into each client's ATS is the main job. An RPO team runs the same first round for many clients, in each client's own systems, against a submission deadline. RPO interview software has to start the interview as soon as a candidate applies and hand the recruiter something a client will accept. Six tools are compared below. Choose Tracy when: - you want every candidate interviewed, not only the ones a recruiter has time to call; - you want the evidence for every candidate of a requisition in one place, with a package for the client in one record; - you want to open a new client requisition with no seats and no upfront cost. Choose HeyMilo or Alex when ATS write-back decides the purchase. For agencies without an RPO contract, see [AI interview tools for staffing agencies](/blog/ai-interview-tools-for-staffing-agencies). ## What does RPO interview software have to do? The buyer is a head of delivery or recruiting lead, and the pain is speed to submit across every client's system. Five needs follow. 1. Depth. A client asks why this candidate, so the interview has to follow up on what the candidate said, in the role's own field. 2. A same-day start, on click, so evening and weekend applicants are not lost to scheduling. 3. Results that land in each client's ATS, or are easy to carry there. 4. A package the client accepts: resume match, assessment, strengths, gaps and a recording. 5. No seats and no upfront cost per client. A recruiter stands behind each submit, so the evidence goes to a person. Our guide to [screening when there are too many applicants](/blog/how-to-screen-candidates) covers the first two. ## How do the six tools rank, and why in this order? We ranked on one question: how much of a same-day, depth-first interview does each tool cover for a firm that works across several client systems? A published price breaks ties. Tracy is first here on a same-day start on click, an evidence package for the client and a published price. She is not first on every line: HeyMilo and Alex list Bullhorn and longer ATS lists, and HireVue has the deepest governance material. ## Which RPO interview software is the better pick, and when? ### 1. Tracy AI Interviewer: a same-day interview with a package for the client Tracy interviews each candidate by two-way video. After every answer she asks about what was said, in the candidate's own field, and waits for them to finish. You describe the role or bring your own questions and choose the skills to evaluate for each requisition as tags. Each candidate starts on click, at any hour, from a browser link with no account. The [participant experience](/participant-experience) page shows the candidate's side. On candidate experience, Steven Perry, Account Executive at IT Resources, an IT staffing firm in Tampa, US, says: "By the time they finish with this platform, they're even laughing and joking." Each candidate gets a written assessment for every question, with the reasoning, and every point links to the moment in the recording. The package for the client holds resume match, the assessment per question, strengths and gaps, the recording and the fit rationale in one record, shared by link or PDF. A recruiter can move a candidate forward or invite a candidate again. Integrity signals (tab changes, gaze, live face detection, delay before answers, fullscreen exits, external screens) reach the recruiter as evidence, not proof, and nothing is decided in the interview. [AI interview reports](/blog/ai-interview-reports) shows what a reviewer reads. Tracy connects through API and MCP, Slack, WhatsApp and Zapier, beside the Workable and Breezy HR chat connectors. A Bullhorn connector and ATS write-back are not stated on Tracy's pages, so test that first. [Pricing](/pricing) is public: $1 per credit, one credit per started 20 minutes, the first 3 results free, no seats, so 100 interviews of 20 minutes cost $100. ![Tracy list of eight candidates for one role: six finished with a View result link, one invited and one with an expired link.](/blog/images/ai-interview-tools-for-rpo-1.png) *Every candidate for a requisition sits in one list, showing who has finished and who still has a link to use. Illustration with sample data* Every candidate on a requisition sits in one list with a status, so the recruiter sees who has finished and who has not. ### 2. HeyMilo: the staffing ATS pick HeyMilo screens by voice, video, SMS and form, with follow-ups the recruiter caps and ties to a scoring rubric. Its integrations page lists a Bullhorn integration that syncs reports to Notes and creates JobSubmissions automatically, and says it is built for staffing agencies. It lists 20+ ATS partners. No price is published; its pricing page leads to a demo form. It is the better pick when the Bullhorn write-back decides the purchase. See [Tracy vs HeyMilo](/compare/tracy-vs-heymilo). ### 3. Alex: phone, WhatsApp and fraud checks Alex interviews by video, phone, SMS and WhatsApp and asks follow-ups based on what the candidate says. It lists 33+ ATS integrations, shows Bullhorn on its integrations page and has an RPO and Staffing section. Its Verify, Proctor and Identity layers check application fraud, interview help and a government ID. There is no public price. It suits a firm whose candidates answer on phones and whose clients care about identity checks. See [Tracy vs Alex](/compare/tracy-vs-alex). ### 4. HireVue: enterprise governance HireVue's AI interviewer asks probing follow-ups, and its pages tie each score to evidence from the interview and the competency rubric. Its materials list annual independent audits and ISO 42001 alignment, and state "AI Informs. Humans decide." The pricing page lists no prices. It fits when a large client's procurement team asks for audit papers. See [Tracy vs HireVue](/compare/tracy-vs-hirevue). ### 5. Paradox: hourly hiring by chat Paradox runs apply, screening and scheduling by SMS, WhatsApp and Facebook Messenger, and says interviews are scheduled within 10 minutes. Its pages describe screening for basic qualifications and scheduling, and an AI interview with follow-ups is not stated. There is no public price. It is the better pick for frontline requisitions where the work is scheduling many people fast. See [Tracy vs Paradox](/compare/tracy-vs-paradox). ### 6. Sapia: text interviews priced per hire Sapia runs five structured questions in an untimed text chat and ranks candidates with a fit score. It is text only, and follow-ups are not stated. It says it is independently audited for bias, with the results published. Pricing is per hire, at a tailored rate, for companies that hire more than 500 people a year. It is the better pick for text screening at high volume, or when a published bias audit is a client requirement. ## How do the six tools compare on depth, volume, stack and evidence? Rival facts: vendors' own pages. "Not stated" and "not verified" mean the page does not say or we have not checked. | Tool | Depth | Volume | Stack | Evidence | Price | | --- | --- | --- | --- | --- | --- | | 1. Tracy AI Interviewer | Two-way video, follow-ups in the candidate's field | Starts on click, any hour; assessment per question | API, MCP, Zapier, Slack, WhatsApp; Bullhorn not stated | Package: resume match, assessment, strengths, gaps, recording | $1 per credit, first 3 results free, no seats | | 2. HeyMilo | Voice, video, SMS; follow-ups tied to a rubric | Automatic invites; drop-offs re-engaged | 20+ ATS partners including Bullhorn, with write-back | Score and rationale per question, timestamps | No public price; demo form | | 3. Alex | Video, phone, SMS, WhatsApp; follow-ups from the answer | Outreach and reminders; says 48% of interviews run after hours | 33+ ATS integrations, Bullhorn listed | Transcripts, role-fit scores, fraud flags | No public price; demo form | | 4. HireVue | Probing follow-ups; on-demand and live video | Mid-market to enterprise first rounds | Scores in the ATS; list not verified | Scores tied to interview evidence | None published | | 5. Paradox | Chat and text screening for basic qualifications | Says scheduling within 10 minutes; SMS, WhatsApp | Workday, SAP SuccessFactors; open API | Qualified or not; scored report not described | No public price; demo-led | | 6. Sapia | Five text questions; follow-ups not stated | Untimed; built for 500+ hires a year | Customer's ATS or HRIS; list not verified | Ranked shortlist with fit score | Per hire, tailored rate, none published | ## What does the table leave out? Alex, HeyMilo, Sapia and HireVue state bias audits or independent audits; Tracy publishes no bias audit, and a person decides from the evidence report. Tracy is SOC 2 Type II attested, with the report on request, and the other five list SOC 2 on their pages. Paradox answers candidate questions in 100+ languages, while Tracy runs a spoken-language check and publishes no language count. Sapia lists screen-reader support; accessibility is not stated for Tracy. Tracy Enterprise includes a dedicated customer manager and support channel, and a setup customised in days. Implementation time, support and contract terms: not verified for the other five. Reasoning, not a rival figure: at low volume, a per-seat, per-hire or contract model likely costs more. ![Three steps: same-day interviews on click, an assessment per question for each candidate, and a package with resume match, assessment and recording.](/blog/images/ai-interview-tools-for-rpo-2.png) *Tracy gives every client requisition the same same-day interview and client-ready package, so the recruiter submits with evidence. Illustration. Source: Tracy Product and Value records.* One interview layer serves every client requisition: same-day interviews, evidence per candidate and a package to send. ## How do you test them on two live requisitions? A demo shows the vendor's best case, so rank tools on your own candidates. Pick two live requisitions from two clients on different ATS, send the same 20 candidates per requisition through each tool, and score hours from application to a submit-ready package, depth of follow-ups, client acceptance of the package and how the result reached the ATS. On Tracy a 34-minute interview is 2 credits, or $2, so the test costs little. Run your first requisition in Tracy and you see the package your client would get. See the [cost guide](/blog/ai-interview-cost). ## Which path fits your firm? - Hundreds or thousands of applicants a month: [volume hiring](/use-cases/volume-hiring). - A wider list: [AI interview tools for high-volume skilled hiring](/blog/ai-interview-tools-for-high-volume-hiring). - Your ATS's interviewer or a dedicated one: [the comparison](/blog/ats-ai-interviewer-vs-dedicated). ## Common questions ### What is RPO interview software? It interviews candidates for an RPO firm's client requisitions before a recruiter submits them, by video, voice, text or chat. Tracy AI Interviewer interviews by video and leaves a package for the client. ### Can one tool run interviews for several clients? With Tracy, yes. Each role has its own interview, with its own skills chosen as tags, so every requisition gets its own plan and results. Test how each client's ATS receives the result. ### Does RPO interview software work with Bullhorn? HeyMilo and Alex list Bullhorn. A Bullhorn connector is not stated on Tracy's pages, so connect through the API, Zapier or MCP and test the hand-off first. ### How much does RPO interview software cost? Tracy costs $1 per credit with the first 3 results free and no seats. HeyMilo, Alex, HireVue and Paradox publish no price, and Sapia prices per hire. ### Will a client accept an AI interview? Candidates are told on the join screen that the conversation is recorded and transcribed, and a recruiter decides from the evidence report. --- Run one client requisition in Tracy and compare the follow-ups and the package with your current screen. Send us the role and we build the interview plan, customised in days, with a recording and report you can check. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=ai-interview-tools-for-rpo&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=ai-interview-tools-for-rpo&utm_content=try-free) ## Sources - [Tracy AI Interviewer, Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Volume hiring](https://tracyinterview.ai/use-cases/volume-hiring), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Connect](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Home](https://tracyinterview.ai), retrieved 9 Nov 2026. - [HeyMilo, Integrations (Bullhorn)](https://heymilo.ai/integrations), retrieved 9 Nov 2026. - [Tracy AI Interviewer, customer proof on the home page](https://tracyinterview.ai), retrieved 9 Nov 2026. # Askable Reviews: What Reviewers Say (Nov 2026) > Askable reviews counted, not picked: 20 recent Trustpilot reviews with quotes and dates, what they praise, what they flag, and who it suits. - URL: https://tracyinterview.ai/compare/askable-reviews - Published: 2026-11-13 - Author: Tracy team > **In short** Tracy AI Interviewer drafts the interview guide from what you want to learn, customers join in the browser when it suits them, and each finding links back to the words and the moment in the recording. > > Choose Askable when you need participants found for you: verified people in 50+ countries, with recruitment included. > > The most common complaint in the 20 newest Trustpilot reviews is a late or refused participant payout: 5 of 20. > > Tracy's price is public: the first 3 results are free, then $1 per credit, with no seats. Askable reviews tell you more about a participant panel than about interviews. Most of the 20 newest Trustpilot reviews read as written by people who take part in studies for pay, and five of them describe payouts that arrived late, were refused or never came. If you lead research and are choosing between a panel and your own customers, that shapes what these reviews can tell you. [Tracy AI Interviewer](/use-cases/customer-research) is the second route: you bring the customers, and she runs the interviews. ## Where do these Askable reviews come from? The corpus is the 20 newest of 279 Trustpilot reviews of Askable (TrustScore 4.0), dated 15 April to 5 November 2026. Counts cover Trustpilot only. Twenty is a small base, so read each count as a direction. Most of these writers are panel participants, not the researchers who buy. ## What do Askable reviewers praise? Eight of the 20 reviews give five stars. Three praise support: "The customer service team is friendly, easy to approach and very responsive" (Trustpilot, 28 August 2026). Two say they were paid in good time, and one writes that Askable "really is the best in the business" (18 September 2026). A reviewer who gave two stars still calls the app "intuitive and simple to use" (21 July 2026). Askable's own pages describe the panel as verified people in 50+ countries and say recruitment carries no fee. Those are vendor statements, and the panel is what most research leads buy Askable for. ## What do Askable reviewers complain about? Counts are out of the 20 reviews, with one quote each. Counts: our reading of the 20 newest Trustpilot reviews. A review can raise more than one issue. Quotes are word for word. | Complaint | Reviews (of 20) | One quote | Date | | --- | --- | --- | --- | | Payout late, refused or missing | 5 | "Askable will sometimes refuse to pay for completed surveys without giving any explanation." | 30 Sep 2026 | | Login problems | 2 | "I cannot login my account has my mobile number" | 25 May 2026 | | Slow human support | 1 | "After nearly a week without a response from a human support agent, Askable finally got back to me" | 1 Oct 2026 | | Few studies to join | 1 | "Availability of surveys not very high." | 23 Apr 2026 | | Recruitment handled badly | 1 | "I am extremely frustrated and disappointed with how this recruitment process was handled." | 27 Oct 2026 | ![Bar chart of 20 Askable reviews by complaint: payouts 5, login 2, support 1, few studies 1, recruitment 1.](/blog/images/askable-reviews-2.png) *Reviews that raise each issue, of the 20 newest. A review can raise more than one. Trustpilot, Askable page, 20 of 279 reviews, April to November 2026.* Payouts are the issue raised most: 5 of 20, against 2 or fewer for every other complaint. ## What do the reviews mean for a research lead? These are participant complaints, so they speak to panel health, not to moderation or analysis. As reasoning, not a finding: a panel works only while participants trust they will be paid, so raise payouts on your demo call. Ask how Askable handles unpaid sessions and no-shows, what a replacement costs, and for references from teams that recruit in your market. Askable's pricing page lists Pro and Enterprise with no fixed price, so ask for a figure for your study volume in the same call. ## Who is Askable right for, and who should look at Tracy? Askable is the better pick when you need people you do not have: strangers who match a niche profile, recruited for you. Askable says its platform runs AI-moderated interviews, and Pro adds a Certified Researcher. Tracy does not supply participants. She is the better pick when the customers are already in your CRM, a churned-accounts list or a win-loss queue, and the bottleneck is booking and running the calls. ## When does a panel beat your own customers? The two routes answer different questions. A panel suits discovery in a market you do not know yet, where you need strangers who fit a screener, and where Askable's reach across 50+ countries counts. Your own customers suit questions only they can answer: why accounts churned, why a deal was lost to a rival, what confused people in the first week. Those people already know your product, and they are often in your CRM with no one free to call them. Each Tracy interview there takes the time of a link, not of a calendar, and each finding links back to the words and the moment in the recording. Many research teams will want both. Askable column: Askable's published pages. Tracy column: Tracy's published capabilities. | Need | Askable | Tracy | | --- | --- | --- | | Finding participants | Verified people in 50+ countries; recruitment included. | Not a panel. You send one link to customers you already have. | | Running the interview | Remote sessions and AI-moderated interviews (listed on its home page); Pro adds project delivery with its AI Moderator and a Certified Researcher. | Tracy drafts the guide from what you want to learn, or uses yours. Customers join in the browser when it suits them. | | Evidence | Session evidence feeds an AI repository you can ask across. | Each finding links back to the words and the moment in the recording. | | Price | Custom, on Pro and Enterprise tiers; no fixed price listed. | $1 per credit, one credit per started 20 minutes, first 3 results free, no seats. | ![Tracy asks how approvals are handled, the customer says chat, and Tracy asks why chat, each line timestamped.](/blog/images/askable-reviews-1.png) *Tracy asks why after a one-line answer, and each line is timestamped in the transcript. Illustration with sample data* Tracy follows a short answer with a question about the reason. ## How does Tracy run a customer interview? Surveys show what people chose, not why. A customer interview gets at the reason because the interviewer follows the answer: after a customer says approvals now happen in chat, Tracy asks why chat and what stopped happening in email. Each finding links back to the words and the moment in the recording. If your customers are already in a list, [create an interview for your own research question](/connect) and compare the follow-ups and the report with what you get from a panel study. ## What does Askable cost, in brief? Askable shows no fixed price, so budget by asking for a quote. Tracy's price is on the [pricing page](/pricing): the first 3 results are free, then $1 per credit, one credit for each started 20 minutes, with no seats and no upfront cost. At 20 minutes each, 100 interviews cost $100. For the wider picture, see [what an AI interview costs](/blog/ai-interview-cost). ## Which path fits you? Choose Askable if you need participants found and screened for you. Choose Tracy if you already have the customers. If you want to know how a participant sees a Tracy interview, read the [participant experience](/participant-experience). The [customer research page](/use-cases/customer-research) shows the use case in full. For other tools that run interviews, see [the best AI interview software](/blog/best-ai-interview-software). ## Common questions ### Are Askable reviews good? The 20 newest Trustpilot reviews are mixed: 8 give five stars and 6 give one. Most writers are panel participants, so they speak to payouts and support, not to the studies research teams run. ### What are the main complaints about Askable? Late, refused or missing payouts (5 of 20), login problems (2), slow human support (1), few studies to join (1) and a badly handled recruitment (1). ### What do Askable app reviews say? The app is the participant side. One reviewer calls it intuitive and simple to use, two have trouble logging in, and several are about whether payouts arrive. ### What are the alternatives to Askable? If you need a panel, look at other panel vendors. If you already have the customers, Tracy runs the interviews from one link and leaves a recording and timestamped quotes behind each finding. ### How much does Askable cost? Askable lists Pro and Enterprise tiers with custom pricing and no fixed price. Tracy costs $1 per credit, with the first 3 results free and no seats. --- Create one interview for your own research question, then compare the follow-ups and the report with what a panel study gives you. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=askable-reviews&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=askable-reviews&utm_content=demo) ## Sources - [Trustpilot: Askable reviews](https://www.trustpilot.com/review/askable.com), retrieved 9 Nov 2026. - [Tracy AI Interviewer: Customer research](https://tracyinterview.ai/use-cases/customer-research), retrieved 9 Nov 2026. - [Tracy AI Interviewer: Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. # Best AI Tools for Recruiters (Nov 2026) > The best AI tools for recruiters, by job: Tracy, HireVue, HeyMilo, Paradox, Alex, Sapia, Metaview and BrightHire, with what each does and costs. - URL: https://tracyinterview.ai/blog/best-ai-tools-for-recruiters - Published: 2026-11-13 - Author: Tracy team > **In short** Tracy AI Interviewer is the pick for the first-round interview: every applicant interviews on click at any hour, with follow-ups, and you get a package to send on: resume match, interview score, strengths and gaps, recording and fit rationale. > > Her price is public: the first 3 results are free, then $1 per credit, with no seats. > > HireVue, HeyMilo, Alex and Sapia run AI interviews too. Paradox schedules, and Metaview and BrightHire take notes on human-run interviews; Tracy runs the interview itself. > > Pick the tool for the job that costs you the most hours. The best AI tools for recruiters each do one job well, so the right list depends on where your hours go. Sourcing finds people, screening and interviewing decide who moves on, scheduling books the time, and notes record what was said. Most recruiters lose the most time on the first-round interview: the calls that fill the diary before anyone has seen a good candidate. [Tracy AI Interviewer](/use-cases/volume-hiring) takes that job. She interviews every applicant from a link, asks follow-ups in their own field, and hands you a package you can send on. This list covers screening, interviewing, scheduling and notes. Sourcing is covered where a tool lists it: Metaview and Alex. ## What should AI tools do for a recruiter? Four questions separate the tools. **Depth:** does it follow up on what the person said, or read from a script? **Volume:** can every applicant start at any hour without a calendar? **Stack:** does it work with your ATS and chat tools? **Evidence:** can you check a result against what was said? Price is the tie-breaker, because staffing margins make a fixed contract hard to justify before volume is proven, and because a public price lets you check the cost before you book a demo. ## How do the tools rank, and why in this order? The ranking is for the first-round interview, the job with the biggest recruiter-hours cost, judged on depth, volume, stack and evidence. Tracy is first for the first-round interview because it combines follow-ups, any-hour start, a recording with transcript and a public price. HireVue, HeyMilo and Alex follow because each runs an AI interview. HireVue has probing follow-ups and scores in the ATS, with no published price. HeyMilo interviews by voice, video, SMS and forms, with a demo form for price. Alex lists 33+ ATS integrations, with price on request. Sapia comes next as a text-only screen built for large employers. Paradox, Metaview and BrightHire rank lower for interviewing because their strength lies elsewhere: scheduling and notes. They rank first for those jobs. Each row is what the vendor's own pages state. Price is as published on the vendor's pages on 9 November 2026; Alex's figure is from its February 2026 blog. | Tool | Main job | Interview format | ATS and stack | Price | | --- | --- | --- | --- | --- | | Tracy AI Interviewer | First-round interview | Interview with follow-ups, recording and transcript | ATS connector, Workable, Breezy HR, Zapier, Slack, WhatsApp, API and MCP | First 3 results free, then $1 per credit, no seats | | HireVue | AI interviews, video platform | Questions from the job description, probing follow-ups | Scores visible in the ATS | Not published | | HeyMilo | AI recruiter for high-volume hiring | Voice, video, SMS and forms | Widget in Bullhorn, Greenhouse, Lever and Ashby | Not published; demo form | | Alex | Outreach, screening, interviewing, scheduling | Video and phone, 26+ languages by phone | 33+ ATS | Pricing page: demo form. Blog, Feb 2026: from $36K a year | | Sapia | Text-chat screen | Five questions, text only, untimed | Works with your ATS or HRIS | Per hire; no rate published | | Paradox | Apply, screen and schedule by chat | Chat and text screening | Not stated on the pages read | Not published; demo request | | Metaview | Interview notes, sourcing, application review | Notes on human-run interviews; a voice screening agent | Greenhouse, Ashby, Lever, Workday and more | Notes: free, then $60 per user per month; sourcing Pro $100 per user per month | | BrightHire | Interview notes and an AI screening interviewer | Voice-only or video, asynchronous | Greenhouse, Ashby, Workday, SmartRecruiters | No dollar amounts published; priced by volume | ## Tracy AI Interviewer: the first-round interview Tracy interviews the applicant from one link, so a recruiter does not book a slot. Each applicant starts on click, at any hour, and a dropped applicant can rejoin within 10 minutes. She asks follow-ups in the person's own field, and reads the resume when the CV check is on. What you receive is a package you can send on: resume match, interview score, strengths and gaps, and the recording and fit rationale in one record, shared by link or PDF. Every result opens on a summary and an assessment per question, so you do not read every transcript. Use the tools above for sourcing, scheduling and notes on calls your team runs; Tracy is built to run the interview. You pay per interview: the first 3 results are free, then $1 per credit, with one credit per started 20 minutes and no seats. At 20 minutes each, 100 interviews cost $100. On the candidate side, Steven Perry, Account Executive at IT Resources, an IT staffing firm in Tampa, US, says: "By the time they finish with this platform, they're even laughing and joking." Read more on the [hiring at volume page](/use-cases/volume-hiring) and the [pricing page](/pricing). ![Three steps: start on click at any hour, follow-ups on what the applicant says, a package to send on by link or PDF.](/blog/images/best-ai-tools-for-recruiters-2.png) *Tracy's first-round interview in three steps: it starts on click, asks follow-ups, and hands the recruiter a package to send on. Source: Tracy product pages.* One link replaces the scheduling, and the recruiter receives a package to send on. ## Which tool fits which job? - **HireVue** builds questions and rubrics from the job description, asks probing follow-ups and shows scores in the ATS, according to its AI Interviewer page. It publishes no price. It fits an enterprise team already on HireVue. - **HeyMilo** is an AI recruiter for high-volume hiring that interviews by voice, video, SMS or form. Its pricing URL leads to a demo form. It fits a team that wants several channels in one tool. - **Alex** handles outreach, screening, interviewing and scheduling, with 33+ ATS integrations. Its pricing page shows a demo form, and its February 2026 blog post says annual agreements start at $36K. It fits a larger staffing firm. - **Sapia** runs a five-question text chat and charges per hire. It says it is designed for companies hiring more than 500 people a year. - **Paradox** applies, screens and schedules by chat and text, and shows no prices. It is the better pick when your time goes on scheduling, not on first-round interviews. - **Metaview** and **BrightHire** take notes on interviews your team runs, which is the better pick when recruiters keep running the calls. Metaview publishes prices, including a free notetaker plan and a $100-per-user sourcing plan. BrightHire lists no amounts and prices by volume. ![Tracy asks what managing a hospital account involved, the candidate answers, and Tracy asks about a short-notice call-out.](/blog/images/best-ai-tools-for-recruiters-1.png) *With the CV check on, Tracy asks about a claim on the resume and follows the answer. Illustration with sample data* Tracy follows a resume claim with a question about the detail, and each line is timestamped. ## What should you ask about price? Six of the eight tools show no price on their pricing pages, so budget by asking. As reasoning, not a finding: a low-volume agency under a per-hire, per-seat or annual contract likely pays more per interview than a firm at high volume. On a quote call, ask what the minimum term is, whether pricing counts seats, hires or interviews, what a failed or abandoned interview costs, and whether setup and ATS connection are extra. Tracy's pricing page answers most of these: per interview, no seats, no upfront cost, and a session with no answers is never charged. Use the same questions for every vendor and compare the answers on one sheet. ## How to choose Choose by the job that costs you the most hours. If it is first-round calls, [create one interview in Tracy](/connect) for a role you are filling now. If it is scheduling, look at Paradox. If it is interview notes, look at Metaview or BrightHire. If it is the wider market, see [the best AI interview software](/blog/best-ai-interview-software). Staffing teams will also want [how AI interviews work](/blog/how-do-ai-interviews-work) and [AI interview reports](/blog/ai-interview-reports). ## Common questions ### What are the best AI tools for recruitment agencies? For first-round interviews, Tracy; for scheduling, Paradox; for notes on human-run interviews, Metaview or BrightHire. Alex and HireVue suit larger firms that run a quote process. ### What are the best AI sourcing tools for recruiters? Metaview lists a sourcing agent with a free tier for the first 100 profiles and a Pro plan at $100 per user per month. Alex lists outreach among its jobs. Tracy is built for the interview, not sourcing. ### What are the best free AI tools for recruiters? Metaview has a free notetaker plan with 25 calls per user per month. Tracy's first 3 results are free, with no card. ### Can AI tools replace recruiters? No. A recruiter decides: they can move a candidate forward or invite them to interview again, and the history stays on the record. --- Create one interview for the first-round calls that fill your diary: send us one role and we will build the interview plan, with a recording and report you can check. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=best-ai-tools-for-recruiters&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=best-ai-tools-for-recruiters&utm_content=try-free) ## Sources - [Metaview: Pricing](https://www.metaview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer: Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer: Hiring at volume](https://tracyinterview.ai/use-cases/volume-hiring), retrieved 9 Nov 2026. # BrightHire Alternatives (Nov 2026) > BrightHire alternatives and competitors: Tracy, Metaview, HireVue, HeyMilo, Ribbon and VidCruiter on who runs the interview, evidence, stack and price. - URL: https://tracyinterview.ai/compare/brighthire-alternatives - Published: 2026-11-13 - Author: Tracy team > **In short** Tracy AI Interviewer is first for teams that want the first round run for every applicant: two-way video, own-field follow-ups, a recording and assessment per question. > > She costs $1 per credit, with the first 3 results free and no seats. > > BrightHire stays the pick for notes on interviews your team runs, 18 ATS connectors and a bias audit. Metaview is the nearest notes swap at a listed price. > > Next: pick your top two and test them on one role. Teams look for BrightHire alternatives when the quote, the split across products or the notes-first design stops fitting. When the first round has to run for every applicant and a reviewer must be able to open the evidence, Tracy AI Interviewer is the first tool to try. The list puts Tracy first, then five other tools, each with who it suits and where it falls short. ## Why do teams look for BrightHire alternatives? Three reasons come from BrightHire's own pages. The price is a demo. Its pricing page lists no amount, and each of its three products is priced on a different unit: company size, interview volume or applicant volume. Tracy shows hers: $1 per credit, on the [pricing page](/pricing). The starting point is the ATS. BrightHire calls its ATS integrations the deepest in the category, and its notes, scorecards and Screen invites are built around them. That suits a supported ATS and helps less when your first round sits outside an ATS stage. The interviewer may not be what you need. If you want the first round run for every applicant, Screen is one answer and there are others below. Some teams should stay. If interviewers run most of your calls and you want notes and highlights in the ATS, a third-party bias audit and 30+ languages, BrightHire lists all of it. ## Which BrightHire alternative suits which team? - Tracy AI Interviewer: technical or expert roles where the first round should run for every applicant, at a published price. - Metaview: notes and scorecard fill for interviews you run, with a published Notetaker price. - HireVue: large enterprise teams that want governance documents and one suite, by quote. - HeyMilo: high-volume voice, video, SMS and forms with two-way ATS sync, by quote. - Ribbon: sourcing, outreach and a first-round interview in one tool. - VidCruiter: live panels, scheduling and native HR connectors. ## Tracy AI Interviewer: follow-ups and evidence on every result Tracy interviews every applicant by two-way video. After each answer she asks about what was said, in the person's field, and waits for them to finish. Behind every result sit a recording, a timestamped transcript, a written assessment per question and integrity signals shown to a human reviewer. A person decides. A Director at a top-tier early-stage VC says: "Tracy sometimes brings up exactly those questions I might have forgotten to ask, which has been a pleasant surprise." You can bring your own questions or let Tracy draft them, and a coding task can sit inside the interview (see [technical interviews](/use-cases/technical-interviews)). She starts on click with no scheduling, connects through API and MCP, Slack, WhatsApp and Zapier. The price is on the [pricing page](/pricing), and for uneven [volume hiring](/use-cases/volume-hiring) the cost follows the interviews you run. Where others are better: BrightHire lists native connectors to 18 ATS, including Workday and SAP SuccessFactors, which are not stated on Tracy's Connect page. It also states a third-party bias audit and 30+ languages, and Tracy publishes no bias audit. ![Sample interview: Tracy asks about duplicate payment webhooks, then follows up on where the event id is stored.](/blog/images/brighthire-alternatives-1.png) *Tracy asks a follow-up on the candidate's last answer, so a reviewer sees depth, not a recited answer. Illustration with sample data* The reviewer reads the follow-up in the exchange, with timestamps. ## Metaview: notes and scorecards at a listed price Metaview is the nearest swap if your interviewers run the calls. Its notes page says Metaview "maps interview answers straight into your scorecards on Greenhouse, Ashby, Lever, Workday, and more." Its Notetaker has a free plan with 25 calls per month and a Pro plan at $60 per user per month. It also sells a screening agent: "Every call is recorded, transcribed, and scored against your rubric," and "A human always makes the final decision." Screening price: not stated on its pricing page. Pick Metaview for notes at a set per-seat price. ## HireVue: the enterprise suite with governance documents HireVue is the closest like-for-like swap for an enterprise buyer. Its AI agents "ask and record candidate responses, including probing follow-up questions, in a natural two-way conversation," it lists more than 200 on-demand coding assessments, and it publishes governance detail including adverse impact testing and an annual SOC 2 Type 2 audit. It lists no price and points to a plan conversation. Pick it when governance documents are a procurement requirement. See [HireVue alternatives](/compare/hirevue-alternatives). ## HeyMilo: voice, video, SMS and forms for high volume HeyMilo screens applicants over voice, video, SMS and forms in 20+ languages. It lists two-way sync with 20+ ATS partners, with scores, transcripts and recordings written back to the candidate record, and says it runs monthly independent bias audits by Warden AI. It publishes no price. Its own comparison page says follow-ups can only come from the recruiter's scoring rubric. Pick it for high-volume screening in many languages with ATS write-back. See [HeyMilo alternatives](/compare/heymilo-alternatives). ## Ribbon: sourcing, outreach and a first-round interview Ribbon runs structured voice and video interviews in 10+ languages, lists 60+ integrations and a 7-day free trial, and adds AI sourcing and outreach for high-volume and frontline hiring. Its pricing URL currently returns a 404, and its home page says it has signed an agreement to be acquired by Dayforce. Pick it for sourcing plus a first-round screen on frontline roles. See [Ribbon alternatives](/compare/ribbon-alternatives). ## VidCruiter: live panels, scheduling and native connectors VidCruiter runs pre-recorded and live interviews with a human, including panels, with self-serve scheduling and 75+ integrations including Workday and SAP SuccessFactors. Its pricing page is a request form for nine modules. Pick it if you run live panels or need scheduling and references from one vendor. ## BrightHire and its alternatives compared on four pillars and price Six tools on four pillars and price, from each vendor's own pages, Nov 2026. | Tool | Depth | Volume | Stack | Evidence | Price | What to test | | --- | --- | --- | --- | --- | --- | --- | | Tracy AI Interviewer | Two-way video, follow-ups in the person's field | Starts on click, any hour, no scheduling | API, MCP, Slack, WhatsApp, Zapier | Recording, transcript, assessment per question, integrity signals | $1 per credit, first 3 results free | Count the follow-ups that landed on your role | | BrightHire | Notes for interviews people run; Screen runs async voice or video with follow-ups | Screen is async, no scheduling | 18 ATS, Zoom, Google Meet, Teams | Notes, highlights, transcripts, audio playback | By demo, per product | Ask for a live Screen run and a first-year total | | Metaview | Notes in any interview format; a screening agent | Screening without scheduling | Greenhouse, Ashby, Lever, Workday and more | Recorded, transcribed, scored; a human decides | Notetaker $0 or $60 per user per month | Ask what screening costs | | HireVue | AI interviewer with probing follow-ups; 200+ coding assessments | Mid-market to enterprise first rounds | Scores shown in the ATS | Scores tied to evidence; governance detail | None published | Ask what sets price and for audit documents | | HeyMilo | Voice, video, SMS and forms; follow-ups from the rubric | 20+ languages, 24/7 | 20+ ATS, two-way sync | Scores with rationale, transcripts, recordings | No public price | Ask how many follow-ups your rubric allows | | Ribbon | Structured voice and video interviews | Quick to start; 10+ languages | 60+ integrations | Recording, transcript, summary, scorecard | Not stated; its pricing page returns a 404 | Ask about the Dayforce deal | | VidCruiter | Pre-recorded and live; same questions in order | Pre-recorded link 24/7; live needs booking | 75+ integrations | Transcript; AI screening rationale | By request | Ask for the quote in writing | ![Grid of who runs the interview against public price for Tracy, BrightHire, HireVue, HeyMilo and Metaview.](/blog/images/brighthire-alternatives-2.png) *Of the five tools in this grid, Tracy is the AI interviewer that shows a price. Source: each vendor's pricing and product pages, Nov 2026.* Of the tools whose pages show an AI interviewer, only Tracy also shows a price. ## What do the four pillars leave out? Implementation, security, accessibility and contract terms; Not stated where a vendor is silent. | Criterion | Tracy AI Interviewer | BrightHire | Others | | --- | --- | --- | --- | | Implementation and support | Enterprise: a dedicated customer manager, a dedicated support channel, a setup customised in days. | Most customers set up in under a day with no extra fees; dedicated customer success. | HeyMilo lists a named customer success manager and a forward deployed engineer. | | Security | SOC 2 Type II attested; report on request. | Says SOC 2 Type II certified, GDPR and CCPA compliant. | HireVue lists an annual SOC 2 Type 2 audit and FedRAMP. | | Accessibility | Not stated. | Not verified. | Not verified. | | Contract terms | Pay as you go, no seats, no upfront cost. | Each product priced on its own; terms not stated. | Metaview Notetaker Free is $0 with 25 calls a month. | ## How do you choose a BrightHire alternative? 1. Who runs the interview? If your interviewers run the calls, you need notes and scorecards: BrightHire or Metaview. If you want it run for every applicant, you need an AI interviewer: Tracy, BrightHire Screen, HireVue, HeyMilo or Ribbon. See [one-way vs two-way video interviews](/blog/one-way-vs-two-way-video-interviews). 2. What must the first round show? If you need to see how people reason in their field, ask each vendor for a live run on your own role. 3. How will you pay? Quote-only prices cannot be compared until you know the unit and the first-year total. Tracy's is $1 per credit. 4. What must it connect to? Name your ATS and ask for the connector. The [Connect page](/connect) lists Tracy's. Which path fits you: [technical interviews](/use-cases/technical-interviews), [volume hiring](/use-cases/volume-hiring), [expert vetting](/use-cases/expert-vetting). For wider lists, read [best AI interview software](/blog/best-ai-interview-software). ## Common questions ### What is the best BrightHire alternative? It depends on the job. Tracy suits roles where the first round should run for every applicant and show evidence, at a published price. Metaview suits notes for interviews your team runs, HireVue is the nearest enterprise suite, and HeyMilo and Ribbon run high-volume AI screens. ### Who are BrightHire's competitors? Metaview on interview notes and screening, HireVue, HeyMilo, Ribbon and Tracy on AI interviews, and VidCruiter where live interviews and scheduling matter. ### Is there a BrightHire alternative with public pricing? Yes. Tracy costs $1 per credit with the first 3 results free. Metaview publishes Notetaker plans at $0 with 25 calls a month and $60 per user per month. BrightHire, HireVue, HeyMilo and VidCruiter publish no price. ### Which alternative works for technical interviews? Tracy can run a coding task inside the interview and ask follow-ups on approach and trade-offs. HireVue lists more than 200 on-demand coding assessments. --- Create one interview for the workflow your current tool misses. Send us the role and we build the interview plan, customised in days, with a recording and report you can check. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=brighthire-alternatives&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=brighthire-alternatives&utm_content=try-free) ## Sources - [BrightHire, Pricing](https://brighthire.com/pricing/), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Connect](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. # BrightHire Pricing (Nov 2026) > BrightHire pricing is by quote, per product. Tracy's is public: $1 per credit. See the cost at 100, 500 and 1,000 interviews and what to ask on a call. - URL: https://tracyinterview.ai/compare/brighthire-pricing - Published: 2026-11-13 - Author: Tracy team > **In short** Tracy AI Interviewer costs $1 per credit, one credit per started 20 minutes, so 100 interviews of up to 20 minutes cost $100. The first 3 results are free, with no seats. > > Each result has a recording, a timestamped transcript and an assessment per question, and a person decides. > > BrightHire lists no amount; it stays the pick if your interviewers run the calls and you need notes in the ATS. > > Next: try three interviews free, then calculate your own cost. BrightHire pricing is not public. Tracy AI Interviewer publishes hers, on [Tracy's pricing](/pricing): $1 per credit, the first 3 results free, no seats and no upfront cost. A public price means you can budget a first round before any call, and an applicant can start on click at any hour with no recruiter booking it. - Choose BrightHire when your interviewers run the calls and you want notes, highlights and scorecards in the ATS, native connectors to its 18 listed ATS, or a third-party bias audit. - Choose Tracy when you want the first round run for every applicant at a price you can compute today. ## What is public about BrightHire pricing? BrightHire's pricing page lists no amounts and sends you to a demo. This table lists everything its pages say about cost, with the exact words. Source: BrightHire pricing and product pages, Nov 2026. Source: BrightHire pricing and AI Interviewer pages, Nov 2026. | What the page covers | What BrightHire says | Exact words | Page (Nov 2026) | | --- | --- | --- | --- | | Plans and prices | No plans and no prices. A demo request gets a plan. | "Request a demo and we'll put together a plan that fits your hiring volume and team size." | Pricing | | Interview Intelligence | Priced by company size. | "Priced by company size" | Pricing | | BrightHire Screen (AI interviewer) | Priced by interview volume. Its own page adds no additional setup fees. | "Priced by interview volume" | Pricing and AI Interviewer | | BrightHire Inbound (application review) | Priced by applicant volume. | "Priced by applicant volume" | Pricing | | Structure and bundling | Each product is priced on its own, and you can buy one or bundle. No discount amount is stated. | "Each product is priced on its own, based on how your team uses it." | Pricing | | Setup and fees | Most customers set up in under a day, with no extra training or implementation fees. | "Most customers are set up in less than a day, with no extra fees for training or implementation." | Pricing | | Free trial and contract terms | Not stated. | Not stated | Pricing | ## What does Tracy cost at 100, 500 and 1,000 interviews? Tracy charges one credit, $1, for each started 20 minutes of an interview that produced a result. A 15-minute interview is $1 and a 34-minute one is $2. Credits never expire, a session with no answers is never charged, and API and MCP access (see [Connect](/connect)) is included. The first 3 results are free whatever their length; the totals ignore that. Tracy cost by interview count, from the published rate of $1 per credit; BrightHire is quote-only, Nov 2026. | Interviews | Tracy, up to 20 minutes each (1 credit) | Tracy, 21 to 40 minutes each (2 credits) | BrightHire | | --- | --- | --- | --- | | 100 | $100 | $200 | Not public. Quote on request. | | 500 | $500 | $1,000 | Not public. Quote on request. | | 1,000 | $1,000 | $2,000 | Not public. Quote on request. | ![Three figures: $100 for 100 interviews, $500 for 500 and $1K for 1,000, each up to 20 minutes.](/blog/images/brighthire-pricing-1.png) *What Tracy costs at 100, 500 and 1,000 interviews of up to 20 minutes, each with a recording, transcript and assessment. $1K is $1,000. Source: Tracy pricing page, Nov 2026.* Tracy's cost is a number you can compute today: $1 per interview of up to 20 minutes at any volume. ## How does a BrightHire quote convert to Tracy interviews? Our reasoning, not a BrightHire price: a plan priced by company size or volume comes down to one yearly number, and each $1,000 of first-year quote equals 1,000 Tracy interviews of up to 20 minutes. So divide your quote by the interviews you will complete and compare it with $1 per interview of up to 20 minutes. If you only want notes on interviews your team runs, you are buying Interview Intelligence, and the per-result comparison does not apply. For the wider market, see [AI interview cost](/blog/ai-interview-cost). ## What should you ask on a BrightHire quote call? Say which of the three products you need and have your yearly interview volume, your ATS and your regions ready. Then ask: 1. Which products are in the quote, and what does bundling save, in dollars? 2. What is the unit for each, and what is the first-year total at our volume and at twice our volume? 3. What does implementation include, and what are the term and renewal terms? 4. Are ATS connectors and the MCP server in the price? 5. Can we run BrightHire Screen live on our own role before we sign? 6. Can a reviewer get from a score to the sentence the candidate said, and is the recording kept as video or only audio? Tracy answers the money questions on her pricing page, so the call can go to fit. See also [HireVue pricing](/compare/hirevue-pricing). ## How do BrightHire and Tracy compare on the four pillars? Price is one line. The bigger question is what the interview gives you. Tracy cells come from her own pages. BrightHire cells are its statements. | Criterion | Tracy AI Interviewer | BrightHire | What to ask or who leads | | --- | --- | --- | --- | | Depth | Two-way video interview. She asks follow-ups in the person's own field and waits for the answer. Questions come from your brief or your own list, and a coding task can sit inside the interview. | Screen runs async voice or video interviews, generates custom questions and follow-ups, and you approve the questions first. Interview Intelligence records interviews that people run. | Ask for a live run of Screen on your own role, and how many answers get a second question. | | Volume | Every applicant starts on click, at any hour, with no scheduling. Each interview is ready in under 10 seconds. A dropped applicant can rejoin within 10 minutes (configurable per customer). | Screen is async: candidates respond anytime, with no scheduling, on any device. | Ask what happens when 200 applicants start in one day, and how that volume is priced. | | Stack | API and MCP connections (ChatGPT, Claude, Mistral, Gemini, Copilot Studio), Slack, WhatsApp and Zapier, and use beside the Workable and Breezy HR connectors. | 18 ATS named, including Greenhouse, Workday and SAP SuccessFactors, plus Zoom, Google Meet and Teams. Data flows both ways with the ATS for Screen. | Name your ATS and ask which connector is included. | | Evidence | A recording of both sides, a timestamped transcript, a written assessment per question with its reasoning, and integrity signals with their evidence. A person decides. | Screen gives a candidate ranking, scoring, summaries, transcripts and audio playback. Interview Intelligence adds notes and highlight clips. | Ask to open one result and get from the assessment to the sentence the candidate said. | | Price | $1 per credit, one credit per started 20 minutes. First 3 results free, no seats, no upfront cost, and an empty session is never charged. | Not public. Each product is priced on its own and you request a demo. | Ask for the unit, the first-year total at your volume and at twice your volume. | | Implementation and service | Enterprise includes a dedicated customer manager, a setup customised to your workflow and a walkthrough for your team. Tracy is customised in days per customer need. | Says most customers are set up in under a day with no extra fees, and lists dedicated customer success. | Level on service. BrightHire states the faster setup. | | Multilingual | Language count: not stated. A spoken-language check scores fluency, coherence, grammar, pronunciation and vocabulary from 0 to 100 beside the transcript. | Interviews in 30+ languages, live and async. | BrightHire on published languages; Tracy adds a spoken-language check. | | Security and compliance | SOC 2 Type II attested, with the report on request at hello@tracyinterview.ai. | Says SOC 2 Type II certified, and GDPR and CCPA compliant. | Level on SOC 2 Type II. BrightHire also states GDPR and CCPA. | | Validation, bias controls and human oversight | A person reviews the evidence report and decides, and the full history stays on the record. Tracy publishes no bias audit. | Says it has annual independent third-party AI bias audits and an AI governance committee. Recruiters and hiring managers decide. | BrightHire documents more. Tracy publishes no bias audit. | ![Recreated Tracy result report: recording, timestamped transcript and four fields marked Stated, each with a quoted answer.](/blog/images/brighthire-pricing-3.png) *Tracy gives a reviewer each stated point with its quote and the recording behind it, so the $1 buys a result you can check. Illustration with sample data* One credit buys a result a reviewer can check: the person's own words, a timestamp and the recording. ## Which path fits you? [Technical interviews](/use-cases/technical-interviews), [volume hiring](/use-cases/volume-hiring) and [expert vetting](/use-cases/expert-vetting) each have a page. Tracy is free to test: the first 3 results cost nothing and there is no seat to buy. Enterprise is priced to how many interviews you run; book a demo for a figure. ## Common questions ### How much does BrightHire cost? BrightHire publishes no price. Its pricing page names three products, each priced on its own, and you request a demo. We do not estimate a figure. ### What does BrightHire charge by? BrightHire Inbound is priced by applicant volume, BrightHire Screen by interview volume and Interview Intelligence by company size. You can buy one product or bundle them. ### Does BrightHire charge setup fees? It says most customers are set up in less than a day with no extra fees for training or implementation. Get that in writing. ### Is there a free way to try Tracy? Yes. The first 3 results are free, whatever their length, and no card is needed. There are no seats and no upfront cost. --- Try three real interviews free, then calculate your completed-interview cost. Bring your annual volume, interview length and stack and we build the plan, customised in days, with a recording and report you can check. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=brighthire-pricing&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=brighthire-pricing&utm_content=try-free) ## Sources - [BrightHire, Pricing](https://brighthire.com/pricing/), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Connect](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. # Mercor Pricing: Cost vs Tracy (Nov 2026) > Mercor pricing for buyers: what its pages state, how to estimate a cost, and Tracy's published $1 per credit at 100 to 1,000 interviews. - URL: https://tracyinterview.ai/compare/mercor-pricing - Published: 2026-11-13 - Author: Tracy team > **In short** Tracy AI Interviewer publishes its price: the first 3 results free, then $1 per credit, one credit per started 20 minutes, no seats and no upfront cost. > > Mercor's pages read state no buyer price; they name expert staffing and task-based pricing and send buyers to a contact form. > > Tracy interviews applicants you already have, in their field, with a recording and a timestamped transcript behind every result. > > Choose Mercor when you need vetted experts supplied. Mercor pricing for buyers is not stated on the pages read, so a cost comparison has to start from what the pages do say, and from what you would ask on a quote call. This page does both, then sets Tracy's published price beside it. Mercor and [Tracy AI Interviewer](/product) sell different things: Mercor supplies vetted experts, and Tracy interviews the experts you already have. Which one fits depends on which problem you have. Mercor is the better pick when you have no applicants and need people placed with your team. Tracy is the better pick when the applicants exist and the cost you are watching is the cost of interviewing and checking each one. ## What does Mercor pricing look like on its pages? Mercor's enterprise pages send buyers to a contact form and name two ways of working, with no amounts. The table lists what is and is not stated. Source: Mercor's own pages, read 9 Nov 2026. The expert rates are what experts are shown, not a price list for buyers. | What a buyer asks | What Mercor's pages state | Where | | --- | --- | --- | | Price for expert supply | Not stated; buyers are told to "Talk to research" or book a demo | mercor.com home | | How the work is bought | Two ways: expert staffing ("you direct the work") and task-based pricing, "priced per task rather than per head"; no amounts | mercor.com enterprise evals page | | How a buyer starts | "Tell us what you're working on and the right person on our team will be in touch." | mercor.com partner page | | Rates on the expert page | An average contracted rate of $90/hr, and role listings such as $60-$180/hr for a Finance Expert. These are expert-facing; the pages do not say what a buyer pays | mercor.com experts page | | Fees on its data offer | None for a business licensing its data; a different offer from expert supply | mercor.com data page | ## How do you estimate a Mercor cost? You cannot size a Mercor bill from the public pages, so estimate it from your own plan. This is reasoning, not a Mercor figure. **Count the unit.** Under staffing you pay for people over time, so cost follows hours worked and the number of experts. Under task-based pricing cost follows the tasks you order. Neither is described as following the number of applicants you screen, which is the number that drives an interview tool's cost. **Count the volume.** If a vendor sets a minimum or a contract, a few dozen applicants may cost more per applicant than a pay-as-you-go tool. At hundreds of applicants a month the gap depends on terms you cannot see yet. **Add the time inside your team.** Any expert you take on still needs onboarding and review on your side. A fair comparison uses cost per vetted person who passes, not cost per hour or per task, because the number you carry into a budget is the number of people you end up using. ## What should you ask on a Mercor quote call? Put these to Mercor before you compare numbers. - What is the unit I pay for: hours, tasks, placements or seats? - Is there a minimum commitment, a term or an upfront payment? - What is included in the price: sourcing, vetting, replacement if an expert leaves? - How long does an expert stay available, and what happens if a project pauses? - What do I receive as evidence of an expert's skill: a score, a transcript, a recording? - What do renewals and a second project cost? Get the answers in writing, then use the same list for any vendor, Tracy included. ## What does Tracy cost at 100, 500 and 1,000 interviews? Tracy's price is a rule you can read: the first 3 results free, then $1 per credit, one credit for each started 20 minutes of an interview that produced a result. Credits never expire, a session with no answers is never charged, and there are no seats and no upfront cost. Enterprise pricing is set to your volume. The table applies the rule at three volumes. Tracy usage cost at $1 per credit; the first 3 results are free. A 34-minute interview is 2 credits. Cost of a plan, not a quote. | Interviews | 20 minutes each (1 credit) | | --- | --- | | 100 | $100 | | 500 | $500 | | 1,000 | $1,000 | ![Bar chart of Tracy cost: $100, $500, $1,000 for 20-minute interviews; $200, $1,000, $2,000 for 40-minute.](/blog/images/mercor-pricing-1.png) *Tracy's usage cost at 100, 500 and 1,000 interviews of 20 or 40 minutes, so you can size a budget before you commit. Tracy pricing, Nov 2026.* The cost grows in a straight line with interviews and length, so you can budget before you start. ## What is inside the Tracy price? The price buys a result a reviewer can check, not a minute of an interview. Each result comes with a recording of both sides, a timestamped transcript and a written assessment per question. Tracy asks follow-ups in the applicant's own field and waits for the answer, and a person keeps the last word. Applicants start from a link at any hour, so no scheduling cost sits beside the credit. Try three real interviews first and calculate your own cost: [start free](/connect). The [pricing page](/pricing) has more worked examples, and [what an AI interview costs](/blog/ai-interview-cost) covers the wider picture. ## How do Mercor and Tracy compare beyond price? Price is half the decision. The table compares what each vendor does, using only what Mercor's pages state. Mercor column: its own pages, read 9 Nov 2026; Not verified means the pages read do not say. Tracy column: Tracy's published capabilities. | | Mercor, as it describes itself | Tracy | | --- | --- | --- | | Depth | Experts take AI assessments on joining its network | Interviews each applicant in their own field, follows up and waits for the answer | | Volume | 5M+ experts and placement in days (vendor claims) | Each applicant starts on click at any hour; no scheduling, no recruiter present | | Stack | Not verified | API and MCP, Slack, WhatsApp, Workable, Breezy HR, Zapier | | Evidence | Not stated on the pages read | Recording, timestamped transcript and an assessment per question | | Security | Not verified | SOC 2 Type II attested | | Contract and first-year cost | Not stated; ask on a quote call | $1 per credit, no seats, no upfront cost | ## Which fits you, Mercor or Tracy? Mercor wins when you need people, not interviews: a very large expert pool, placement in days and a research team for human evaluation. If sourcing is your bottleneck, a quote from Mercor is the right next step. Tracy wins when sourcing is not the problem and screening cost and evidence are. You can use both: buy supply where you lack it, and interview your own applicants with Tracy. See [AI interview tools for expert vetting](/compare/ai-interview-tools-for-expert-vetting) for a wider shortlist. ## Common questions ### How much does Mercor cost for buyers? Mercor does not state a price on the pages read. Its enterprise pages name expert staffing and task-based pricing, with no amounts, and send buyers to a contact form. ### Is Mercor pricing per hour? The $90/hr average contracted rate and role rates on its experts page are shown to experts. The pages do not say what a buyer pays, or whether staffing is billed hourly. ### What does Tracy cost? $1 per credit, with the first 3 results free. ### How do I compare Mercor with an interview tool? Compare cost per vetted person, not per hour or per task. Ask Mercor for its unit, minimum and what is included, then compute Tracy's cost on your own applicants. ### Can Tracy replace Mercor? Only if you already have the applicants. Tracy interviews them in their field and leaves a recording, transcript and assessment per question. If you need experts supplied, use Mercor. --- Try real results and calculate your cost: bring your volume, interview length and stack, and we build the interview plan for one role in days. The first 3 results are free. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=mercor-pricing&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=mercor-pricing&utm_content=try-free) ## Sources - [Mercor: Enterprise evals](https://mercor.com/enterprise-evals/), retrieved 9 Nov 2026. - [Tracy AI Interviewer: Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. # Metaview Alternatives (Nov 2026) > Metaview alternatives for technical hiring: Tracy, BrightHire, Greenhouse Voice AI, HeyMilo and Alex compared on depth, volume, stack, evidence and price. - URL: https://tracyinterview.ai/compare/metaview-alternatives - Published: 2026-11-13 - Author: Tracy team > **In short** When the first round has to be run, not only noted, Tracy AI Interviewer runs it by two-way video: follow-ups in the person's own field and a report a person decides from. > > $1 per credit, the first 3 results free, no seats. > > BrightHire is the nearest swap. Greenhouse Voice AI, HeyMilo's interview product and Alex lead to quote forms. Stay with Metaview if your team runs its own interviews. > > Next: create one interview for the workflow your current tool misses. Teams look for Metaview alternatives when the job is different from the one Metaview does best. Metaview takes notes on interviews people run; running the first round for every applicant is a different job, and Tracy AI Interviewer is the first tool to try for it. The list puts Tracy first, then four tools for other needs, each with who it suits and where it falls short. ## Why do teams look for Metaview alternatives? These reasons come from Metaview's own pages. The notetaker records interviews people run, and the notes are only as deep as the interviewer's questions. Automating the interview itself means its separate Screening agent. The price stops at the notes. Notetaker, Application Review and Sourcing have list prices, and Screening and the platform bundle do not. Tracy shows hers on her pricing page. Seats scale with headcount: Notetaker Pro is $60 per user per month, so cost follows how many interviewers you have, not how many interviews you run. The stack is ATS-first: Ashby, Greenhouse, Lever, Workday, iCIMS and SmartRecruiters, with no API, MCP server or chat-tool connections stated. Some teams should stay. If your interviewers run the calls and you need ATS scorecards filled from them, Metaview does that well and the tools below mostly do not. ## Which Metaview alternative suits which team? - Tracy AI Interviewer: technical or expert roles where the first round must show depth, at a published price per result. - BrightHire: teams that want notes on their own interviews and an AI interviewer from one vendor, and accept a quote. - Greenhouse Voice AI: teams that want a voice interview inside, or beside, the Greenhouse ATS. - HeyMilo: high-volume hiring that wants voice, video, SMS and forms in one tool. - Alex (formerly Apriora): hiring that needs video, phone, SMS and WhatsApp interviews with identity checks. The sections below give the detail and the limits of each. ## Tracy AI Interviewer: the first round, run and evidenced Tracy interviews every applicant by two-way video. After each answer she asks about what was said, in the person's own field, and waits for them to finish. You can bring your own questions or describe the role and let Tracy draft them, and a coding task can sit inside the interview for technical roles (see [AI technical interviews](/blog/ai-technical-interviews)). Behind every result sit a recording of both sides, a timestamped transcript, a written assessment per question with its reasoning, and integrity signals shown to a human reviewer. A person decides. A Director at a top-tier early-stage VC said Tracy "sometimes brings up exactly those questions I might have forgotten to ask." She starts on click at any hour with no scheduling and connects through API and MCP (ChatGPT, Claude, Mistral, Gemini, Copilot Studio), Slack, WhatsApp, Workable, Breezy HR and Zapier. The price is $1 per credit, the first 3 results free, no seats. See [Tracy's pricing](/pricing). Where others are better: Metaview and BrightHire put notes and scorecards straight into the ATS record for interviews your team runs, and Tracy is not a notetaker for your calls. Metaview names more ATS connectors than Tracy's Connect page does; ATS write-back is not stated there. ![Sample Tracy exchange: a webhook retry question, the candidate's answer, then a follow-up at 03:48.](/blog/images/metaview-alternatives-1.png) *A person reviews the recording, transcript and written assessment per question. Illustration with sample data* Tracy asks the next question from the candidate's answer. A recording and a timestamped transcript sit behind every result, plus a written assessment per question. ## BrightHire: notes and an AI interviewer from one vendor BrightHire sells three products that you can buy separately or together: Inbound, Screen and Interview Intelligence. Its interview side offers live interviews with real-time guidance, AI notes and auto-filled scorecards, interviewer coaching, and fraud signals in live interviews with candidate consent. Its AI interviewer "runs structured screening interviews, at scale," with candidates able to interview 24/7 in a two-way conversation scored against your rubrics. It integrates with Greenhouse, Lever, Ashby and Workday, and with Zoom, Microsoft Teams, Google Meet and Slack. BrightHire lists no dollar amount: Screen is priced by interview volume and Interview Intelligence by company size. Pick it when you want notes and an AI interviewer from one vendor. ## Greenhouse Voice AI: a voice screen with a wide ATS list Greenhouse Voice AI is a structured, two-way voice interview scored against your own rubric, and it works with or without the Greenhouse ATS. It asks follow-up questions drawn from the rubric, answers candidate questions from team-approved information, and leaves a score, recording, transcript and per-question explanations. It conducts interviews in 11 languages and integrates with 60+ ATS providers. It is a two-way voice interview, and video is not stated. Its price is an annual subscription scaled to your organization with no public figure. Pick it when your team already runs on Greenhouse or wants a voice screen with a wide ATS list. See [Greenhouse AI interviewer vs Tracy](/compare/greenhouse-ai-interviewer-vs-tracy). ## HeyMilo: voice, video, SMS and forms in one tool HeyMilo screens and interviews every applicant over voice, video, SMS and forms. The recruiter sets questions, rubrics and weights, and the AI asks follow-ups tied to the rubric, with a per-question score, rationale, transcript and recording. It syncs with 20+ ATS partners, supports 20+ languages and has an MCP server and public API. The recruiter makes the final call. HeyMilo's interview pricing page leads to a demo form and states no price; plans appear in the product after billing is set up. Pick it for high-volume screening across several channels, then ask for the rate. See [Tracy vs HeyMilo](/compare/tracy-vs-heymilo). ## Alex: video, phone, SMS and WhatsApp with identity checks Alex, formerly Apriora, runs two-way interviews over video, phone, SMS and WhatsApp, asks follow-ups based on what the candidate says and keeps a transcript and score as the audit trail. It layers fraud checks (application fraud risk, in-interview proctoring and a government ID plus selfie check), lists 33+ ATS integrations and 30+ languages, and does not automate hiring decisions. Alex publishes no price: its pricing page leads to a demo form. Pick it when WhatsApp or phone interviews with identity verification are the requirement. See [Tracy vs Alex](/compare/tracy-vs-alex). ## Metaview and its alternatives compared on four pillars and price Source: Tracy approved product records; vendor pages read 3 to 9 Nov 2026. A blank price means no published figure. | Tool | Depth | Volume | Stack | Evidence | Price | What to test | | --- | --- | --- | --- | --- | --- | --- | | Tracy AI Interviewer | Two-way video, follow-ups in the person's field, coding task inside | Starts on click, any hour, no scheduling | API, MCP, Slack, WhatsApp, Zapier | Recording, timestamped transcript, assessment per question | $1 per credit, first 3 free | Second question after a thin answer | | Metaview | Notetaker on human interviews; Screening is a voice agent with follow-ups | Screening starts when the candidate chooses | ATS: Ashby, Greenhouse, Lever, Workday and more | Recorded, transcribed, rubric-scored; risk score | Notetaker $60 per user per month; Screening by quote | Screening run on your role | | BrightHire | AI notes on live interviews; AI interviewer, two-way | Candidates can interview 24/7 | Greenhouse, Lever, Ashby, Workday, Slack | Scored against your rubrics; auto-filled scorecards | By interview volume, no figure | Quote for Screen | | Greenhouse Voice AI | Voice, rubric-anchored follow-ups | Annual plan with a fair-use allowance | 60+ ATS providers | Score, recording, transcript, explanations | Quote, no figure | Run on a technical role | | HeyMilo | Voice, video, SMS, forms; follow-ups tied to rubric | Re-engages drop-offs by email, SMS or callback | 20+ ATS partners, MCP server, API | Score, rationale, transcript, recording | Interview pricing page: demo form, no figure | Coding is not described | | Alex | Video, phone, SMS, WhatsApp; follow-ups on answers | Calls applicants within minutes | 33+ ATS integrations | Transcript, score, proctor flags with timestamps | Demo form, no figure | Coding is not described | ![Three steps: who runs the round, what price you can see, does it run on your role.](/blog/images/metaview-alternatives-2.png) *Illustration, no data: three questions that sort a shortlist of interview tools. Illustration, no data.* Three questions sort the shortlist: who runs the round, what price you can see, and whether it runs on your role. ## How do you choose a Metaview alternative? Start with who runs the interview. If your recruiters run it and you want notes and scorecards, stay with Metaview or look at BrightHire. If you want it run for every applicant, shortlist Tracy, Greenhouse Voice AI, HeyMilo and Alex. Then check the price. Only Tracy states a per-interview price. For the rest, ask for the cost of a 20-minute screen at 100, 500 and 1,000 candidates, in writing, and set it beside $1 per credit. Finally, ask each vendor to run one real role. Read the second question the AI asks, and open the recording behind one result. Create one interview for your role and read [what an AI interview report contains](/blog/ai-interview-reports) and [the best AI interview software](/blog/best-ai-interview-software) for the wider list. ## Which path fits you? - Technical hiring: [technical interviews](/use-cases/technical-interviews) covers role-specific interviews with follow-ups. - High-volume hiring: [volume hiring](/use-cases/volume-hiring) covers applicant waves. - Integrations: [Connect](/connect) lists the integrations. ## Common questions ### What is the best Metaview alternative? It depends on the job. Tracy AI Interviewer suits technical and expert roles that need an AI-run first round with evidence at a published price. BrightHire is the nearest like-for-like swap, Greenhouse Voice AI fits Greenhouse teams, HeyMilo covers several channels and Alex covers WhatsApp and phone. ### Who are Metaview's competitors? Among the tools on this page, BrightHire competes most directly on notes plus an AI interviewer. Greenhouse Voice AI, HeyMilo, Alex and Tracy compete with Metaview's Screening agent. ### Is Tracy a notetaker like Metaview? No. Tracy runs the interview and writes the report. For notes on calls your team runs, use a notetaker. ### Which Metaview alternative publishes its price? Tracy: $1 per credit, one credit per started 20 minutes. BrightHire, Greenhouse Voice AI, HeyMilo's interview product and Alex lead to a quote or demo form, and Metaview states no Screening price. --- Create one interview for the workflow your current tool misses: send us one role and we will build the interview plan, with a recording and report you can check. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=metaview-alternatives&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=metaview-alternatives&utm_content=try-free) ## Sources - [Metaview, Screening](https://www.metaview.ai/screening), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Connect](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. # Metaview Pricing (Nov 2026) > Metaview pricing in Nov 2026: Notetaker Pro $60 per user, Application Review $150, Screening unpriced. Tracy's AI interview is $1 per credit. - URL: https://tracyinterview.ai/compare/metaview-pricing - Published: 2026-11-13 - Author: Tracy team > **In short** Tracy AI Interviewer is priced per completed interview: $1 per credit, one credit per started 20 minutes, the first 3 results free, no seats. 100 interviews of up to 20 minutes cost $100 and 1,000 cost $1,000. > > Each result carries a recording, a timestamped transcript, an assessment per question and integrity signals, and a person decides. > > Metaview lists $60 per user per month for Notetaker and no Screening price. If your own people run every interview, the Notetaker is cheaper. > > Next: try three interviews free and calculate your cost per completed interview. Metaview pricing is public for Notetaker, Application Review and Sourcing, and not for Screening or the platform bundle. You can budget the notes side today. What an AI-run first round costs, you cannot, because Screening has no stated price. **Choose Metaview when** your recruiters and hiring managers run the interviews and you want notes, scorecards and ATS sync across them, at $60 per user per month or free for light use. **Choose Tracy AI Interviewer when** you want the interview itself run for every applicant, with a price you can compute before any call: $1 per credit, the first 3 results free, no seats. She starts on click at any hour, with no scheduling, and the cost follows interviews, not headcount. ## Metaview pricing table: every plan the pages list Source: Metaview pricing and interview-notes pages, 9 Nov 2026. Prices are Metaview list prices; billing terms and discounts are not stated on the pages. | Plan | Price | What it lists | Page words | | --- | --- | --- | --- | | Notetaker Free | $0 | 25 calls per month, basic transcription; calls from the last 14 days | "25 calls per month" | | Notetaker Pro | $60 per user per month | Unlimited calls, priority processing | "Pro is $60 per user per month with unlimited calls." | | Notetaker Enterprise | Tailored | Dedicated CSM and enterprise security features | "priced on request" | | Application Review Free | $0 | 50 applications per month, basic screening | "50 applications per month" | | Application Review Pro | $150 per month | 500 reviews per month, unlimited users, advanced AI evaluation | "500 reviews / month" | | Sourcing Free | $0 per user | First 100 profiles sourced | "First 100 profiles sourced free" | | Sourcing Pro | $100 per user per month | 200 profiles sourced per month | "200 profiles sourced per month" | | Sourcing Max | $300 per user per month | Unlimited profiles sourced | "Unlimited sourcing for high-volume hiring." | | Screening | Not stated | No pricing tab and no price on its page | "Screen every candidate without picking up the phone." | | Agentic Recruiting Platform | Custom | All agents and premium features; button reads Speak to Sales | "all agents and premium features" | ## What costs extra on Metaview Metaview says every user gets the full platform but "you only use and pay for the agents you need." So the table above is a menu, and the bill is the sum of the agents you pick. Notetaker Free gives access to calls from the last 14 days; the Pro retention window is not stated. Reports is an add-on for "a flat fee," with no figure shown. Assistant questions and note regenerations are included. ## What a typical team pays for Metaview Notetaker Take a team where five people run interviews and all five need the notetaker. Five users at $60 is $300 a month, or $3,600 over twelve months at the monthly rate. A team of 20 interviewers is $1,200 a month. Add Application Review Pro and it is $150 more a month for the whole team, since that plan lists unlimited users. This is our arithmetic from list prices, not a Metaview quote. Notetaker Pro lists unlimited calls, so 100 calls and 1,000 calls cost the same $300 for those five seats. ## What Tracy costs at 100, 500 and 1,000 interviews A credit costs $1 and covers each started 20 minutes of an interview that produced a result. The first 3 results are free with no card, and there are no seats. The full model is on [Tracy's pricing page](/pricing), and [AI interview cost](/blog/ai-interview-cost) works through more examples. Enterprise is priced to your volume and adds a dedicated customer manager, a dedicated support channel and a setup customised to your workflow in days. ![Three steps: one credit per started 20 minutes, evidence per result, a person decides.](/blog/images/metaview-pricing-2.png) *Illustration: what one Tracy credit buys. A recording and a timestamped transcript sit behind every result, with a written assessment per question, and a person decides. Illustration with sample data* One credit covers an interview with a recording, a transcript and a written assessment per question behind the result. ## Cost at 100, 500 and 1,000 interviews a month: Tracy next to Metaview Notetaker Interviews of up to 20 minutes on Tracy. The two columns buy different work: Tracy runs the interview, the notetaker records one a person runs. Arithmetic from list prices. | Interviews per month | Tracy runs the interview | Metaview Notetaker Pro, 5 interviewers, people run the calls | | --- | --- | --- | | 100 | $100 | $300 a month, unlimited calls | | 500 | $500 | $300 a month, unlimited calls | | 1,000 | $1,000 | $300 a month, unlimited calls | ## Is a seat price cheaper than an interview price? Our reasoning, not a Metaview price: the seat price leaves out the hours. 1,000 twenty-minute first-round calls take a person about 333 hours before scheduling. If your people run those calls anyway, $300 for the notes is cheap. If the hours are the problem, Tracy's $1,000 buys the interview itself, started on click at any hour. See [hiring at volume](/use-cases/volume-hiring). Try your first 3 results free and see the cost of your own completed interviews. ## What to ask on a Metaview Screening quote call Metaview's Screening agent is a voice agent that talks to candidates, asks follow-ups and scores the call against your rubric. Its page lists no price, and the pricing page has no Screening tab. Ask these on the quote call: 1. Is Screening priced per candidate, per call, per role or per seat? 2. What does a 20-minute screen cost at 100, 500 and 1,000 candidates? 3. Is Notetaker included, or billed per user? 4. Which agents does the platform price include, and which are add-ons such as Reports? 5. What are the annual terms, minimum commitment and renewal uplift? 6. Can you run Screening on one of our roles before we sign? A written answer to the second question gives you a per-interview figure to set beside $1 per credit. ## Beyond price: how the published workflows differ on four pillars Source: Tracy approved product records; Metaview pages, 9 Nov 2026. | Pillar | Tracy AI Interviewer | Metaview | Who leads | | --- | --- | --- | --- | | Depth | Two-way video interview. After each answer she asks about what was said, in the person's own field, and a coding task can sit inside the interview. | Screening is a voice agent that asks follow-ups and can show code for the candidate to explain aloud. Notetaker records interviews people run. | Even on the published pages. Ask both for a live run on your role. | | Volume | Starts on click at any hour, no scheduling, no recruiter present. | Candidates take the Screening call when it suits them on any device. Metaview says it has the most impact where 50+ candidates apply per opening. | Even on the published pages. Metaview says Screening fits roles with 50+ applicants per opening; Tracy starts on click at any hour. | | Stack | API and MCP (ChatGPT, Claude, Mistral, Gemini, Copilot Studio), Slack, WhatsApp, Workable, Breezy HR, Zapier. | ATS and CRM: Ashby, Greenhouse, Lever, Gem, SmartRecruiters, Workday, iCIMS and more; scorecards fill from notes. API and MCP are not stated. | Metaview for ATS scorecards. Tracy for API, MCP and chat tools. | | Evidence | Recording of both sides, timestamped transcript, written assessment per question, integrity signals shown to a person. | Screening calls are recorded, transcribed and scored against a rubric; a risk score comes with an explanation; an audit trail of AI reasoning. | Even. Both leave the decision to a person. | ## What the four pillars leave out Source: Tracy approved product records; Metaview pages, 9 Nov 2026. "Not stated" and "Not verified" mean the page does not say, not that the feature is absent. | Criterion | Tracy AI Interviewer | Metaview | Who leads | | --- | --- | --- | --- | | Implementation | Setup customised to your workflow in days; questions drafted from a job description. | Most teams are up and running the same day. | Metaview on stated speed. | | Accessibility | Not stated. | Not stated on the pages read. | Neither. Ask both. | | Multilingual | A spoken-language check scores fluency, coherence, grammar, pronunciation and vocabulary. | Screening lists most major languages, with branding and a welcome video. | Metaview on its published list. | | Security | SOC 2 Type II attested; report on request. | SOC 2 Type II, GDPR, CCPA, LGPD and PIPEDA badges. | Metaview on the regimes listed. | | Validation and bias | Tracy publishes no bias audit. A person decides from the evidence report. | Not verified. Metaview says a human makes the final decision. | Not verified. | | Reporting | A report with resume match, score, strengths and gaps, recording and fit rationale, shared by link or PDF. | Reports on question coverage, interview adherence and talk-time, as a flat-fee add-on. | Metaview across many interviews. | | Contract and first-year cost | $1 per credit, no seats, no upfront cost, credits never expire. | Per-user list prices for three agents; Screening and platform by quote. | Tracy for a number you can compute today. | | Support | Enterprise adds a dedicated customer manager and support channel. | Forward Deployed Talent Engineers; dedicated CSM on Enterprise. | Even. | ## Is Metaview worth it, and for whom? On speed, not price: one customer cut its average time to hire from 44 days to 7, and Tracy did the work of 53 full-time recruiters in one month. The verdict for a budget owner: if the plan is to keep humans on the call, price the Notetaker. If the plan is to have the first round run for you, put the Screening quote next to [Tracy's pricing](/pricing). ## Which path fits you? - Technical hiring: [technical interviews](/use-cases/technical-interviews) covers role-specific interviews with follow-ups. - High-volume hiring: [volume hiring](/use-cases/volume-hiring) covers applicant waves. - Integrations: [Connect](/connect) lists what Tracy connects to. - The report: [AI interview reports](/blog/ai-interview-reports) shows what a reviewer gets. ## Common questions ### How much does Metaview cost? Notetaker is free for 25 calls a month and $60 per user per month for unlimited calls. Other agents and Screening are in the table above. ### How does Tracy's pricing compare with Metaview's? Tracy charges $1 per credit, one credit per started 20 minutes, with the first 3 results free and no seats. 100 interviews of up to 20 minutes cost $100. The two buy different work: Tracy runs the interview, and Notetaker records interviews your team runs. ### What does Metaview Screening cost? Metaview does not state a Screening price. Ask for the cost of a 20-minute screen at 100, 500 and 1,000 candidates, in writing. --- Try three real interviews and calculate your completed-interview cost: send us one role, your annual volume, interview length and stack, and we will build the interview plan, with a recording and report you can check. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=metaview-pricing&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=metaview-pricing&utm_content=try-free) ## Sources - [Metaview, Pricing](https://www.metaview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Connect](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. # Tracy vs BrightHire: Run the First Round or Record It? (Nov 2026) > BrightHire vs Tracy: notes on interviews your team runs, or an AI interviewer for the first round. Where each wins on depth, stack, evidence and price. - URL: https://tracyinterview.ai/compare/tracy-vs-brighthire - Published: 2026-11-13 - Author: Tracy team > **In short** Tracy AI Interviewer runs the first round for every applicant: a two-way video interview on click, with follow-ups in the person's field and a recording, transcript and assessment per question. > > She costs $1 per credit, with the first 3 results free and no seats. > > BrightHire is the better pick if your team runs the interviews and wants notes, ATS scorecards and a bias audit. It also sells BrightHire Screen. > > Next: test both on one role. BrightHire calls itself the leader in interview intelligence: its core product records and writes up interviews that people run. It also sells BrightHire Screen, an AI interviewer. Tracy AI Interviewer runs the first round herself, for every applicant, and leaves a recording, a transcript and an assessment per question. The BrightHire vs Tracy choice comes down to who runs the first round; background is in [what an AI interviewer is](/blog/what-is-an-ai-interviewer). ## What does BrightHire do well? Interview Intelligence "records and transcribes candidate interviews and turns the conversations into structured notes." It keeps highlight clips that play in your ATS, fills your ATS scorecard fields from the notes, opens inside Zoom with no per-call bot, and sends consent notifications before interviews. BrightHire Screen runs async voice or video interviews, generates custom questions and follow-ups (you approve the questions first) and scores each interview against your rubric. BrightHire lists 18 ATS, says it is SOC 2 Type II certified, states annual independent third-party AI bias audits, supports 30+ languages, and sets up most customers in under a day with no extra fees. If your interviewers run most calls and the pain is notes and consistency, that is a mature answer. ## BrightHire vs Tracy: buyer criteria side by side Tracy cells come from Tracy's own pages. BrightHire cells are BrightHire's statements. | Criterion | Tracy AI Interviewer | BrightHire | Who leads | | --- | --- | --- | --- | | Who runs the interview | Tracy runs the first round. Each applicant gets a two-way video interview with no recruiter present. | Interview Intelligence records interviews that people run. Screen runs async voice or video interviews. | Depends on the job. | | Depth | Follow-ups in the person's own field, after she waits for the end of the answer. Your own questions or ones drafted from the role, and a coding task inside the interview. | Screen generates custom questions and follow-ups; you approve the questions first. Coding task: not stated on its AI interviewer page. | Level on follow-ups. Tracy for a coding task. | | Volume and start | Every applicant starts on click, at any hour, with no scheduling. Ready in under 10 seconds; a dropped applicant can rejoin within 10 minutes (configurable per customer). | Screen is async: candidates respond anytime, with no scheduling, on any device. | Level. | | Evidence per result | Recording of both sides, a timestamped transcript and a written assessment per question with its reasoning. A person decides. | Screen gives a ranking, scoring, summaries, transcripts and audio playback. Interview Intelligence adds notes and highlight clips. | Different. Open one result in each. | | Integrity signals | Tab changes, gaze, live face detection, delay before answers, fullscreen exits and external screens, shown as evidence to a reviewer. Not proof. | Signals from applications, AI interviews and live interviews. They trigger human review, never automated rejection. | BrightHire covers more stages. | | Price | $1 per credit, one credit per started 20 minutes. First 3 results free, no seats, no upfront cost. | No price published. Each product is priced on its own: Screen by interview volume, Interview Intelligence by company size. | Tracy for a published price. | | ATS and stack | API and MCP (ChatGPT, Claude, Mistral, Gemini, Copilot Studio), Slack, WhatsApp and Zapier, and use beside the Workable and Breezy HR connectors. Workday and SAP SuccessFactors connectors: not stated on Tracy's Connect page. | 18 ATS named, including Greenhouse, Workday and SAP SuccessFactors, plus Zoom, Google Meet and Teams. An MCP server for AI tools. | BrightHire on ATS depth. | | Implementation and support | Enterprise: a dedicated customer manager, a dedicated support channel, a setup customised to your workflow and a team walkthrough. Customised in days. | Most customers set up in under a day with no extra fees; dedicated customer success. | Level. BrightHire states the faster setup. | | Security, bias and oversight | SOC 2 Type II attested; report on request at hello@tracyinterview.ai. A person reviews the evidence and decides. Tracy publishes no bias audit. | Says SOC 2 Type II certified, GDPR and CCPA compliant, annual third-party AI bias audits. Recruiters and hiring managers decide. | BrightHire documents more. | | Languages and accessibility | A spoken-language check scores fluency, coherence, grammar, pronunciation and vocabulary from 0 to 100. Accessibility accommodations: Not stated. | Interviews in 30+ languages. Accessibility: Not stated. | BrightHire on languages. | | Reporting | A report per interview with an assessment per question, shareable by link or PDF. | Scorecards filled from notes, a ranking from Screen, insights across interviews. Audit logs: Not verified. | BrightHire on insights across interviews. | ## How the published workflows differ for the same technical role Take an invented role, a data engineer on a reporting team. BrightHire Screen's published description is a structured interview from your job description, async, with follow-ups. Here is a chain Tracy could run, with an invented candidate. The figure is an illustration, not a recording. > Tracy: Your nightly load sometimes misses rows that arrive late. How do you design for that? > > Candidate: We re-read the last two days and merge on the key, so late rows land in the right day. > > Tracy: You said the nightly load re-reads two days. What happens to a report that was already sent? a Director at a top-tier early-stage VC describes the same effect: "Tracy sometimes brings up exactly those questions I might have forgotten to ask, which has been a pleasant surprise." For engineers, a coding task inside the interview gets follow-ups on approach and trade-offs; see [AI technical interviews](/blog/ai-technical-interviews). ![Recreated Tracy screen: a follow-up asks what happens to a report that was already sent after a two-day re-read.](/blog/images/tracy-vs-brighthire-1.png) *Tracy asks a follow-up on the detail the candidate just gave, so a reviewer sees depth, not a recited answer. Illustration with sample data* At 06:12 Tracy asks about the detail the candidate just gave, and the exchange is recorded with a timestamp. ![Three takeaways on recording interviews your team runs versus having Tracy run the first round.](/blog/images/tracy-vs-brighthire-2.png) *Tracy runs the first round and leaves a recording, transcript and assessment; BrightHire records interviews your team runs. Illustration with sample data. Source: BrightHire and Tracy product pages, Nov 2026.* Your team runs the calls, BrightHire's notes fit; applicants outnumber the calls, Tracy runs the first round. ## How does Tracy handle integrity and volume? Integrity signals (tab changes, gaze, live face detection, delays before answers, fullscreen exits, external screens) are shown to a reviewer as evidence, not proof; see [AI interview cheating detection](/blog/ai-interview-cheating-detection). Every applicant starts on click at any hour; see [hiring at volume](/use-cases/volume-hiring). ## What do BrightHire and Tracy cost? BrightHire publishes no price. Its pricing page says "Request a demo and we'll put together a plan that fits your hiring volume and team size." Tracy's is published: $1 per credit covers each started 20 minutes of an interview that produced a result, so 100 interviews of up to 20 minutes cost $100. The first 3 results are free, there are no seats, and a session with no answers is never charged. Enterprise is priced to how many interviews you run; book a demo for a figure. Our reasoning, not a BrightHire price: BrightHire says Screen is priced by interview volume, so ask for the price per automated interview and multiply by your next 100 applicants; on Tracy the same 100 interviews of up to 20 minutes are $100. See [Tracy's pricing](/pricing). ## How do you test both before you commit? Pick one role with at least 20 expected applicants. Ask BrightHire for a live Screen run on it and get the quote in writing, and ask whether it covers Screen and Interview Intelligence separately. In Tracy, describe the role and edit the drafted questions, or send us the role and we build the plan; the first 3 results are free. Run one role in Tracy and compare the follow-ups and the report with BrightHire Screen's. Have one reviewer read the reports from your next 20 applicants and count the follow-ups that landed, the results verified from the recording in under a minute, and the cost per interview. For the mechanics, read [how AI interviews work](/blog/how-do-ai-interviews-work). Which path fits you: [technical interviews](/use-cases/technical-interviews), [volume hiring](/use-cases/volume-hiring), [expert vetting](/use-cases/expert-vetting), or [Connect](/connect) for integrations. ## Common questions ### Does BrightHire have an AI interviewer? Yes. BrightHire Screen runs async interviews by voice or video, generates custom questions and follow-ups, and scores them against your rubric. Its Interview Intelligence product records interviews that people run. ### How much does BrightHire cost? BrightHire publishes no price. Each product is priced on its own and you request a demo. Tracy's price is on her pricing page: $1 per credit. ### Is BrightHire better than Tracy? For teams whose interviewers run the calls, yes: notes, highlights, ATS scorecards, a bias audit and 30+ languages. Tracy is better when the first round should run for every applicant, with follow-ups, a coding task and a published price. ### Does Tracy connect to my ATS? Tracy connects through API and MCP, Slack, WhatsApp and Zapier, and works beside the Workable and Breezy HR connectors. Native Workday or SAP SuccessFactors connectors are not stated on her Connect page; ask us about your ATS. --- Run one role in Tracy and compare the follow-ups and report. Send us the role and we build the interview plan, customised in days, with a recording and report you can check. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=tracy-vs-brighthire&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=tracy-vs-brighthire&utm_content=try-free) ## Sources - [BrightHire, Pricing](https://brighthire.com/pricing/), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Connect](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. # Tracy vs Metaview: Run the First Round or Take Notes? (Nov 2026) > Tracy vs Metaview: notes on interviews your team runs, or an AI interviewer that runs the first round. Where each wins on depth, stack, evidence, price. - URL: https://tracyinterview.ai/compare/tracy-vs-metaview - Published: 2026-11-13 - Author: Tracy team > **In short** Tracy AI Interviewer runs the first round for every applicant: two-way video, follow-ups in the person's own field, started on click at any hour. A person decides from her recording, transcript and assessment per question. > > Pricing is $1 per credit with the first 3 results free and no seats. > > Choose Metaview when recruiters keep running the interview and you want notes that fill Greenhouse, Ashby, Lever or Workday scorecards. > > Next: test both on one role. The choice in Tracy vs Metaview is between running the first round and taking notes on it. Metaview's notetaker joins interviews that people run, writes the notes and fills the scorecard. Tracy AI Interviewer runs the interview itself and writes the report. Metaview also sells an AI Screening agent, so the line is thinner than the headline suggests. Metaview is the stronger choice for teams that keep humans on the call. ## What does Metaview do well? Metaview wins when your interviews are people talking to people. The notetaker supports Google Meet, Microsoft Teams, Webex and Zoom, joins scheduled interviews from your calendar, handles phone interviews, and can take notes without a bot in the call. It maps answers into scorecards on Greenhouse, Ashby, Lever, Workday and more, and Metaview says interviewers turn feedback around 48 hours faster. Coding and system design interviews are captured with what was shared on screen. It goes past notes: agents for sourcing, application review, screening and reports, plus reporting on question coverage and talk-time. It lists SOC 2 Type II and GDPR, says most teams are up the same day, and has a free Notetaker plan. For a team that runs its own interviews it is the stronger tool, because Tracy does not join your calls, and ATS write-back is not stated on her Connect page. ## Does Metaview run the first round too? Metaview's Screening agent talks to candidates by voice, asks follow-ups, can show code for the candidate to explain aloud, scores the call against your rubric, and syncs the human decision to your ATS. Its page does not describe a video interview and lists no price. Tracy runs a two-way video interview. After each answer she asks about what was said, in the person's own field, and waits for the person to finish. For technical roles a coding task sits inside the same interview with follow-ups on approach and trade-offs. See [AI technical interviews](/blog/ai-technical-interviews) for how that works. ## Tracy vs Metaview: buyer criteria side by side Source: Tracy approved product records; Metaview pages, 9 Nov 2026. "Not stated" and "Not verified" mean the page does not say, not that the feature is absent. | Criterion | Tracy AI Interviewer | Metaview | Who leads | | --- | --- | --- | --- | | Interview depth | Two-way video, follow-ups in the person's own field, a coding task inside the interview. Questions from your brief or your own list. | Notetaker records interviews people run. Screening is a voice agent with follow-ups, a rubric and code read aloud. | Even on the pages. Run both on one role. | | Who runs the call | Tracy runs it, with no recruiter present. | Your people run it and Metaview takes notes; or the Screening agent runs it. | Metaview if your team keeps the call; Tracy if it should not. | | Volume | Starts on click at any hour, no scheduling. | Screening starts when the candidate chooses; strongest where 50+ candidates apply per opening. | Level on start. Metaview states Screening works best from 50+ applicants per opening. | | Stack | API and MCP (ChatGPT, Claude, Mistral, Gemini, Copilot Studio), Slack, WhatsApp, Workable, Breezy HR and Zapier. ATS write-back is not stated on Tracy's Connect page. | Ashby, Greenhouse, Lever, Workday, iCIMS, SmartRecruiters and more; notes and scorecards post to the candidate record. | Metaview on ATS depth. Tracy on API and chat tools. | | Evidence | Recording of both sides, timestamped transcript, assessment per question with reasoning, integrity signals shown to a person. | Recorded, transcribed, scored on a rubric; High, Medium or Low risk score with an explanation; audit trail of AI reasoning. | Even. | | Implementation | Setup customised to your workflow in days. | Most teams are up and running the same day. | Metaview on stated speed. | | Accessibility | Not stated. | Not stated on the pages read. | Neither. | | Multilingual | A spoken-language check scores fluency, coherence, grammar, pronunciation and vocabulary. | Screening lists most major languages. | Metaview on its published list. | | Security | SOC 2 Type II attested. | SOC 2 Type II, GDPR, CCPA, LGPD, PIPEDA. | Metaview on the regimes listed. | | Validation and bias | Tracy publishes no bias audit. A person decides from the evidence report. | Not verified. A human makes the final decision. | Not verified. | | Reporting | A report with resume match, score, strengths and gaps, recording and fit rationale, shared by link or PDF. | Reports on question coverage, interview adherence and talk-time. | Metaview across many interviews. | | Contract and first-year cost | $1 per credit, first 3 results free, no seats, no upfront cost. | Notetaker free, then $60 per user per month. Screening and platform by quote. | Tracy on a price you can compute. | | Support | Enterprise adds a dedicated customer manager and support channel. | Forward Deployed Talent Engineers; dedicated CSM on Enterprise. | Even. | ## How the published workflows differ for the same technical role Take one invented role, a site reliability engineer on a payments team. The table sets Metaview's published description of its notetaker beside an illustration of how a Tracy interview would run. The candidate is fictional, and the Tracy column is an illustration, not a recording of a real interview. ![Sample Tracy exchange: an alert question, the candidate's answer, then a follow-up on the alert rule.](/blog/images/tracy-vs-metaview-1.png) *The third question follows from the second answer. Illustration with sample data* Tracy asks the next question herself, from what the candidate just said. ## The same role, step by step Metaview column: published descriptions. Tracy column: illustration of the product workflow. | Step | Metaview notetaker (from its pages) | Tracy (illustration) | | --- | --- | --- | | Setup | Connect your calendar and ATS; the notetaker joins scheduled interviews. Notes follow your templates and scorecards. | You describe the role and Tracy drafts the questions. You edit them or bring your own. | | Who is on the call | An interviewer from your team. The notetaker joins, with or without a bot. | Tracy and the applicant. No recruiter is present. | | Questions | The interviewer chooses them; Metaview tracks question coverage against the template. | Questions drafted from the role or your own list, which you can edit. | | Follow-up | The human interviewer asks it. | Tracy asks about what was said and waits for the answer. | | Output | Notes, a filled scorecard and optional video snippets in your ATS. | Recording, timestamped transcript, assessment per question and integrity signals in one report. | | Decision | The interviewer and hiring team. | A person, from the report. A recruiter decides: they can move the person forward or invite them again. | ## Do follow-ups matter if the notes are good? A notetaker's depth is the interviewer's depth. If your recruiters know the field, that is fine, and Metaview makes their calls easier to review. When they screen a role they do not know, or applicants outnumber the calls people can make, the depth has to come from the follow-up: the second question comes from the first answer, in the person's own field. A Director at a top-tier early-stage VC said Tracy "sometimes brings up exactly those questions I might have forgotten to ask, which has been a pleasant surprise." Metaview's Screening agent also asks follow-ups, so ask both vendors for a live run on your role and read the second question. ## Can a reviewer check a Metaview or a Tracy result? Both can leave a reviewer something to check, in different ways. Metaview records and transcribes Screening calls, scores them against your rubric and keeps an audit trail of the AI's reasoning. Tracy leaves the recording of both sides, a timestamped transcript and an assessment per question with its reasoning, so a reviewer can open the moment behind a result. See [AI interview reports](/blog/ai-interview-reports). Integrity works differently. Metaview gives a High, Medium or Low risk score with an explanation. Tracy captures tab changes, gaze, live face detection, delays before answers, fullscreen exits and external screens, and shows the evidence to a person. These are signals for a human reviewer to interpret, not proof, and nothing is decided in the interview. Candidates are told on the join screen that the conversation is recorded and transcribed. See [AI interview cheating detection](/blog/ai-interview-cheating-detection). ## Does Tracy hold up at volume? One customer cut its average time to hire from 44 days to 7, and Tracy did the work of 53 full-time recruiters in one month. Metaview says Screening has the strongest impact where 50+ candidates apply per opening, which is the same situation. ![Three steps: your team runs the call, volume outgrows the calendar, or both tools together.](/blog/images/tracy-vs-metaview-2.png) *Which job each tool does: Tracy runs the first round, a notetaker covers calls your team runs. Illustration, no data. Illustration, no data.* Metaview takes notes on your calls and Tracy runs the first round; you can use both. ## What do Metaview and Tracy cost? Metaview prices per seat and per agent: Notetaker is free for 25 calls a month and $60 per user per month; Screening is not priced on its pages. Tracy's price is published, and it prices a different unit. A credit costs $1 and covers each started 20 minutes of an interview that produced a result. The first 3 results are free with no card and there are no seats. 100 interviews of up to 20 minutes cost $100. Enterprise is priced to your volume and adds a dedicated customer manager, a dedicated support channel and a setup customised to your workflow in days. See [Tracy's pricing](/pricing). ## How do you test both before you commit? Run one role, one team and two weeks. For the mechanics first, read [how AI interviews work](/blog/how-do-ai-interviews-work). 1. Pick one technical role with at least 20 expected applicants. 2. In Metaview, connect the notetaker to your next interviews, and ask for a live run of Screening on that role plus a written quote for it. 3. In Tracy, describe the role and edit the drafted questions, or send us the role and we build the plan. The first 3 results are free, so start with three internal volunteers. 4. Have one reviewer read the output of your next 20 applicants in each. Count the follow-ups that landed, the applicants who finished and the cost per interview. ## Which path fits you? - Technical hiring: [technical interviews](/use-cases/technical-interviews) covers role-specific interviews with follow-ups. - High-volume hiring: [volume hiring](/use-cases/volume-hiring) covers applicant waves. - Recruiting teams: [people teams](/use-cases/people-teams) covers screening for HR-led teams. - Integrations: [Connect](/connect) lists the integrations. ## Common questions ### What is Metaview? Metaview calls itself the Agentic Recruiting Platform. It sells a notetaker for interviews alongside sourcing, application review, screening and reports agents. ### Is Metaview an AI notetaker or an AI interviewer? Both. Its Notetaker records interviews people run, and its Screening agent is a voice agent that talks to candidates. Tracy AI Interviewer runs the interview itself, by two-way video. ### Can I use Metaview and Tracy together? They do different jobs, so you can run both: Tracy for the first round, a notetaker on later calls your team runs. Tracy connects by API, MCP, Slack, WhatsApp and Zapier. ### Which is cheaper, Metaview or Tracy? They price different units. Metaview Notetaker is $60 per user per month. Tracy is $1 per credit, with the first 3 results free. Metaview does not state a Screening price. --- Run one role in Tracy and compare the follow-ups and report: send us the role and we will build the interview plan, with a recording and report you can check. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=tracy-vs-metaview&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=tracy-vs-metaview&utm_content=try-free) ## Sources - [Metaview, Interview notes](https://www.metaview.ai/interview-notes), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Connect](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. # Codility Pricing: What It Costs vs Tracy (Nov 2026) > Codility pricing as of 9 Nov 2026: packs from $4,800 a year or $800 a month. Cost at 100, 500 and 1,000 interviews against Tracy. - URL: https://tracyinterview.ai/compare/codility-pricing - Published: 2026-11-12 - Author: Tracy team > **In short** Tracy AI Interviewer charges $1 per credit, one credit per started 20 minutes, so a 45-minute technical interview is $3. The first 3 results are free, with no seats and no upfront cost. > > She follows up on approach and trade-offs inside the coding task and leaves a recording and a written assessment per question. > > At 100 interviews a year Tracy costs $300 against $4,800 on Codility's smallest annual pack. Codility is the better pick for timed coding assessments, its task library and proctoring. > > Next step: try three real interviews free and work out your own cost. Codility pricing is public: online packs start at $4,800 a year for 200 invites or $800 a month for 25, as listed on 9 Nov 2026. Enterprise is a custom quote with no figure. This page prices a technical screen at 100, 500 and 1,000 candidates and sets it beside Tracy AI Interviewer, which charges for each interview that produced a result. See [Tracy's pricing](/pricing) for her full list. Choose Codility when: - You want timed coding assessments and live technical interviews from a library of 1,300+ validated tasks (the full library from 1,000 invites a year). - Plagiarism detection and proctoring, included in every plan, matter more than the price per candidate. Choose Tracy when: - You want the first round to be a conversation that follows up on how the candidate reasoned, in their own field. - Your volume swings and you would rather pay for each finished interview than buy a pack of invites in advance. ## What does Codility pricing look like? Codility sells online packs of invites that you buy by card. An annual pack runs from 200 invites for $4,800 to 1,400 invites for $18,000, and Codility notes the same allowance would cost more on monthly plans. A monthly pack runs from 25 invites at $800 a month to 100 at $2,750. Every online plan covers Screen assessments and live Interviews from the same allowance, with 5 platform users and unlimited collaborators. Teams that need more than 1,400 invites a year, single sign-on or more integrations move to Enterprise, which is quoted for products, integrations and support. Source: Codility pricing page, as of 9 Nov 2026. | Pack | Price | If bought monthly instead | | --- | --- | --- | | 200 invites a year | $4,800 a year | Would be $6,400 | | 400 invites a year | $6,800 a year | Would be $12,000 | | 600 invites a year | $9,000 a year | Would be $18,000 | | 1,000 invites a year | $14,000 a year | Would be $27,500 | | 1,400 invites a year | $18,000 a year | Would be $38,500 | | 25 invites a month | $800 a month | Billed monthly | | 50 invites a month | $1,500 a month | Billed monthly | | 100 invites a month | $2,750 a month | Billed monthly | | Enterprise | Custom quote | No figure listed | ## What counts as an invite, and what costs extra? An invite counts when you send it. If the candidate has not started, you can cancel it and it returns to your allowance within the same billing period. When you reach your invites on an online plan, you move to the next pack, and assessments in progress carry on. Whether a test and a live interview for the same candidate use one invite or two is not stated on the page; ask. Some things sit outside the online packs: interview recording and transcripts, ID verification, video proctoring, SSO, every ATS integration beyond one standard connector and an interview environment with databases, caches and queues are listed as Enterprise features. Online plans include one standard integration, such as Greenhouse, Lever or Ashby. ## What does Codility cost at 100, 500 and 1,000 interviews? This is our arithmetic from the listed prices, not a Codility quote. It assumes one invite per candidate and picks the cheapest listed annual pack that covers the volume: the 200-invite annual pack for 100 candidates, the 600-invite pack for 500 and the 1,000-invite pack for 1,000. Monthly packs cost more for the same yearly volume at 500 and 1,000 interviews; at 100, one month of the 100-invite pack is $2,750 if you do not renew (monthly cancellation terms are not stated), and past 1,400 invites a year Codility moves you to Enterprise. Tracy charges one credit for each started 20 minutes, so a 45-minute interview is 3 credits. A Codility invite can start a test or a live interview and its pack also pays for the task library and proctoring, so the comparison is cost per candidate screened, not like for like. Codility figures from its pricing page, as of 9 Nov 2026. Tracy figures are $1 per credit before the first 3 free results. | Interviews in a year | Codility, cheapest listed annual pack | Tracy, up to 20 minutes | Tracy, 41 to 60 minutes | | --- | --- | --- | --- | | 100 | 200-invite annual pack, $4,800 (a 100-invite monthly pack is $2,750 for a single month) | $100 | $300 | | 500 | 600-invite annual pack, $9,000 | $500 | $1,500 | | 1,000 | 1,000-invite annual pack, $14,000 | $1,000 | $3,000 | ![Bar chart: Codility $4,800, $9,000, $14,000 against Tracy $300, $1,500, $3,000 at 100, 500, 1,000 interviews.](/blog/images/codility-pricing-1.png) *Tracy's usage cost for 45-minute interviews beside the cheapest listed Codility annual pack at each volume, so you can size a technical screening budget. Plan prices, not a quote. Source: Codility pricing page (as of 9 Nov 2026) and Tracy pricing page, Nov 2026.* At 100, 500 and 1,000 interviews a year, Tracy's usage cost is a fraction of Codility's cheapest listed annual pack. ## What does the interview itself look like? The price per interview is the first difference; the second is what the interview does. Tracy asks questions in the person's own field, follows up on what they said and waits for the answer. Inside a coding task she follows up on approach and trade-offs: which input to test next, and how the loop changes when it fails. Questions come from the job description or your own list, and you set the skills and their weights per role. Codility lists AI follow-up questions as included on every plan, and its pages describe a coding environment where candidates can work with an AI assistant, with every prompt saved for review. Both go beyond scoring a test. See [how Tracy works](/product). ![Sample interview: Tracy asks which input to test next, the candidate names edge cases, Tracy asks how to fix the loop.](/blog/images/codility-pricing-2.png) *A follow-up chain inside a coding task: Tracy asks about an edge case the candidate has just raised, so the reviewer sees reasoning, not only a result. Illustration with sample data* Tracy answers a coding result with a follow-up on the reasoning behind it. ## How do Codility and Tracy compare beyond price? Codility column from its pricing page, as of 9 Nov 2026. | Criterion | Tracy AI Interviewer | Codility | Who leads | | --- | --- | --- | --- | | Depth | Follow-ups in the person's own field, a coding task inside the interview, questions from the job description or your own. | 1,300+ validated tasks in 80+ languages and frameworks (full library from 1,000 invites a year), Screen and live Interview, AI follow-up questions. | Codility on coding environment and task library; Tracy on own-field questions. | | Volume | Starts on click at any hour, no scheduling; ready in under 10 seconds. | Packs of 200 to 1,400 invites a year or 25 to 100 a month; more than 1,400 goes to Enterprise. | Tracy: no allowance to outgrow. | | Stack | API and MCP (ChatGPT, Claude, Mistral, Gemini, Copilot Studio), Slack, WhatsApp and Zapier, and use beside Workable and Breezy HR. | One standard ATS integration on online plans; every integration including Workday and SAP on Enterprise. | Codility on listed ATS depth. | | Evidence | Recording of both sides, timestamped transcript, assessment per question with reasoning. A person decides. | Every prompt, command and code change saved for review; interview recording and transcripts on Enterprise. | Level on review trail; Tracy on recording and transcript for every result. | | Integrity | Signals for a human reviewer to interpret, with their evidence: tab changes, answer delays, fullscreen exits, external screens. Applicants are told on the join screen that the interview is recorded. No automated decision. ID check: Not stated. | Plagiarism detection, proctoring with snapshots and task leakage protection; ID verification and video proctoring on Enterprise. | Codility on plagiarism and proctoring. | | Security | SOC 2 Type II attested; report on request. No bias audit published. | SOC 2 Type II and ISO 27001 listed, GDPR processing, EU or US hosting on Enterprise. | Codility on listed standards. | | Price | $1 per credit; first 3 results free; no seats. | From $4,800 a year or $800 a month; Enterprise custom. | Tracy at the volumes above. | ## Is Codility worth its price, and for whom? If you run timed coding assessments from a large validated library, want plagiarism detection and proctoring included in every plan, and need a coding environment for live interviews, Codility is built for it, and its published packs make the budget easy to plan. If you want the first round to show how each applicant reasons, at a cost that follows the number of interviews, at the volumes above Tracy's interview costs less than Codility's cheapest listed annual pack, though Codility's price also buys assessments, a task library and proctoring: the first 3 results are free, there is no seat to buy, and the next 100 interviews of up to 20 minutes cost $100. Which path fits you: - Engineers and other technical roles: [technical interviews](/use-cases/technical-interviews). - Hundreds or thousands of applicants: [volume hiring](/use-cases/volume-hiring). - A first-round budget: [AI interview cost](/blog/ai-interview-cost). ## What should you ask Codility before you buy? 1. Does a live interview use the same invite as a test for the same candidate? 2. What is the first-year total at our volume and at twice our volume, and what happens at renewal if we outgrow the pack? 3. Which Enterprise features, such as recording and transcripts or ID verification, do we need, and what do they add? 4. Which ATS connector comes with our plan, and does it write results back? 5. What does a trial scoped to our roles include? Tracy answers the money questions on her pricing page. ## Common questions ### How much does Codility cost? As of 9 Nov 2026, online annual packs run from $4,800 for 200 invites to $18,000 for 1,400. Monthly packs run from $800 for 25 invites to $2,750 for 100. Enterprise is a custom quote. ### How much does Codility cost per candidate? Codility lists packs of invites, not a per-candidate rate. On the 200-invite annual pack, $4,800 spread over 200 candidates is $24 each, our arithmetic. Tracy charges $3 for a 45-minute interview. ### Does Codility have a free trial? Its pricing page has a Request a free trial option for Enterprise and says its team scopes trials to your roles, volume and products. Tracy's first 3 results are free and no card is needed. ### Is Tracy cheaper than Codility? At the listed prices, yes at 100, 500 and 1,000 interviews a year: $300, $1,500 and $3,000 for 45-minute interviews against $4,800, $9,000 and $14,000. Codility's price also buys assessments, a task library and proctoring. ### Does Tracy charge per seat or per month? No. There are no seats, no subscription and no upfront cost, and a session with no answers is never charged. --- Try three real interviews free. Send us one technical role and we build the interview plan, customised in days, with a recording and report you can check. Then $1 per credit. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=codility-pricing&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=codility-pricing&utm_content=try-free) ## Sources - [Codility, pricing](https://www.codility.com/pricing/), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. # iMocha Alternatives (Nov 2026) > iMocha alternatives for technical hiring: Tracy, HackerRank, CodeSignal, HireVue and Karat compared on depth, evidence, stack and price. - URL: https://tracyinterview.ai/compare/imocha-alternatives - Published: 2026-11-12 - Author: Tracy team > **In short** Among iMocha alternatives, Tracy AI Interviewer is the pick when a test score is not enough: she interviews every applicant by video, asks follow-ups in the person's own field, and leaves a recording, transcript and assessment per question, with a person deciding. > > Her price is published: $1 per credit, the first 3 results free, no seats. iMocha publishes none. > > iMocha is the better pick for breadth: a 10,000+ test library, proctoring logs, 50+ languages and named ATS connectors. > > Next step: create one interview for the workflow your current tool misses. Tracy AI Interviewer is the strongest iMocha alternative when the first round has to show how a person reasons, and iMocha is the better pick when you want a wide library of tests and proctoring around the interview. Choose Tracy when: - You want a conversation that follows up on what the person said, in their own field. - You want to read the price today: $1 per credit, the first 3 results free, no seats. - You want a recording, a timestamped transcript and an assessment per question behind every result. Choose iMocha when: - You want skill tests, coding problems, English assessment and proctoring logs from one vendor. - You need 50+ languages, WCAG 2.1 and a named ATS connector. ## Why do teams look for iMocha alternatives? Three differences show on iMocha's own pages. The price is a demo: its pricing address redirects to a page for booking one, so you cannot budget before a sales conversation. Tracy shows hers on the [pricing page](/pricing). The platform is wide: hiring is one use case beside skill-gap analysis, upskilling and internal mobility. And much of what it lists is assessment, with 10,000+ skill tests and 5,000+ coding problems, which screens against a bar but is not the same as hearing someone talk through a trade-off. Tara is described as an adaptive agent with structured scoring; the page does not say whether she asks a follow-up on what the candidate just said. ## Tracy AI Interviewer: follow-ups in the person's field Tracy interviews every applicant by two-way video. After each answer she asks about what was said, in the person's field, and waits for them to finish. Behind every result sit a recording, a timestamped transcript, a written assessment per question and integrity signals shown to a human reviewer. A person decides. A coding task can sit inside the interview (see [technical interviews](/use-cases/technical-interviews)), and she connects through API, MCP, Slack, WhatsApp and Zapier (see [Connect](/connect)). a Director at a top-tier early-stage VC said: "Tracy sometimes brings up exactly those questions I might have forgotten to ask, which has been a pleasant surprise." Where iMocha is stronger: a 10,000+ test library, 50+ languages, WCAG 2.1 and more attestations. Tracy publishes no language count, accessibility statement or bias audit. Send us one role and we build the interview plan, customised in days, so you can compare the follow-ups and the report with your current tool. ![Sample interview: Tracy asks about retried payments, the candidate answers, and Tracy asks a follow-up about disagreement.](/blog/images/imocha-alternatives-2.png) *Tracy follows up on the candidate's own answer, and each result keeps a recording and a timestamped transcript plus a written assessment per question, so a person can decide. Illustration with sample data* Tracy asks the second question on what the person actually said, and a recording and timestamped transcript sit behind every result. ## HackerRank: a live coding canvas on every plan Its AI interviewer, Chakra, is included on Starter, Pro and Enterprise, and it says Chakra "opens a live canvas like a dev environment and watches candidates actually do the work. Then it asks follow-up questions on the candidate's work". Starter is $79 a month ($948 billed annually) for 60 candidate attempts a year, Pro is $419 a month for 360, extra attempts are $20 each, and Enterprise is custom. Pick it when the work sample is code in a real environment. It is built around engineering, a narrower fit than Tracy for expert, applications or customer-research interviews. ## CodeSignal: assessments and AI interviews on one credit balance Build is $79 a month billed annually for 60 annual credits, Grow is $479 a month for 420, and Pro is custom. A credit is used when a candidate starts an assessment, an AI interview or a live interview, and extra credits are $20 each. Its AI interviewer runs a short structured interview, typically 15 minutes, and "if answers are thin, it probes and asks follow up questions based on your criteria"; follow-ups stay within the competencies you define, and humans decide. Pick it when you want assessments, AI interviews and live interviews drawn from one balance. ## HireVue: the enterprise suite with governance documents HireVue says its AI agents "ask and record candidate responses, including probing follow-up questions, in a natural two-way conversation," it lists more than 200 on-demand coding skills assessments, and it says "every score ties back to evidence from the interview and your competency rubric." It publishes adverse impact testing and an annual SOC 2 Type 2 audit. Like iMocha, it lists no price and points to a conversation. Pick it when governance documents are a procurement requirement. See [HireVue alternatives](/compare/hirevue-alternatives). ## Karat: trained human interviewers for engineering Karat runs technical interviews with its own Interview Engineers, who are people, not an AI agent. It offers 24/7 on-demand interviewing, including same-day, evenings and weekends, and each interview ends in a report with a recommendation, a score and written feedback per question. It has no pricing page; its calls to action are Request a Consult and Request demo. It describes evaluating engineering talent only. Pick it when you want a person to run a technical screen for you and accept a quote. ![Four cards: Tracy at $1 per credit; HackerRank and CodeSignal plans from $79 a month; iMocha and HireVue demo; Karat consult.](/blog/images/imocha-alternatives-1.png) *Tracy shows a price per credit you can read today; the others show plans, a demo or a consult. Illustration with sample data. Source: each vendor's pricing and product pages, Nov 2026.* Tracy's price is readable today; iMocha and HireVue ask you to book a demo. ## iMocha and its alternatives compared on four pillars and price Vendors' own pages and approved Tracy records, Nov 2026. Not stated means the page is silent. | Tool | Depth | Volume | Stack | Evidence | Price | | --- | --- | --- | --- | --- | --- | | Tracy AI Interviewer | Two-way video, follow-ups in the person's field | Starts on click, any hour | API, MCP, Slack, WhatsApp, Zapier | Recording, transcript, assessment per question | $1 per credit, first 3 free | | iMocha | Tara, adaptive AI interviewer; 10,000+ tests | On-demand AI interviews | Greenhouse, Workable, Lever, iCIMS, Workday | Scoring, transcripts, proctoring logs | Not public | | HackerRank | Coding canvas with follow-ups | Attempts per plan | 22+ enterprise ATS integrations | A why for every score | From $79 a month, annual | | CodeSignal | Short AI interview that probes thin answers | First pass for volume roles | Gem, Ashby; Workday, iCIMS on enterprise | Score, transcript with rationale | From $79 a month, annual | | HireVue | Probing follow-ups; 200+ coding assessments | Mid-market to enterprise first rounds | Scores shown in the ATS | Scores tied to evidence | None published | | Karat | Human Interview Engineers | 24/7 on demand | Not stated | Score, feedback, timestamped markers | No pricing page | ## What do the four pillars leave out? Other columns: vendors' own pages. Tracy column: approved Tracy product records. "Not stated" and "Not verified" mean the page does not say or we have not checked, not that it is absent. | Criterion | Tracy AI Interviewer | iMocha | Verified edge elsewhere | | --- | --- | --- | --- | | Accessibility | Not stated. | States WCAG 2.1 compliance. | iMocha. | | Multilingual | A spoken-language check scores fluency, coherence, grammar, pronunciation and vocabulary from 0 to 100. No language count. | 50+ global languages; CEFR-aligned English test. | HackerRank lists 55+ languages. | | Security and compliance | SOC 2 Type II attested, report on request at hello@tracyinterview.ai. | SOC 2 Type II, ISO 27001:2022, GDPR. | iMocha lists more; HireVue lists FedRAMP. | | Bias controls and oversight | A person reviews the evidence and decides. Tracy publishes no bias audit. | Describes a bias-free experience; no audit linked. | HireVue publishes adverse impact testing. | | Implementation, contract and support | Pay as you go, no seats. Enterprise adds a dedicated customer manager and support channel, customised in days. | Demo first, then a pilot or rollout plan. Support: not verified. | HackerRank and CodeSignal publish annual plans. | ## How do you choose an iMocha alternative? Decide four things. Tests or a conversation: if you screen against a bar, iMocha is built for it; if you need to see how people reason, pick a tool that asks the second question (see [AI technical interviews](/blog/ai-technical-interviews)). Who runs it: an AI agent scales to any hour, a person as at Karat suits teams that want a human on every screen. How you pay: a per-result price suits uneven volume, a plan by attempts or credits suits steady volume, and a quote cannot be compared until you know the unit and first-year total. What it connects to: name your ATS and ask each vendor for the connector. Which path fits you: - Engineers and other technical roles: [technical interviews](/use-cases/technical-interviews). - Hundreds or thousands of applicants: [volume hiring](/use-cases/volume-hiring). - Subject experts to vet: [expert vetting](/use-cases/expert-vetting). - A first-round budget: [Tracy's pricing](/pricing). ## Common questions ### What is the best iMocha alternative? It depends on the job. Tracy AI Interviewer suits technical and expert roles that need follow-ups and evidence at a published price. HackerRank and CodeSignal suit engineering teams that want an AI interviewer with plans from $79 a month, HireVue is the nearest enterprise suite, and Karat uses human interviewers. ### Who are iMocha's competitors? Among the tools we checked, HackerRank and CodeSignal compete for technical assessment and AI interviews, HireVue for enterprise first rounds and Karat for human-led technical screens. Tracy competes where the first round needs follow-ups. ### Is there an iMocha alternative with public pricing? Yes. Tracy costs $1 per credit with the first 3 results free. HackerRank and CodeSignal start at $79 a month billed annually. iMocha, HireVue and Karat publish no price. ### Which alternative works for technical interviews? Tracy can run a coding task inside the interview and ask follow-ups on approach and trade-offs. HackerRank opens a live coding canvas and asks follow-ups on the work. --- Create one interview for the workflow your current tool misses: send us one role and we build the interview plan, customised in days, with a recording and report you can check. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=imocha-alternatives&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=imocha-alternatives&utm_content=try-free) ## Sources - [iMocha, Skills Assessment](https://www.imocha.io/products/skills-assessment), retrieved 9 Nov 2026. - [Tracy AI Interviewer](https://tracyinterview.ai/), retrieved 9 Nov 2026. # iMocha Pricing: Tara AI Interviewer Costs (Nov 2026) > iMocha pricing is not public. Tracy's is: $1 per credit. See the cost at 100, 500 and 1,000 interviews and what to ask on an iMocha quote call. - URL: https://tracyinterview.ai/compare/imocha-pricing - Published: 2026-11-12 - Author: Tracy team > **In short** Tracy AI Interviewer has a price you can read and test today: $1 per credit, the first 3 results free, no seats. 100 interviews of up to 20 minutes cost $100, and each credit buys a reviewable interview with follow-ups, a recording and a report. > > iMocha pricing is not public: its pricing address redirects to a book-a-demo page, and Tara has no figure. > > iMocha fits when you want a 10,000+ test library, proctoring and named ATS connectors from one vendor. > > Next step: bring your volume and run three real interviews. iMocha pricing is not public, so no page can give you the cost of Tara today. Tracy AI Interviewer's is public, and this page sets it beside what iMocha's pages say about buying. Choose Tracy when: - You want the cost before you talk to anyone: $1 per credit, the first 3 results free, no seats. - The first round should be a conversation that asks follow-ups in the person's own field, with a recording and report you can check. - Your volume is uneven, so a per-result price fits better than a contract sized in advance. Choose iMocha when: - You want skill tests, coding problems, English assessment and proctoring from the same vendor as the interview. - You need a named connector for Workday, Greenhouse, Lever or iCIMS, or you run skills-based hiring, upskilling and internal mobility on one platform. Tracy's price is on her [pricing page](/pricing). ## What does iMocha pricing look like? Nothing on iMocha's site is a price. The pricing address redirects twice and lands on a customised product demo page, where the next step is a demo, pilot or rollout plan "only if it's a fit". The Tara page has one call to action, Book a Demo. iMocha describes itself as a skills and work intelligence platform, so a quote can span tests, interviews and proctoring, and it does not say how Tara is charged. Source: iMocha pricing, demo and Tara pages, Nov 2026. | What the page covers | What iMocha says | | --- | --- | | Plans and prices | None. The pricing address redirects to a book-a-demo page. | | How you buy | A customised product demo, then a demo, pilot or rollout plan. No term, unit or minimum is stated. | | Tara, the AI interviewer | Book a Demo. No price, plan or credit unit on the page. | ## What does Tracy cost at 100, 500 and 1,000 interviews? She charges one credit for each started 20 minutes of an interview that produced a result, and a credit costs $1. A 15-minute interview is $1 and a 34-minute interview is $2. Credits never expire, a session with no answers is never charged, and API and MCP access (see [Connect](/connect)) is part of Pay as you go. The first 3 results are free whatever their length; the totals ignore that. Each credit includes the follow-ups, the recording, the transcript and the assessment per question, so the price is per reviewable interview. Tracy at $1 per credit, one credit per started 20 minutes. iMocha: no public price, Nov 2026. | Interviews | Tracy, up to 20 minutes each | Tracy, 21 to 40 minutes each | iMocha | | --- | --- | --- | --- | | 100 | $100 | $200 | Not public | | 500 | $500 | $1,000 | Not public | | 1,000 | $1,000 | $2,000 | Not public | ![Three figures: $1 per credit per started 20 minutes, $0 for the first 3 results, and zero seats and upfront cost.](/blog/images/imocha-pricing-2.png) *Tracy costs $1 per credit, with the first 3 results free and no seats, so you can budget a first round before any call. Source: Tracy pricing page, Nov 2026. Source: Tracy pricing page, Nov 2026.* Tracy's cost is a number you can compute today, with no form to fill in before you see it. ## Why can a quote cost more per interview at low volume? Our reasoning, not an iMocha price: a platform quote usually comes down to one yearly number. Divide it by the interviews you complete. Each $1,000 of first-year quote equals 1,000 Tracy interviews of up to 20 minutes, so if you will complete fewer interviews of up to 20 minutes than your quote has dollars, pay as you go costs less per interview. A team that interviews 200 applicants a month, each up to 20 minutes, spends $2,400 a year. Bring your annual volume and interview length and we will work out your cost per completed interview. For a wider view, see [AI interview cost](/blog/ai-interview-cost) and [HireVue pricing](/compare/hirevue-pricing). ![Three figures: $1 for a 15-minute interview, $2 for a 34-minute interview, $100 for 100 interviews of 20 minutes.](/blog/images/imocha-pricing-1.png) *Tracy credits add up simply: $1 for a 15-minute interview, $2 for a 34-minute one, $100 for 100 interviews of 20 minutes. Worked examples from Tracy's pricing page. Source: Tracy pricing page, Nov 2026.* A credit is one started 20 minutes, so you can work out the cost of any interview length yourself. ## What should you ask on an iMocha quote call? 1. Which products are in the quote, and is Tara a separate line? 2. What is the unit: per candidate, per assessment, per interview, per user or a flat annual fee? 3. What is the first-year total at our volume, and at twice our volume? 4. Is implementation included, and is a pilot paid? 5. Are the ATS connectors (Greenhouse, Workable, Lever, iCIMS) in the price? 6. Can a reviewer get from a score to the sentence the candidate said, and is the recording kept? Bring the same list to every vendor you shortlist. Tracy answers the money questions on her pricing page, so the call can go to fit. ## How do iMocha and Tracy compare on the four pillars? Price is one line. Compare the two on Depth, Volume, Stack and Evidence, then on the criteria the pillars leave out. For how an interview becomes a report, see [how AI interviews work](/blog/how-do-ai-interviews-work). iMocha column: iMocha's own pages. Tracy column: approved Tracy product records. "Not stated" and "Not verified" mean the page does not say or we have not checked, not that it is absent. | Criterion | Tracy AI Interviewer | iMocha | Edge | | --- | --- | --- | --- | | Depth | Two-way video. She asks follow-ups in the person's own field and waits for the answer. A coding task can sit inside the interview. | Tara is an adaptive AI interview agent with structured scoring, around 10,000+ skill tests and 5,000+ coding problems. Follow-ups: not verified. | Tracy on follow-ups; iMocha on test breadth. | | Volume | Starts on click, at any hour, no scheduling. Ready in under 10 seconds. A dropped applicant can rejoin within 10 minutes (configurable). | On-demand AI interviews answered at the candidate's convenience. | Level on what is published. | | Stack | API and MCP connections, Slack, WhatsApp and Zapier, and use beside the Workable and Breezy HR connectors. | Lists Greenhouse, Workable, Lever, iCIMS, SmartRecruiters, Workday Recruiting and Oracle. | iMocha on named ATS connectors. | | Evidence | Recording of both sides, timestamped transcript, an assessment per question with reasoning, and integrity signals as evidence for a person, who decides. | Scoring, insights and full transcripts; proctoring with timestamped violation logs. | iMocha on proctoring logs; Tracy on the recording and report. | | Price and contract | $1 per credit, first 3 results free, no seats, no upfront cost. Enterprise is priced to how many interviews you run. | Not public. Demo first, then a pilot or rollout plan. | Tracy to start without a conversation. | | Accessibility and languages | Accessibility not stated. A spoken-language check scores fluency, coherence, grammar, pronunciation and vocabulary from 0 to 100. | States WCAG 2.1 compliance and 50+ global languages, plus a CEFR-aligned English test. | iMocha, on what is published. | | Security and oversight | SOC 2 Type II attested, report on request at hello@tracyinterview.ai. A person reviews the evidence and decides. Tracy publishes no bias audit. | Lists SOC 2 Type II, ISO 27001:2022 and GDPR, and describes a bias-free experience; no audit is linked. | iMocha lists more attestations. | | Implementation and support | Enterprise includes a dedicated customer manager, a dedicated support channel and a setup customised to your workflow. Tracy is customised in days. | Not verified. | Not verified for iMocha. | ## Is iMocha worth the quote, and for whom? For a large employer that wants one platform for tests, proctored assessments, interviews and skills data, with connectors for its HR suite, iMocha is built for the job and its quote may be worth chasing. It has a 10,000+ test library that Tracy does not match. For a team that wants the first round to show how each applicant reasons, Tracy is the one you can price before a call: the first 3 results are free and the next 100 cost $100. One customer cut its average time to hire from 44 days to 7. Enterprise is priced to how many interviews you run; book a demo for a figure. Which path fits you: - Engineers and other technical roles: [technical interviews](/use-cases/technical-interviews). - Hundreds or thousands of applicants: [volume hiring](/use-cases/volume-hiring). - Subject experts to vet: [expert vetting](/use-cases/expert-vetting). ## Common questions ### How much does iMocha cost? iMocha does not publish a price. Its pricing address redirects to a page for booking a customised demo. We do not estimate a figure. ### How much does Tara cost? Tara is iMocha's AI interviewer, and its page carries a Book a Demo button and no price. Ask whether it is a separate line in the quote and what unit it is charged in. ### Is iMocha cheaper than Tracy at low volume? iMocha publishes no price, so we cannot say. Each $1,000 of first-year quote equals 1,000 Tracy interviews of up to 20 minutes, so if you will complete fewer interviews of up to 20 minutes than your quote has dollars, pay as you go costs less per interview. ### Is there a free way to try Tracy? Yes. The first 3 results are free, whatever their length, and no card is needed. There are no seats and no subscription. --- Try three real interviews free and calculate your completed-interview cost: send us one role, your annual volume, interview length and stack, and we build the interview plan, customised in days, with a recording and report you can check. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=imocha-pricing&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=imocha-pricing&utm_content=try-free) ## Sources - [iMocha, Pricing (redirects to its demo page)](https://www.imocha.io/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. # Mercer Mettl Alternatives for Hiring Teams (Nov 2026) > Mercer Mettl alternatives for hiring teams: Tracy, HackerEarth, TestGorilla, iMocha and Talview, with published price and the case each fits. - URL: https://tracyinterview.ai/compare/mercer-mettl-alternatives - Published: 2026-11-12 - Author: Tracy team > **In short** Mercer Mettl alternatives depend on whether you need the test or the interview. Tracy AI Interviewer is first for the interview: follow-ups in the candidate's own field, a coding task in the same conversation, and a recording and written assessment per question. > > Her price is public: $1 per credit, the first 3 results free. Mercer Mettl, Talview and iMocha publish no price. > > Others by need: HackerEarth (live IDE, ID checks), TestGorilla (skills tests; AI interviews on Plus). > > Stay with Mercer Mettl for proctored tests at campus and enterprise scale. Mercer Mettl alternatives start with what you use Mercer Mettl for. Its pages describe an online assessment and proctoring platform: psychometric, aptitude and coding tests, AI and human proctoring, and a live coding interview tool. A team that bought it to run proctored tests at scale needs a different shortlist from one that wants to see how a candidate thinks. A test shows the answer. An interview shows the reasoning behind it. This page lists five options, [Tracy AI Interviewer](/product) first, with price where it is public and the case where each is the better pick. ## Why do teams look for Mercer Mettl alternatives? Three reasons, from Mercer Mettl's own pages. - **No public price.** Its pricing page offers personalised plans and a quote, with no amount listed. - **No AI interviewer stated.** The pages read describe tests, proctoring and a live coding interview tool, and do not state an AI interviewer that follows up on answers. - **Tests first.** Its pages lead with assessments and proctoring; interviews appear as a live coding tool. ## Which Mercer Mettl alternative fits which job? Best-for and published price for each option, from each vendor's own pages, read 9 Nov 2026. "Quote" means the vendor publishes no price. | Option | Best for | Published price | | --- | --- | --- | | Tracy AI Interviewer | Follow-ups in the candidate's own field, a coding task inside the interview, a recording and report a reviewer can check. | $1 per credit, one credit per started 20 minutes. First 3 results free. | | HackerEarth | AI technical interviews with code written in a live IDE, ID checks and native ATS connectors. | OnScreen from $129 a month for 10 invites; $499 for 50. | | TestGorilla | Skills tests plus AI interviews with follow-ups and your own scoring criteria. | Core from $215 a month (tests), Plus from $520, billed annually; AI interviews are on Plus, 2 credits per candidate. | | iMocha | A skills-test library with an AI interviewer called Tara and AI proctoring. | Quote. | | Talview | Identity verification, proctoring and an AI interviewer called Ivy, for high-stakes hiring. | Quote, typically from 150 to 200 candidates. | | Mercer Mettl | Proctored assessments and coding tests at campus and enterprise scale. | Quote. | ![Illustration: map of interview or test format against published price, with Tracy in the follow-ups, price shown quadrant.](/blog/images/mercer-mettl-alternatives-1.png) *Where the shortlist sits on format and published price, from each vendor's own pages, Nov 2026. HackerEarth and TestGorilla appear twice because they sell both interviews and tests. Source: vendors' public pages, Nov 2026.* Tracy, HackerEarth and TestGorilla publish a price for interviews with follow-ups; Mercer Mettl, Talview and iMocha ask for a quote. ## Tracy AI Interviewer: the interview behind the test Tracy is a two-way video interview run by an AI interviewer. You describe the role and she drafts the questions from your job description, or you bring your own. She asks follow-ups in the person's own field and waits for the answer, reads the CV when your CV check is on, and can run a coding task inside the same interview, with follow-ups on approach and trade-offs. A Director at a top-tier early-stage VC says "Tracy sometimes brings up exactly those questions I might have forgotten to ask, which has been a pleasant surprise." See [how AI interviews work](/blog/how-do-ai-interviews-work). After each interview you get a recording of both sides, a timestamped transcript, a written assessment for every question with its reasoning, and integrity signals with their evidence: tab changes, gaze, live face detection, delay before answers, fullscreen exits and external screens. A person decides. She has no seats and no upfront cost, and connects through API and MCP, Slack, WhatsApp and Zapier ([Connect](/connect)). **Where Tracy does not replace Mercer Mettl.** She does not run proctored exams or psychometric tests. Her integrity signals are evidence for a person to review, not a proctoring suite. ![Sample interview: Tracy asks a candidate what changes if lookup keys no longer fit in memory.](/blog/images/mercer-mettl-alternatives-2.png) *Tracy AI Interviewer follows up on a technical answer by changing the constraint, so a reviewer sees how the candidate reasons. Illustration with sample data* A follow-up asks the candidate to defend the answer a test would only mark. ## HackerEarth: AI technical interviews with ID checks HackerEarth's OnScreen runs adaptive AI interviews in which the candidate writes real code in a live IDE. It lists government-grade ID verification, behavioural monitoring, native Workday, Greenhouse, Lever and Ashby connectors, and SOC 2 Type II. Plans start at $129 a month for 10 invites. Pick HackerEarth when ID checks, proctoring and native ATS decide. ## TestGorilla: skills tests with AI interviews TestGorilla sells a library of 350+ tests on its Core plan. AI interviews, with dynamic follow-ups and scoring criteria you define, are a Plus feature at 2 credits per candidate; Core is from $215 a month and Plus from $520, billed annually; ATS and API integrations are Plus features. Pick it when you want broad skills tests and an AI interview from one vendor at a listed price. ## iMocha and Talview: skills tests and proctoring, priced by quote iMocha pairs a library of 10,000+ skill tests with an AI interviewer called Tara and AI proctoring with timestamped violation logs. Its pricing page leads to a demo, with no price. Talview sells identity verification, proctoring and an AI interviewer called Ivy to high-stakes hirers, and asks for a quote, typically for 150 to 200 candidates or more. Pick either when proctoring and verification matter as much as the interview. ## Where is Mercer Mettl the better pick? Mercer Mettl is the better pick for proctored assessments at campus and enterprise scale. It says it is trusted by 6,000+ organizations and 700+ educational institutions in over 100 countries, and its pages list psychometric, aptitude and communication tests, AI and human proctoring, a technical question repository of 100,000+ questions for 300+ skills, and live coding interviews with pair programming and a whiteboard. If your first round is a proctored test, keep it, and compare only the interview step. ## How do Tracy and Mercer Mettl compare criterion by criterion? Mercer Mettl column: its own pages, read 9 Nov 2026. "Not stated" means the pages read do not say. | Criterion | Tracy AI Interviewer | Mercer Mettl | Who leads | | --- | --- | --- | --- | | Depth | Follow-ups in the person's own field; a coding task inside the interview. | Auto-graded coding assessments and a live coding interview tool; an AI interviewer is not stated. | Tracy on follow-ups; Mercer Mettl on test breadth | | Volume | Each applicant starts on a click, at any hour, with no scheduling. | Schedule and invite multiple candidates; real-time group reports. | Level | | Stack | API and MCP, Slack, WhatsApp and Zapier, beside Workable and Breezy HR. | Prebuilt ATS and LMS integrations on its partners page (Greenhouse, Lever, SmartRecruiters). | Level; Tracy adds API and MCP, Slack, WhatsApp. | | Evidence | A recording, a timestamped transcript and a written assessment per question; integrity signals with their evidence. | Section and question-level reports; AI and human proctoring. | Mercer Mettl on proctoring; Tracy on interview evidence | | Security | SOC 2 Type II attested; report on request. | A "SOC2 Type 2" tile and ISO 27001 on its pricing page. | Level | | Price and terms | $1 per credit; first 3 results free; no seats. | Quote. | Tracy | ## How to choose among Mercer Mettl alternatives 1. **Start from what the test does for you.** If it proctors a campus drive, keep that and add an interview after it. If it screens technical skills, compare the interview. 2. **Decide whether you need follow-ups.** For roles where a correct answer can hide shallow reasoning, pick a two-way interview: Tracy, HackerEarth, TestGorilla, iMocha or Talview. 3. **Settle price before features.** Tracy, HackerEarth and TestGorilla publish a price. For Mercer Mettl, iMocha and Talview, ask for the unit, the minimum and the first-year total at your volume. 4. **Check the evidence.** Ask each vendor to open one completed result and show how a score ties to what the candidate said. 5. **Run one role.** Pick a role with at least 20 applicants and run two options on it. To start with Tracy, send us the role and we build the plan. For technical roles, see [technical interviews](/use-cases/technical-interviews). [The best AI interview software](/blog/best-ai-interview-software) compares the wider market, and [AI interview cheating detection](/blog/ai-interview-cheating-detection) covers what to ask any vendor about integrity signals. ## Common questions ### What are the best Mercer Mettl alternatives? For follow-up interviews at a published price, Tracy AI Interviewer. For AI technical interviews with ID checks, HackerEarth. For skills tests with AI interviews, TestGorilla. For proctoring and verification, iMocha or Talview. ### What is Mercer Mettl? An online talent assessment and proctoring platform. Its pages list psychometric, aptitude, communication and coding tests, AI and human proctoring, and a live coding interview tool. ### Does Tracy replace Mercer Mettl's proctoring? No. Tracy shows integrity signals, with their evidence, to a person who reviews them. If you need proctored exams, keep that layer and compare only the interview. ### How do I try Tracy before paying? The first 3 results are free, no card. Create one interview for a technical role and compare the follow-ups with your test results. --- Create one interview for the workflow your current tool misses. Send us the role and we will build the plan, customised in days, with a recording and report you can check. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=mercer-mettl-alternatives&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=mercer-mettl-alternatives&utm_content=try-free) ## Sources - [Mercer Mettl: Pricing](https://mettl.com/en/pricing/), retrieved 9 Nov 2026. - [Tracy AI Interviewer: Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. # Mercer Mettl Pricing: Cost vs Tracy (Nov 2026) > Mercer Mettl pricing is a quote with no public price. What to ask, how to estimate, and Tracy's $1 per credit cost at 100, 500 and 1,000 interviews. - URL: https://tracyinterview.ai/compare/mercer-mettl-pricing - Published: 2026-11-12 - Author: Tracy team > **In short** Mercer Mettl publishes no price. Its pricing page offers personalised plans and a quote. Tracy AI Interviewer publishes hers: $1 per credit, one credit per started 20 minutes, the first 3 results free, no seats, no upfront cost. > > A 45-minute technical interview is $3 on Tracy, with follow-ups and a written assessment per question. > > Mercer Mettl is the better pick for proctored tests at campus and enterprise scale; ask it for a first-year total at your volume. > > Next step: try three real interviews free and set the figure beside your quote. Mercer Mettl pricing is not public. Its pricing page asks for a quote, with no amount, plan tier or unit listed. That makes the first job of a pricing page practical: say what is known, how to estimate, and what to ask on the call. [Tracy AI Interviewer](/product) is the contrast. She charges for each interview that produced a result, the price is on her [pricing page](/pricing), and you can work out a budget before you speak to anyone. ## What does Mercer Mettl pricing look like? Mercer Mettl's pages describe an online assessment and proctoring platform and sell by quote. The table lists every price signal on the pages read. Source: Mercer Mettl pages, read 9 Nov 2026. Quoted phrases are word for word. No amount, plan name or unit appears on any of them. | Page | What it says about price | Read on | | --- | --- | --- | | Pricing page | "Personalized Pricing Plans" and "Customised plan as per your organisation". A Get quote now button with a product selector, and a demo request. | 9 Nov 2026 | | Tech screening page | Offers "Try for free", a demo and a call with an expert. No price. | 9 Nov 2026 | | Home page | Lists Pricing in its main menu. No amount. | 9 Nov 2026 | ## How can you estimate what Mercer Mettl costs? What follows is reasoning, not a Mercer Mettl figure. Its pages do not say how it bills, so the unit could be the candidate, the test, the proctored session or the user. Whatever the unit, a quote built for large assessment programmes is likely to carry a minimum, and a low volume of interviews is then likely to cost more per interview than a high one. Treat any figure you are given as a starting point and ask for the first-year total at 100, 500 and 1,000 candidates, so you can compare it with a per-result price. Also ask whether the quote covers proctoring, the live coding interview tool and the question library, because its pages list them as separate products. ## What should you ask on a Mercer Mettl quote call? 1. What is the unit we pay for: candidate, test, proctored session or user? 2. Is there a minimum commitment or a minimum number of candidates? 3. Which of the pieces we need are priced separately: proctoring, the live coding interview tool, the question library, reports? 4. Does ATS integration cost extra? Its partners page lists prebuilt integrations (Greenhouse, Lever, SmartRecruiters); ask which are in the quote. 5. What does the free trial on the tech screening page include? 6. What is the first-year total at 100, 500 and 1,000 candidates, and what happens at renewal? 7. What support is included, and who answers a candidate's call on test day? ## What does Tracy cost at 100, 500 and 1,000 interviews? Tracy charges one credit, $1, for each started 20 minutes of an interview that produced a result. The first 3 results are free, credits never expire, a session with no answers is never charged, and there are no seats. Worked examples: 15 minutes is 1 credit, 34 minutes is 2 credits, and 100 interviews of 20 minutes cost $100. Tracy figures are $1 per credit before the first 3 free results. Mercer Mettl: quote, no price published. Bring your yearly volume and interview length and we will work out your own cost on a demo call. | Interviews in a year | Mercer Mettl | Tracy, up to 20 minutes | Tracy, 41 to 60 minutes | | --- | --- | --- | --- | | 100 | Quote | $100 | $300 | | 500 | Quote | $500 | $1,500 | | 1,000 | Quote | $1,000 | $3,000 | ![Bar chart of Tracy cost: $100, $500, $1,000 for 20-minute interviews and $300, $1,500, $3,000 for 45 minutes.](/blog/images/mercer-mettl-pricing-1.png) *What Tracy costs at 100, 500 and 1,000 interviews of up to 20 and of 45 minutes. Mercer Mettl publishes no price, so it appears as an outlined label, not a bar. Plan prices, not a quote. Source: Tracy pricing page, Nov 2026.* Tracy's cost is a straight line you can compute before a call: $1 per 20 minutes, 100 to 1,000 interviews. ## What is inside each Tracy credit? A credit buys a finished interview, not a seat. Tracy drafts questions from your job description or uses your own, asks follow-ups in the person's own field, and can run a coding task inside the same interview. Each result comes with a recording of both sides, a timestamped transcript and a written assessment per question with its reasoning. A person decides. Each applicant starts on a click, at any hour, with no scheduling. See [hiring at volume](/use-cases/volume-hiring) and [AI interview reports](/blog/ai-interview-reports). ![Sample interview: Tracy asks how a token bucket behaves across two servers after a candidate's first answer.](/blog/images/mercer-mettl-pricing-2.png) *Each Tracy result includes the follow-ups Tracy asked and the answers, with times, in a timestamped transcript. Illustration with sample data* The follow-up is part of the credit: Tracy asks about the constraint the first answer skipped. ## How do Tracy and Mercer Mettl compare beyond price? Mercer Mettl column: its own pages, read 9 Nov 2026. "Not stated" means the pages read do not say. | Criterion | Tracy AI Interviewer | Mercer Mettl | Who leads | | --- | --- | --- | --- | | Depth | Follow-ups in the person's own field; a coding task inside the interview. | Coding assessments and a live coding interview tool; an AI interviewer is not stated. | Tracy on follow-ups | | Volume | Starts on click at any hour; a dropped applicant can rejoin within 10 minutes. | Schedule and invite multiple candidates; trusted by 6,000+ organizations and 700+ educational institutions. | Mercer Mettl at campus scale | | Stack | API and MCP, Slack, WhatsApp and Zapier, beside Workable and Breezy HR. | Prebuilt ATS and LMS integrations listed on its partners page. | Level | | Evidence | A recording, a timestamped transcript and a written assessment per question; integrity signals with their evidence. | Section and question-level reports; AI and human proctoring. | Mercer Mettl on proctoring; Tracy on interview evidence | | Security | SOC 2 Type II attested; report on request. | A "SOC2 Type 2" tile and ISO 27001 on its pricing page. | Level | | Support | Enterprise adds a dedicated customer manager and a support channel. | Phone, email and a help portal. | Level | ## Is Mercer Mettl worth a quote, and for whom? If your first round is a proctored test, if you run campus drives, or if you need psychometric and aptitude tests with AI and human proctoring, Mercer Mettl is built for it and a quote is worth getting. If you want the first round to show how each applicant reasons, with a cost that follows the number of interviews, Tracy's published price lets you test one role before you commit: the first 3 results are free, and the next 100 interviews of up to 20 minutes cost $100. Enterprise is priced to how many interviews you run. Which path fits you: - Engineers and other technical roles: [technical interviews](/use-cases/technical-interviews). - Hundreds or thousands of applicants: [volume hiring](/use-cases/volume-hiring). - A first-round budget: [AI interview cost](/blog/ai-interview-cost). ## Common questions ### How much does Mercer Mettl cost? Mercer Mettl publishes no price. Its pricing page offers personalised plans and a quote. Tracy costs $1 per credit, with the first 3 results free and no seats. ### Does Mercer Mettl have a free trial? Its pricing page does not mention one. Its tech screening page (mettl.com/tech-screening) offers "Try for free", a demo and a call with an expert; ask what the trial includes. Tracy's first 3 results are free, whatever their length, and no card is needed. ### What is Mercer Mettl? An online talent assessment and proctoring platform with psychometric, aptitude and coding tests, AI and human proctoring and a live coding interview tool. ### Is Tracy cheaper than Mercer Mettl? Mercer Mettl publishes no price, so a comparison in dollars is not possible. Tracy's price is $1 per credit, so 100 interviews of up to 20 minutes cost $100. Ask Mercer Mettl for the first-year total at the same volume. ### What does Tracy cost for 45-minute interviews? A 45-minute interview takes 3 credits, so $3. At 500 interviews that is $1,500. --- Try three real interviews and calculate your cost: bring your yearly volume, interview length and stack, and we build the plan, customised in days, with a recording and report you can check. The first 3 results are free, then $1 per credit. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=mercer-mettl-pricing&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=mercer-mettl-pricing&utm_content=try-free) ## Sources - [Mercer Mettl: Pricing](https://mettl.com/en/pricing/), retrieved 9 Nov 2026. - [Tracy AI Interviewer: Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. # Spark Hire Pricing: Cost vs Tracy (Nov 2026) > Spark Hire pricing: video interviews from $249 a month billed annually. Cost at 100, 500 and 1,000 interviews against Tracy's $1 per credit. - URL: https://tracyinterview.ai/compare/spark-hire-pricing - Published: 2026-11-12 - Author: Tracy team > **In short** Tracy AI Interviewer prices by result: $1 per credit, one credit per started 20 minutes, so a 45-minute interview is $3. The first 3 results are free, with no seats and no upfront cost. > > Each result includes own-field follow-ups, a recording and a written assessment per question. > > Spark Hire's annual plan starts at $249 a month, $2,988 a year, unlimited interviews: it costs less past about 1,000 interviews of 45 minutes. > > Next step: try three real interviews free, then calculate your own cost. Spark Hire pricing is public: the video interview plan starts at $249 a month when billed annually, with unlimited interviews. That is a flat subscription for recorded one-way and live video. [Tracy AI Interviewer](/product) charges for each interview that produced a result. This page sets out Spark Hire's plans, what costs extra, and what 100, 500 and 1,000 interviews cost on each, so you can see where a flat plan beats pay per result and where it does not. See [Tracy's pricing](/pricing) for the full list. ## What does Spark Hire pricing look like? Spark Hire sells video interviews, a behavioral assessment and an applicant tracking system, each priced separately. Both video interview plans carry the note "Pricing varies by company size", so the figures below are starting prices. Annual billing adds live training and a Customer Success Manager, which are annual only. Recruit, the ATS, is sold only on an annual contract. Recruit prices also scale with headcount. Source: Spark Hire pricing page, read 9 Nov 2026. Yearly cost is our arithmetic from the monthly figure. Prices are in USD plus applicable sales tax. | Plan | Starting price | Yearly cost | What the page lists | | --- | --- | --- | --- | | Video interviews, monthly | $299 a month | $3,588 | Unlimited live and one-way interviews, up to 5 jobs at a time, up to 5 users | | Video interviews, annual | $249 a month, billed annually | $2,988 | Unlimited interviews, jobs and users; 40+ ATS integrations | | Recruit Pro (ATS) | $335 a month, billed annually | $4,020 | Applicant tracking with one-way video, AI summaries and transcripts | | Recruit Growth (ATS) | $499 a month, billed annually | $5,988 | Recruit Pro plus customizable scorecards and panel interview scheduling | | Recruit Enterprise | Talk to Sales | Quote | Custom terms; quote via sales | ## What costs extra on Spark Hire? API access, single sign-on and additional brands are paid add-ons. Spark Hire says several integrations need a paid add-on too, including Dayforce, UKG Pro and Workday. Its FAQ says that because service starts as soon as you pay, "we do not offer refunds". A free trial is not stated on the pricing page, which offers a demo. SOC 2 Type II is available on request. ## What does Spark Hire cost at 100, 500 and 1,000 interviews? This is our arithmetic from the published prices, not a Spark Hire quote. Spark Hire's annual video plan is a flat $2,988 a year at its starting price, whatever the number of interviews, so its cost per interview falls as volume rises: about $29.88 at 100, $5.98 at 500 and $2.99 at 1,000. The two products also differ in what an interview is, a recorded answer to fixed questions on one side and a two-way conversation with follow-ups on the other. Tracy charges one credit for each started 20 minutes, so 45 minutes is 3 credits. Spark Hire figure: annual video plan at its starting price. Tracy figures are $1 per credit before the first 3 free results. Bring your yearly volume and interview length and we will work out your own cost on a demo call. | Interviews in a year | Spark Hire annual video plan, from | Tracy, up to 20 minutes | Tracy, 41 to 60 minutes | | --- | --- | --- | --- | | 100 | $2,988 | $100 | $300 | | 500 | $2,988 | $500 | $1,500 | | 1,000 | $2,988 | $1,000 | $3,000 | ![Bar chart: Spark Hire $2,988 a year against Tracy $300, $1,500 and $3,000 at 100, 500 and 1,000 interviews.](/blog/images/spark-hire-pricing-1.png) *Tracy's cost at 100, 500 and 1,000 interviews of 45 minutes, beside Spark Hire's annual starting price, so you can see where the flat plan overtakes pay per result. Plan prices, not a quote. Source: Spark Hire pricing page and Tracy pricing page, Nov 2026.* At 100 and 500 interviews a year Tracy costs less than Spark Hire's starting price, and the two meet near 1,000 interviews of 45 minutes. ## Where does Spark Hire's flat price win? For short interviews the flat plan wins late: Tracy at 20 minutes reaches $2,988 only at about 3,000 interviews a year, and at 45 minutes at about 1,000. Beyond that, and for a team that wants one tool for live and recorded video, scheduling and its own ATS, Spark Hire's flat price is the better deal. It is the better pick, too, when recorded answers to fixed questions are all the first round needs. ## How does Tracy's pricing suit volume? Tracy has no seats, no plans to climb and no upfront cost. A credit is $1 and covers each started 20 minutes of an interview that produced a result. Credits never expire, and a session with no answers is never charged. Each applicant starts on a click, at any hour, with no scheduling. One customer cut its average time to hire from 44 days to 7. See [hiring at volume](/use-cases/volume-hiring). ![Sample interview: Tracy asks what an index does to writes on a busy table after a candidate's first answer.](/blog/images/spark-hire-pricing-2.png) *Each result includes the follow-ups Tracy asked and the answers, with times, beside the recording, a transcript and a written assessment per question. Illustration with sample data* One credit buys the whole interview: the follow-up, the evidence and a written assessment, not a bare recording. ## How do Spark Hire and Tracy compare beyond price? Spark Hire column: its own pages, read 9 Nov 2026. "Not stated" means the pages read do not say. | Criterion | Tracy AI Interviewer | Spark Hire | Who leads | | --- | --- | --- | --- | | Depth | Follow-ups in the person's own field; questions from the job description or your own; a coding task inside the interview. | One-way recorded video: every candidate answers the same questions. Follow-ups not stated. | Tracy | | Volume | Starts on click at any hour; a dropped applicant can rejoin within 10 minutes. | Unlimited interviews, jobs and users on the annual plan. | Level; Spark Hire on flat cost at high volume | | Stack | API and MCP, Slack, WhatsApp and Zapier, beside Workable and Breezy HR connectors. | 40+ ATS integrations and its own ATS; some need a paid add-on. | Spark Hire on ATS depth | | Evidence | A recording of both sides, a timestamped transcript, a written assessment per question and integrity signals with their evidence. | Recorded video with AI summaries, transcription in 50+ languages and scoring on key factors. | Tracy on written assessment per question | | Security | SOC 2 Type II attested; report on request. | SOC 2 Type II audit on request. | Level | | Terms | First 3 results free, no card; credits never expire. | No refunds; free trial not stated; best price on annual billing. | Tracy on flexibility | ## What should you ask Spark Hire before you buy? 1. What is the price at our company size, since the page says pricing varies? 2. What is the first-year total with the add-ons we need, such as SSO, API access or a Workday integration? 3. What happens to the price when our headcount band changes? 4. Which plan do we need for live training and a Customer Success Manager? Tracy's [pricing](/pricing) page lists every price in full, so a demo call can go straight to fit. For a wider cost picture, see [what an AI interview costs](/blog/ai-interview-cost), or [VidCruiter pricing](/compare/vidcruiter-pricing) for a suite quoted by module. ## Common questions ### How much does Spark Hire cost? Its video interview plan starts at $299 a month billed monthly, or $249 a month billed annually. The page adds that pricing varies by company size. Recruit, its ATS, starts at $335 a month billed annually. ### Does Spark Hire have a free trial? A free trial is not stated on its pricing page, which offers a demo. Tracy's first 3 results are free, whatever their length, and no card is needed. ### How does Spark Hire work? Candidates record answers to your questions on video, on their own schedule, and your team watches and rates them. Spark Hire also runs live video interviews. Tracy is a two-way interview run by an AI interviewer who asks follow-ups. ### Is Tracy cheaper than Spark Hire? At 100 and 500 interviews a year, yes: $300 and $1,500 for 45-minute interviews against Spark Hire's starting price of $2,988. At about 1,000 interviews of 45 minutes the two meet, and beyond that Spark Hire's flat price is lower. --- Try three real interviews and calculate your cost: bring your yearly volume, interview length and stack, and we build the plan, customised in days, with a recording and report you can check. The first 3 results are free, then $1 per credit. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=spark-hire-pricing&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=spark-hire-pricing&utm_content=try-free) ## Sources - [Spark Hire: Pricing](https://www.sparkhire.com/pricing/?solution=video-interviews), retrieved 9 Nov 2026. - [Tracy AI Interviewer: Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. # Spark Hire Reviews: What Buyers Say (Nov 2026) > Spark Hire reviews counted, not picked: 50 read, with quotes, dates and themes. What buyers praise, what they complain about, and who it suits. - URL: https://tracyinterview.ai/compare/spark-hire-reviews - Published: 2026-11-12 - Author: Tracy team > **In short** Tracy AI Interviewer suits a first round that needs more than a recorded answer: she asks follow-ups in the applicant's own field, with a recording, a timestamped transcript and a written assessment per question behind each result. The first 3 results are free. > > Spark Hire's reviews are warm: 41 of the 50 we read on Software Finder give five stars, and 29 praise ease of use. > > None of the 50 mentions follow-up questions, and only 4 date from 2023 or later. > > Pick Spark Hire if recorded answers to fixed questions are enough. Spark Hire reviews are positive and mostly about convenience. Of the 50 Spark Hire reviews we read, 29 praise ease of use, 21 say it saves time, and one gives a single star. What they cannot say is how much a recorded answer to a fixed question tells you about the person. Spark Hire's own page says every candidate answers the same questions under the same conditions. For first-round roles where a polished answer can hide a shallow one, [Tracy AI Interviewer](/product) asks follow-ups in the applicant's field and waits for the answer. **Choose Spark Hire when:** - you want recorded answers to fixed questions that your own team watches, at a published price; - you want an applicant tracking system from the same vendor, or one of the 40+ ATS integrations it lists; - you want unlimited users on the annual plan. **Choose Tracy when:** - the first round needs follow-ups that test what the applicant just said; - a reviewer must check each result against a recording and a timestamped transcript; - hundreds of applicants should start on a click, at any hour, with no scheduling. ## Where do the Spark Hire reviews come from? This page counts the 50 Spark Hire reviews on Software Finder, dated March 2019 to July 2025 and marked by the site as from verified customers (4.7 overall). A review can raise more than one theme, so read each count as a direction. Twenty come from employers with 500 or more staff, 17 from 101 to 500, and 13 from companies under 100. Forty-two are from 2021 or earlier, so they describe the product as it was then. Reviews on G2 and Capterra are not counted. ## What do buyers praise about Spark Hire? Ease comes first. Twenty-nine of the 50 reviews call it easy to use, set up or navigate. A telecoms reviewer with 1 to 10 staff wrote that "the platform was incredibly easy to navigate" (Software Finder, July 2025). Twenty-one say it saves time. A food and beverage employer with 101 to 500 staff said the scheduling feature "has shaved days off our interview process" (Software Finder, June 2021). Some praise what the format does for fairness and sharing. A non-profit employer with 500 or more staff said the tool "puts all candidates on an equal footing for the initial interview stage" (Software Finder, October 2022). Five describe replaying or sharing recordings with colleagues, four praise support, and four say the ATS connection was easy. ## What do buyers complain about? Twenty reviews name no drawback at all. The rest of the complaints are scattered, and none is repeated by more than seven reviewers. Counts are out of the 50 reviews, with one quote each. Counts: our reading of the 50 Software Finder reviews. A review can raise more than one. Quotes are word for word. One review on ATS depth and one on a single language are left out of the table. | Complaint | Reviews (of 50) | One quote | Reviewer and date | | --- | --- | --- | --- | | Interface or setup friction | 7 | "Initial setup pretty time-consuming for me" | 500+ employer, Feb 2021 | | A feature they wanted | 7 | "I'd like more editing flexibility so I could trim videos" | 101 to 500 employer, Apr 2021 | | Candidates cannot get in, or the link goes astray | 5 | "Occasionally candidates are unable to log in" | 101 to 500 employer, Jun 2022 | | Price, plan limits or billing | 4 | "forced to move up to a higher pricing tier" | 1 to 10 staffing employer, Dec 2020 | | Candidates unfamiliar with, or reluctant about, video | 2 | "somewhat robotic and a little unfamiliar" | 101 to 500 employer, Mar 2022 | | Limited product, poor support | 1 | "the product is extremely limited" | 11 to 50 employer, Sep 2023 (1 star) | ![Bar chart of 50 Spark Hire reviews: ease of use 29, time saved 21, setup 7, missing feature 7, access faults 5.](/blog/images/spark-hire-reviews-2.png) *Reviews that raise each point, of 50 read: praise in blue, complaints in grey. A review can raise more than one. Software Finder, 50 Spark Hire reviews, March 2019 to July 2025.* Praise for ease and time saved outnumbers any single complaint, and no complaint passes 7 of 50. ## What do the reviews not say? None of the 50 mentions follow-up questions, AI scoring or bias. One review mentions scoring, and it means panel scores from the team. One, from November 2021, asks for a transcription feature so a recruiter can read answers instead of rewatching videos. Spark Hire's own pages now list transcription in 50+ languages, AI summaries and scoring against key factors, so that request may have been met. The reviews show the tool is easy and that teams like it. They say little about interview depth, because a one-way format has no moment where the interviewer reacts. Spark Hire says so itself: every candidate answers the same questions under the same conditions. That suits a screen for a standard role. It is thinner for a role where the second question decides. ## How does Tracy answer the complaints? The table sets four complaints against what Tracy does, and what she does not fix. Tracy column: Tracy's published capabilities. Counts: 50 Software Finder reviews. | Complaint in the reviews | What Tracy does | What Tracy does not fix | | --- | --- | --- | | Candidates cannot get in (5) | Each applicant starts on a click, at any hour. A dropped applicant can rejoin within 10 minutes, a reminder goes out after a day, and failed interviews are chased automatically. | An invitation email can still land in a junk folder, as one 2021 reviewer found. | | Price or plan limits (4) | $1 per credit, no seats, no upfront cost, and no plan to climb. | Spark Hire's unlimited annual plan can cost less at high volume. | | Candidates unfamiliar with, or reluctant about, video (2) | A two-way conversation: she asks follow-ups on what the person said. | She is an AI interviewer, and the join screen says the session is recorded and transcribed. | | A feature they wanted (7) | A recording and a timestamped transcript behind every result, with a written assessment per question. One reviewer asked for transcripts. | Tracy is not an applicant tracking system. Candidates stay in your ATS. | ![Sample interview: Tracy asks how a candidate measured a cache speed-up after a vague answer.](/blog/images/spark-hire-reviews-1.png) *Tracy AI Interviewer follows up on a vague answer in the applicant's own field, so a reviewer reads the reasoning behind it. Illustration with sample data* Tracy follows up on a vague answer, so a reviewer reads the reasoning and not only the claim. ## Who is Spark Hire right for, and who is it not? Spark Hire suits teams that want to replace phone screens with recorded video, keep people deciding, and pay a published price: its annual video interview plan starts at $249 a month, with unlimited interviews, jobs and users. The vendor also sells its own ATS, from $335 a month. Reviewers from small teams and 500-person employers both rate it highly. It suits you less if you need follow-up questions, a coding task in the interview, or integrity signals for a person to review. The Spark Hire pages read on 9 Nov 2026 do not state those. For the full cost picture, read [what an AI interview costs](/blog/ai-interview-cost). For a suite with one-way and live interviews, see [Tracy vs VidCruiter](/compare/tracy-vs-vidcruiter). ## What should you take from the Spark Hire reviews? A Director at a top-tier early-stage VC says of Tracy: "Tracy sometimes brings up exactly those questions I might have forgotten to ask, which has been a pleasant surprise." Hold any one-way tool to one test: after the recording, what does the reviewer know that they could not have guessed from the question list? Her price is on the [pricing page](/pricing). To see the difference on your own role, [try Tracy free](/connect), or read how [AI interviews work](/blog/how-do-ai-interviews-work) first. ## Common questions ### Is Spark Hire good, according to reviews? Reviewers rate it highly: 41 of the 50 we read give five stars. They praise ease of use (29) and time saved (21). Only 4 reviews are from 2023 or later. ### What do Spark Hire users complain about? No complaint passes 7 of 50. The most common are interface or setup friction (7), a missing feature (7), and candidates who cannot log in or miss the link (5). Twenty reviews name no drawback. ### What do candidates say about Spark Hire interviews? Almost all reviewers are employers. One employer says the one-way format felt "somewhat robotic and a little unfamiliar" at first, and another says some candidates resist recording a video. One reviewer used the app as a candidate and found the interface clean and the process pleasant. ### Is there a Spark Hire alternative with follow-up questions? Tracy AI Interviewer asks follow-ups in the applicant's own field and keeps a recording and a transcript behind each result. Spark Hire's pages state a one-way format with fixed questions. --- Compare the candidate and reviewer experience: send us one role and we build the interview plan in days, then you check the follow-ups and the report against your Spark Hire recordings. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=spark-hire-reviews&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=spark-hire-reviews&utm_content=try-free) ## Sources - [Software Finder: Spark Hire reviews](https://www.softwarefinder.com/hr/spark-hire/reviews), retrieved 9 Nov 2026. - [Tracy AI Interviewer: Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Spark Hire: Pricing](https://www.sparkhire.com/pricing/?solution=video-interviews), retrieved 9 Nov 2026. # Talview Alternatives (Nov 2026) > Talview alternatives and competitors for AI interviews: Tracy, HireVue, VidCruiter, Sapia.ai, Willo, Hireflix and Interviewer.AI, with price and fit. - URL: https://tracyinterview.ai/compare/talview-alternatives - Published: 2026-11-12 - Author: Tracy team > **In short** Talview publishes no price. Tracy AI Interviewer is first on the shortlist: follow-ups in the candidate's field, a recording and an assessment per question, at $1 per credit with the first 3 results free. > > Others, by need: HireVue or VidCruiter (full suite, quote), Sapia.ai (text chat), Willo or Hireflix (low flat price), Interviewer.AI (admissions). > > Stay with Talview for identity verification, proctoring, Workday or Greenhouse connectors, or many languages. > > Next step: create one interview for the workflow your current tool misses. Talview alternatives depend on which part of Talview you use. Talview sells an AI interviewer called Ivy, live interview rooms, candidate verification, and exam proctoring with an AI proctoring agent called Alvy. A team that bought it for identity and proctoring needs a different shortlist from one that bought it for the interview. This page covers seven options, Tracy AI Interviewer first, with price where it is public and the case where each is the better pick. ## Why do teams look for Talview alternatives? Three reasons, from our reading of Talview's own pages. - **No public price.** Its pricing page is a quote form, unit-based, with a typical minimum commitment of approximately 150 to 200 candidates. - **A suite built around security.** Talview calls itself a "Security partner for high-stakes testing and hiring platforms". A team that wants only a first-round interviewer is buying into a larger identity and proctoring product. - **Report contents.** Ivy's report lists the transcript, rubric scores and evidence; whether a recording sits beside the transcript is not stated. ## Which Talview alternative fits which job? Best-for and published price for each option. "Quote" means the vendor publishes no price. | Option | Best for | Published price | | --- | --- | --- | | Tracy AI Interviewer | Follow-ups in the candidate's own field, with a recording and report a reviewer can check. | $1 per credit, one credit per started 20 minutes. First 3 results free. | | HireVue | Enterprise governance, validated science and one suite for high-volume first rounds. | Quote. | | VidCruiter | Live panels, scheduling, references and testing from one vendor. | Quote, by modules and volume. | | Sapia.ai | Text-chat interviews for companies that hire more than 500 people a year. | Per hire, on request. | | Willo | Low-cost recorded-question screening for a few live roles. | £49 per live role on Lite; Enterprise from £2,999 a year. | | Hireflix | One-way video at a flat, low price for a small team. | $75 a month billed yearly under 50 employees; $150 for 50 to 250. | | Interviewer.AI | Admissions and programme interviews with a branded portal. | $199 a month for 50 credits; $399 a month for 120 credits. | ![Four-quadrant map of interview format against published price, with Tracy in the follow-ups, price-shown quadrant.](/blog/images/talview-alternatives-1.png) *Tracy AI Interviewer sits with Interviewer.AI as a two-way interview with a price on the page; the other options need a quote or are one-way or text. From each vendor's own pages. Source: vendors' public pages, Nov 2026.* Only Tracy and Interviewer.AI pair a two-way interview with a price you can read before a call. ## Tracy AI Interviewer: follow-ups and evidence at a published price Tracy is a two-way video interview run by an AI interviewer. You describe the role and she drafts the questions from your job description, or you bring your own. She asks follow-ups in the person's own field and waits for the answer, reads the CV when your CV check is on, and can run a coding task inside the same interview. A Director at a top-tier early-stage VC says "Tracy sometimes brings up exactly those questions I might have forgotten to ask, which has been a pleasant surprise." See [how AI interviews work](/blog/how-do-ai-interviews-work). After each interview you get a recording of both sides, a timestamped transcript, a written assessment for every question with its reasoning, and integrity signals with their evidence: tab changes, gaze, live face detection, delay before answers, fullscreen exits and external screens. A person decides. The package goes to a hiring manager by link or PDF ([AI interview reports](/blog/ai-interview-reports)). She has no seats and no upfront cost. Tracy connects through API and MCP, Slack, WhatsApp and Zapier ([Connect](/connect)). **Where Talview is the better pick.** Talview verifies identity with face, voice and device checks, runs proctored exams, lists more ATS connectors, says Ivy interviews in 42 languages, and lists WCAG 2.1 AA. Tracy publishes no language count or bias audit, and accessibility is not stated. ![Recreated Tracy signals panel with two flagged signals and the times they occurred, for a reviewer to check.](/blog/images/talview-alternatives-2.png) *Tracy AI Interviewer shows tab changes, delays and other signals with the time each happened, so a reviewer can interpret them; nothing is decided in the interview. Illustration with sample data* Tracy shows tab changes, gaze and other signals with their evidence, and leaves the decision to your team. ## HireVue: the enterprise suite with governance depth HireVue's AI Interviewer runs a two-way conversation with probing follow-ups, in a suite that also sells one-way and live interviews, and it lists SOC 2 Type 2. It publishes no price. Pick HireVue when governance depth and one suite at enterprise scale matter more than a published price; see [Tracy vs HireVue](/compare/tracy-vs-hirevue). ## VidCruiter: live panels, scheduling and structured rubrics VidCruiter covers pre-recorded and live interviews, scheduling, structured interviews and references, lists 75+ integrations including Workday and SAP SuccessFactors, and labels its AI Interviewer "COMING SOON". Pricing is custom by modules and hiring volume. Pick VidCruiter when you need live panels and scheduling from one vendor. ## Sapia.ai: text-chat interviews with a published bias audit Sapia.ai interviews in an untimed text chat of five structured questions, ranks every candidate with a fit score, and says it is independently audited for bias, published openly. Pricing is per hire, and Sapia.ai says it serves companies that hire more than 500 people a year best. Pick it when text is the right format and you hire at that volume. ## Willo: recorded-question screening for a few roles Willo is a recorded-question screening tool with ISO 27001, 18+ languages and 50+ native integrations. Lite is pay as you go at £49 per live role, with 150 assessed candidates per live role, and Enterprise starts at £2,999 a year. Pick Willo for low-cost screening across a few live roles. ## Hireflix: one-way video at a flat price Hireflix is one-way video interviewing that transcribes answers automatically and connects to ATS tools, Zapier and Make. It costs $75 a month billed yearly for companies under 50 employees and $150 for 50 to 250, with unlimited seats, positions and responses. Pick it when recorded answers at a flat, low price are enough. ## Interviewer.AI: admissions and programme interviews Interviewer.AI serves universities, bootcamps and training programmes, and offers a free 14-day trial. Essential is $199 a month for 50 credits and Professional is $399 a month for 120 credits. A conversational AI interview costs 2 credits and a one-way video 1, and adaptive follow-ups are listed under Professional. Pick it for admissions-style programmes. ## How to choose among Talview alternatives 1. **Start from the job Talview was doing.** If it was identity checks and proctoring, keep that layer and compare only the interview. If it was the interview, the shortlist above applies. 2. **Decide whether you need follow-ups.** For roles where a polished answer hides a shallow one, pick a two-way interview: Tracy, HireVue, or Interviewer.AI on Professional. For screening on fixed questions, Willo, Hireflix or Sapia.ai are enough. 3. **Settle price before features.** Tracy, Willo, Hireflix and Interviewer.AI publish a price. For Talview, HireVue, VidCruiter and Sapia.ai, ask for the unit, the minimum and the first-year total at your volume. 4. **Check the evidence.** Ask each vendor to open one completed result and get from the score to what the candidate said. 5. **Run one role.** Pick a role with at least 20 expected applicants and compare two options on it. To start with Tracy, send us the role and we build the interview plan. For technical roles, [technical interviews](/use-cases/technical-interviews) shows how Tracy handles them, and [the best AI interview software](/blog/best-ai-interview-software) compares the wider market. ## Which path fits you? - Engineering roles that need follow-ups: [technical interviews](/use-cases/technical-interviews). - High-volume first rounds: [volume hiring](/use-cases/volume-hiring). - Specialist and expert screening: [expert vetting](/use-cases/expert-vetting). - Tracy's price and Enterprise terms: [pricing](/pricing). ## Common questions ### What are the best Talview alternatives? For follow-up interviews with a published price, Tracy AI Interviewer. For an enterprise suite, HireVue or VidCruiter. For low-cost recorded answers, Willo or Hireflix. For text-chat interviews at 500+ hires a year, Sapia.ai. ### Is there a cheaper alternative to Talview? Talview publishes no price, so a cheaper option cannot be shown in dollars. Tracy, Willo, Hireflix and Interviewer.AI publish theirs, so you can compare them with your own quote. ### Does Tracy replace Talview's proctoring? No. Tracy shows integrity signals, with their evidence, to a person who reviews them. Talview verifies identity and runs proctored exams. If you need proctoring, keep it and compare only the interview. ### How do I try Tracy before paying? The first 3 results are free, no card. --- Create one interview for the workflow your current tool misses. Send us the role and we will build the plan, customised in days, with a recording and report you can check. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=talview-alternatives&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=talview-alternatives&utm_content=try-free) ## Sources - [Talview, Pricing](https://www.talview.com/en/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Home](https://tracyinterview.ai/), retrieved 9 Nov 2026. # Talview Pricing (Nov 2026) > Talview pricing is quote-only, with a typical 150 to 200 candidate minimum. Tracy's is public: $1 per credit. Costs at 100, 500 and 1,000 interviews. - URL: https://tracyinterview.ai/compare/talview-pricing - Published: 2026-11-12 - Author: Tracy team > **In short** Tracy AI Interviewer gives you a price before the sales call: $1 per credit, one credit per started 20 minutes, so 100 interviews of up to 20 minutes cost $100. The first 3 results are free, no seats. > > Each credit buys a checkable result: follow-ups in the candidate's field, a recording and an assessment per question. > > Talview publishes no price: its page is a quote form with a typical minimum of about 150 to 200 candidates. Choose it when proctoring or identity checks decide. > > Next step: try three real interviews free. Talview pricing is not public. The pricing page is a quote form, and a Solutions Expert contacts you afterwards. Tracy AI Interviewer publishes her price on [Tracy's pricing](/pricing): $1 per credit, with each applicant starting on click, at any hour, with no scheduling. This page sets out what Talview's pricing page says, what Tracy costs at 100, 500 and 1,000 interviews, what to ask on a quote call, and how the two compare. ## What is public in Talview pricing? No plan names, prices or tiers. The quote form asks which products you want (video interviews, candidate verification, Ivy, its AI interviewer, interview security, Alvy, its AI proctoring agent and more) and how many candidates you screen a year, from "Less than 200" to "More than 5,000". The FAQ holds the only commercial terms Talview states. What the Talview pricing page states, line by line. | What the page covers | What Talview says | Exact words | | --- | --- | --- | | Pricing model | Unit-based. Each feature has its own unit, and you pay for what you use. | "Our pricing model is unit-based, which means each feature has a specific unit assigned to it." | | Minimum commitment | About 150 to 200 candidates, typically, to get started. | "To get started, a minimum commitment of approximately 150 to 200 candidates is typically required." | | Pilot | Extra charges may apply above 10 candidates. | "For the pilot phase, additional charges may apply for more than 10 candidates." | | Trial or demo | A product walkthrough. No self-serve free trial is stated. | "Talview offers a product walkthrough" | ## What does Tracy cost at 100, 500 and 1,000 interviews? She charges one credit for each started 20 minutes of an interview that produced a result, and a credit costs $1. A 15-minute interview is $1 and a 34-minute interview is $2. Credits never expire, a session with no answers is never charged, and API and MCP access (see [Connect](/connect)) is part of Pay as you go. The first 3 results are free, and the totals ignore that. Every applicant starts on click with no recruiter present, so cost follows interviews finished, not seats or scheduling time. See also [what AI interviews cost](/blog/ai-interview-cost). Tracy's figures follow from $1 per credit and one credit per started 20 minutes. Talview publishes no figure. | Interviews | Tracy, up to 20 minutes each (1 credit) | Tracy, 21 to 40 minutes each (2 credits) | Talview | | --- | --- | --- | --- | | 100 | $100 | $200 | Not public. Quote on request. | | 150 to 200 | $150 to $200 | $300 to $400 | Not public. Its typical minimum is about 150 to 200 candidates. | | 500 | $500 | $1,000 | Not public. Quote on request. | | 1,000 | $1,000 | $2,000 | Not public. Quote on request. | ## What does a minimum commitment mean for a small team? This is our reasoning, not a Talview price. Talview says pricing is unit-based and a minimum commitment of about 150 to 200 candidates is typically required. A team that expects 60 candidates in the first year may be asked to commit to more than it uses. Tracy at 60 interviews of up to 20 minutes costs $60. Units also attach to features, so ask how many units one candidate uses from invite to report. ![Four steps covered by one Tracy credit: interview, recording, assessment per question, and a decision by a person.](/blog/images/talview-pricing-2.png) *Tracy AI Interviewer: what one credit covers, so the price buys a result a reviewer can check. Illustration of Tracy's output. Source: Tracy pricing page, Nov 2026.* One credit covers an interview with follow-ups, a recording and timestamped transcript, a written assessment per question, and a result your team decides on. ## What should you ask on a Talview quote call? 1. Which products and units are in the quote, and how many units does one candidate use? 2. What is the price per unit, and per candidate for the flow we will run? 3. What minimum commitment applies to us, in candidates and dollars, and does unused volume carry over? 4. What does the pilot cost beyond 10 candidates? 5. What are the term, the renewal rule and the price change at renewal? 6. Are implementation, ATS setup and support included? And what do we pay for a candidate who starts and does not finish? Tracy never charges for a session with no answers. ## How do Talview and Tracy compare on the four pillars? The table covers Depth, Volume, Stack and Evidence, then price and service. Talview lists WCAG 2.1 AA, a 42-language claim, SOC 2, ISO 27001 and GDPR, and a log of every proctor and recruiter action. Tracy's accessibility is not stated, and it publishes no bias audit. Tracy is SOC 2 Type II attested. Tracy and Talview on depth, volume, stack, evidence, price and support, from each vendor's own pages. | Pillar | Tracy AI Interviewer | Talview | What to ask | | --- | --- | --- | --- | | Depth | Two-way video interview. She asks follow-ups in the person's own field and waits for the answer. A coding task can sit inside the interview. | Ivy builds a structured plan and asks contextual follow-ups in real time by voice, video or live coding, scored against your rubric. | Ask for a live Ivy run on one of your roles. | | Volume | Every applicant starts on click, at any hour, with no scheduling. Each interview is ready in under 10 seconds. A dropped applicant can rejoin within 10 minutes (configurable per customer). | Says thousands of interviews at once with no scheduling; separate bulk and panel scheduling tools. | Ask what happens when 500 candidates start in one day. | | Stack | API and MCP connections (ChatGPT, Claude, Mistral, Gemini, Copilot Studio), Slack, WhatsApp and Zapier, and use beside the Workable and Breezy HR connectors. | Lists Workday, SuccessFactors, Greenhouse, SmartRecruiters, Taleo, iCIMS and others, plus REST APIs. | Name your ATS and ask which connector is included. | | Evidence | A recording of both sides, a timestamped transcript, a written assessment per question with its reasoning, and integrity signals with their evidence. A person decides. | A report with the full transcript, rubric scores and evidence behind every score; face, voice and device checks; a log of every proctor and recruiter action. | Ask to open one result and get from the assessment to the candidate's words. | | Price | $1 per credit, one credit per started 20 minutes. First 3 results free, no seats, no upfront cost, and an empty session is never charged. | Not public. A quote form; unit-based, with a typical minimum of about 150 to 200 candidates. | Ask for the unit and the first-year total at your volume and at twice your volume. | | Service and support | Enterprise includes a dedicated customer manager, a dedicated support channel, a setup customised to your workflow and a walkthrough. Tracy is customised in days per customer need. | Says it deploys in one week; lists phone lines in the US and India, a help center and ticket status. | Ask each vendor for the onboarding steps and the date the first interview can run. | ## Is Talview worth the quote, and for whom? It is, if proctoring and identity are the purchase. Talview describes itself as a "Security partner for high-stakes testing and hiring platforms", and its pages cover ID checks, face and voice matching, deepfake scans, a secure browser, live proctoring and exam delivery. A team that gives high-stakes exams, or must verify who is on screen for every candidate, should chase the quote. Tracy is not a substitute for proctored exams. Tracy is the better deal when you need a first-round interview with follow-ups, a report your team can check, and a cost you can compute. One customer cut its average time to hire from 44 days to 7, and Tracy did the work of 53 full-time recruiters in one month. To see your own cost, run three real interviews free, then multiply by your yearly volume. ## Which path fits you? - High-volume first rounds: [volume hiring](/use-cases/volume-hiring). - Engineering roles that need follow-ups: [technical interviews](/use-cases/technical-interviews). - Tracy's full price and Enterprise terms: [pricing](/pricing). ## Common questions ### How much does Talview cost? Talview publishes no price. Its pricing page is a quote form, and Talview says pricing is unit-based, with a typical minimum commitment of about 150 to 200 candidates. Tracy's price is $1 per credit, with the first 3 results free. ### Does Talview have a free trial? Its pricing page offers a product walkthrough and states no self-serve free trial. Tracy gives the first 3 results free, with no card needed. ### Is Talview priced per seat or per candidate? Its page says unit-based, with a unit assigned to each feature, and does not say whether seats are charged. Tracy charges no seats. ### What does Tracy cost compared with Talview? Talview's cost is not public, so no comparison in dollars is possible. Tracy costs $100 for 100 interviews of up to 20 minutes, $500 for 500 and $1,000 for 1,000. --- Try three real interviews free and calculate your completed-interview cost. Bring your annual volume, interview length and stack, and we will customise the interview plan in days, with a recording and report you can check. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=talview-pricing&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=talview-pricing&utm_content=try-free) ## Sources - [Talview, Pricing](https://www.talview.com/en/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Home](https://tracyinterview.ai/), retrieved 9 Nov 2026. # TestGorilla AI Interviews: Pricing and What You Get (Nov 2026) > TestGorilla pricing: AI interviews cost 2 credits each inside a plan from $520 a month. See the cost at 100 to 1,000 interviews next to Tracy's $1. - URL: https://tracyinterview.ai/compare/testgorilla-ai-interview-pricing - Published: 2026-11-12 - Author: Tracy team > **In short** Tracy AI Interviewer prices the interview itself: $1 per credit, one credit per started 20 minutes, the first 3 results free, no seats. > > She asks follow-ups in the person's own field. One customer cut its average time to hire from 44 days to 7. > > TestGorilla's AI interviews cost 2 credits each inside a plan, with the full library on Plus from $520 a month billed annually. It suits teams that want validated skills tests in one account. > > Next step: try 3 results free and calculate your cost per completed interview. TestGorilla publishes its plan prices. The catch is where AI interviews sit: inside Gorilla Assessments, a per-account plan, at 2 credits per candidate. Tracy AI Interviewer sells the interview alone and charges per result, so a first round never needs a plan. ## What does TestGorilla pricing include for AI interviews? Gorilla Assessments is priced per account and holds the skills tests, video questions and AI interviews. Gorilla Sourcing, a separate candidate search tool, is priced per seat. Annual billing is shown with a 25 percent saving; monthly figures are not printed. Source: TestGorilla pricing page, Nov 2026. TestGorilla Gorilla Assessments plans, from its pricing page, Nov 2026. | Plan | Price | AI interviews | Also included | | --- | --- | --- | --- | | Free | $0 a month, 10 credits a month. | Limited access: customised interviews only, 2 credits per candidate. | 5 essential skills tests, resume scoring, 1 premium user role. | | Core | From $215 a month per account, billed annually. | Not listed. Video questions from the library only. | 350+ tests, custom questions, a shared credit pool, 2 premium user roles. | | Plus | From $520 a month per account, billed annually. | Full library (100+) and custom AI interviews, 2 credits per candidate. | Job simulations, custom coding challenges, ID verification, ATS and API integrations, personalised onboarding. | ## How do TestGorilla credits work? A skills test costs 1 credit per candidate, video questions 1, an AI interview 2 and ID verification 1 on Plus. Resume scoring and custom questions are free, and only the start of an assessment is charged. Credits are shared across the account, pooled for the year on annual plans. TestGorilla's library AI interviews run about 26 minutes; on Tracy a 26-minute interview is 2 credits, or $2. The page does not state how many credits Core or Plus include. ## What does Tracy cost at 100, 500 and 1,000 interviews? Tracy's [pricing page](/pricing) has no plan to decode. One credit covers each started 20 minutes of an interview that produced a result, and a credit costs $1. Credits never expire, a session with no answers is never charged, and API and MCP access (see [Connect](/connect)) is included. The first 3 results are free, so the totals ignore them. Bring your annual volume and interview length to a demo and we will work out your cost per completed interview. Tracy's cost follows from $1 per credit; TestGorilla AI interviews use 2 credits each. | Interviews | Tracy, up to 20 minutes (1 credit) | Tracy, 21 to 40 minutes (2 credits) | TestGorilla AI interviews (2 credits) | | --- | --- | --- | --- | | 100 | $100 | $200 | 200 credits. Plus from $520 a month; credits included not stated. | | 500 | $500 | $1,000 | 1,000 credits. Plus from $520 a month; credits included not stated. | | 1,000 | $1,000 | $2,000 | 2,000 credits. Plus from $520 a month; credits included not stated. | ## How does a plan fee change the cost per interview? Our arithmetic on TestGorilla's starting price, not a quote: Plus from $520 a month is $6,240 a year, and the same $6,240 buys 6,240 Tracy interviews of up to 20 minutes. The plan also holds tests, simulations, ID checks and ATS sync, so a team that uses those gets more for the fee. At 100 AI interviews a month the fee is $5.20 per interview against $100 in total on Tracy. At 1,000 a month it is $0.52 per interview, and if the plan's credits covered them TestGorilla would cost less than Tracy's $1. ![Three steps: up to 20 minutes for one $1 credit, a result with recording, transcript and assessment, and your team decides.](/blog/images/testgorilla-ai-interview-pricing-1.png) *What one Tracy AI Interviewer credit covers: up to 20 minutes, a result with recording, transcript and assessment, and a decision by your team. Illustration with sample data* One Tracy credit covers an interview of up to 20 minutes and leaves a result a reviewer can check. ## How do TestGorilla and Tracy compare on the four pillars? Price is one line. The table compares what the interview gives you, plus the buying criteria the pillars leave out. TestGorilla leads on validation and ATS depth. Tracy prices by the interview and keeps a recording and timestamped transcript behind each result. For the wider cost picture, see [AI interview cost](/blog/ai-interview-cost). Tracy and TestGorilla compared on depth, volume, stack, evidence, support and security. | Criterion | Tracy AI Interviewer | TestGorilla | | --- | --- | --- | | Depth | Two-way video. She reads the CV when the CV check is on, asks in the person's own field, follows up and waits for the answer. A coding task can sit inside the interview. | An AI interviewer asks follow-ups within 3 to 9 structured questions, or your own questions and scoring criteria. A coding IDE and coding challenges are listed as separate features; coding inside the AI interview: Not stated. | | Volume | Every applicant starts on click, at any hour, with no scheduling. Ready in under 10 seconds; a dropped applicant can rejoin within 10 minutes (configurable per customer). | Candidates complete AI interviews on their own time. Start and rejoin rules: Not verified. | | Stack | API and MCP connections (ChatGPT, Claude, Mistral, Gemini, Copilot Studio), Slack, WhatsApp and Zapier, and use beside the Workable and Breezy HR connectors. | Greenhouse, Workable, Lever, Workday, SAP SuccessFactors and more, plus Zapier, Make and an open API. Scores sync back to the ATS. Plus plan. | | Evidence | A recording of both sides, a timestamped transcript, a written assessment per question, and integrity signals with their evidence. A person decides. | Scores with reasoning against your rubric, editable by the recruiter; a log of tab switches and full-screen exits. Recording and transcript: Not verified. | | Support and implementation | Enterprise: a dedicated customer manager, a dedicated support channel, a setup customised to your workflow in days. | 24/7 live chat and email on every plan; personalised onboarding on Plus. | | Security, validation and bias | SOC 2 Type II attested. A person reviews the evidence and decides. Tracy publishes no bias audit. Accessibility: Not stated. | Lists SOC 2 Type II and GDPR. Test fact sheets with reliability and validity figures and adverse impact monitoring. Accessibility: Not verified. | ## What should you ask TestGorilla before you buy? 1. How many credits do Core and Plus include each month? 2. What does an extra credit cost, and do unused credits roll over? 3. Is the AI interview on Core, or only on Plus and the free plan? 4. What is the monthly price if we do not pay annually? 5. Are the ATS connector and API in our plan? 6. Is a recording and transcript of each AI interview kept? ## Is TestGorilla worth its price, and for whom? TestGorilla is worth it when skills tests are the center of your screening: 350+ tests with validity monitoring, plus video questions and AI interviews in one assessment. Tracy is the better buy when the first round is a conversation. A reviewer opens a recording and timestamped transcript behind every result, and the bill is $1 per started 20 minutes with no plan. See [hiring at volume](/use-cases/volume-hiring) and [technical interviews](/use-cases/technical-interviews). ## Which path fits you? - [Hiring at volume](/use-cases/volume-hiring): in-house teams, staffing firms and graduate programmes. - [Technical interviews](/use-cases/technical-interviews): role-specific interviews with follow-ups. - [Tracy's pricing](/pricing): Pay as you go and Enterprise to your volume. ## Common questions ### How much does TestGorilla cost? Gorilla Assessments is Free at $0, Core from $215 a month and Plus from $520 a month per account, billed annually. Gorilla Sourcing is priced per seat. AI interviews use 2 credits per candidate. ### Does TestGorilla have a free plan? Yes. It gives 10 credits a month and limited access to AI interviews, which cost 2 credits each. ### How much do TestGorilla AI interviews cost? 2 credits per candidate. The full AI interviews library is on Plus, from $520 a month billed annually. ### Is TestGorilla cheaper than Tracy? At 100 interviews a month Tracy costs $100 for interviews of up to 20 minutes, below the starting price of a Plus plan. Ask TestGorilla for the credits each plan includes before comparing at higher volume. ### Is there a free way to try Tracy? Yes. The first 3 results are free, whatever their length, and no card is needed. After that there are no seats and no subscription. --- Try three real interviews free and calculate your cost per completed interview. Bring your annual volume, interview length and stack, or send us one role and we build the interview plan, customised in days, with a recording and report you can check. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=testgorilla-ai-interview-pricing&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=testgorilla-ai-interview-pricing&utm_content=try-free) ## Sources - [TestGorilla, Pricing](https://www.testgorilla.com/pricing/), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. # TestGorilla Alternatives for AI Interviews (Nov 2026) > TestGorilla alternatives for AI interviews: Tracy, HackerRank, CodeSignal, HeyMilo, HireVue and Spark Hire, with prices, fit and where each one wins. - URL: https://tracyinterview.ai/compare/testgorilla-alternatives - Published: 2026-11-12 - Author: Tracy team > **In short** Tracy AI Interviewer is the first to try when the first round must show how someone reasons in their own field: follow-ups in a two-way video interview, a coding task inside it, and a recording and transcript behind every result. She costs $1 per credit, with the first 3 results free. > > HackerRank or CodeSignal fit coding in an IDE. TestGorilla fits validated skills tests. HeyMilo and HireVue need a quote, and Spark Hire fits recorded video from $249 a month. > > Next step: create one interview for the workflow your current tool misses. TestGorilla is built around a library of skills tests, and its AI interview is one tool in an assessment. When the first round has to show depth in a technical or expert role, Tracy AI Interviewer is the first tool to try. Where another tool is the better pick, we say so. ## Why do teams look for TestGorilla alternatives? The full library of 100+ AI interviews and custom interviews are Plus features, from $520 a month per account billed annually, and the page does not state how many credits Core or Plus include. ATS connectors and API access are Plus only. The interview is 3 to 9 structured questions with follow-ups. Coding challenges are listed as separate tests; coding inside the AI interview is not stated. Tracy shows her price on the [pricing page](/pricing): $1 per credit. ## Tracy AI Interviewer: follow-ups in the person's field Tracy interviews every applicant by two-way video. She reads the CV when the CV check is on, asks in the person's own field, follows up on what they said and waits for the answer. Behind every result sit a recording of both sides, a timestamped transcript, a written assessment per question and integrity signals for a human reviewer. A coding task can sit inside the interview for technical roles (see [technical interviews](/use-cases/technical-interviews)). Every applicant starts on click, at any hour. She connects through API and MCP, Slack, WhatsApp and Zapier. The price is $1 per credit, the first 3 results free, no seats. A Director at a top-tier early-stage VC said of her: "Tracy sometimes brings up exactly those questions I might have forgotten to ask, which has been a pleasant surprise." Where she falls short of TestGorilla: she is an interviewer built around the conversation, so keep a test library if tests are part of your screen, and she publishes no bias audit. ![Transcript card: Tracy asks, the candidate answers, and Tracy follows up on what breaks if two changes land on one day.](/blog/images/testgorilla-alternatives-1.png) *Tracy AI Interviewer asks a follow-up on what the candidate just said. Illustration with sample data* Each follow-up comes from the candidate's answer, so a reviewer sees reasoning rather than a preset list. ## HackerRank: an agentic interviewer for engineering teams HackerRank's AI interviewer, Chakra, opens a canvas with a real-world problem and evaluates judgment, critical thinking and AI fluency. It is included on every plan. Starter is $79 a month ($948 billed annually) for 60 candidate attempts a year, about $16 each; Pro is $419 a month ($5,028 annually) for 360 attempts, about $14 each; extra attempts are $20. It offers a free trial. Pick it for software engineering interviews inside a development environment. ## CodeSignal: validated assessments and AI interviewers by role family CodeSignal's Hire plans are credit based. Build is $79 a month billed annually for 60 credits (about $16 each); Grow is $479 a month billed annually for 420 credits (about $14 each); Pro is custom. A credit is used when a candidate starts an assessment, an AI interview or a live interview, and extra credits are $20. Its AI interview is typically 15 minutes, probes thin answers and ends in a report with an overall score. Its assessments are validated by I/O psychologists, a stronger validation story than Tracy publishes. Pick it for validated assessments and AI interviewers across engineering and business roles. ## HeyMilo: voice, video and text screening, priced by demo HeyMilo screens by phone, video, SMS and web form, in 20+ languages, with 20+ ATS partners. Its AI asks follow-ups in real time and its reports carry timestamped candidate quotes. Its pricing page sends you to a demo form, and a coding interview is not stated on its product pages. Pick it for multilingual, high-volume screening by phone or SMS. See [Tracy vs HeyMilo](/compare/tracy-vs-heymilo). ## HireVue: the enterprise suite with governance documents HireVue offers on-demand and live video interviews, AI evaluation with an explanation, annual independent audits and adverse impact testing before models go live. Its pricing page lists no plan or price. Pick it when governance documents are a procurement requirement and you accept a quote. See [Tracy vs HireVue](/compare/tracy-vs-hirevue). ## Spark Hire: recorded and live video at a published price Spark Hire starts from $249 a month billed annually, with unlimited live and one-way video interviews and users. It lists 40+ ATS integrations, transcription in 50+ languages and SOC 2 Type II, and people decide. AI follow-up questions are not stated. Pick it when a lean team wants to replace phone screens with recorded video and review the answers itself. ## TestGorilla and its alternatives compared on four pillars and price Tracy and six alternatives on depth, volume, stack, evidence and price, from each vendor's own pages, Nov 2026. | Tool | Depth | Volume | Stack | Evidence | Price | | --- | --- | --- | --- | --- | --- | | Tracy AI Interviewer | Two-way video, follow-ups in the person's field, a coding task inside the interview | Starts on click, any hour, no scheduling | API, MCP, Slack, WhatsApp, Zapier | Recording, transcript, assessment per question, integrity signals | $1 per credit, first 3 results free | | TestGorilla | AI interviewer with follow-ups, 3 to 9 questions; skills tests and a coding IDE alongside | Candidates complete on their own time | Greenhouse, Workable, Lever, Workday and more, plus API, on Plus | Score reasoning per criterion; behavior log of tab switches and full-screen exits | Plus from $520 a month; AI interview 2 credits | | HackerRank | Agentic interviewer on a live problem canvas; live pair programming | Yearly attempt allowance by plan | ATS and calendar integrations on Pro; API and MCP on Enterprise | Score per competency with a why backed by transcripts and work | From $79 a month; $20 per extra attempt | | CodeSignal | Short structured AI interview that probes thin answers; IDE assessments | Credits used when a candidate starts | Gem and Ashby on core plans; Workday, iCIMS and Oracle on Enterprise | Report with overall score and skill evaluation; Suspicion Score and other fraud layers | From $79 a month; $20 per extra credit | | HeyMilo | Voice, video, SMS and form screening with real-time follow-ups | Around the clock, 20+ languages | 20+ ATS partners | Timestamped candidate quotes; fraud and impersonation flags | Not public | | HireVue | On-demand and live video; AI scoring with an explanation | Mid-market to enterprise first rounds | Not verified | Annual independent audits; adverse impact testing | Not public | | Spark Hire | One-way and live video; AI follow-ups not stated | Unlimited interviews and users | 40+ ATS integrations | Recorded answers with an AI summary; people decide | From $249 a month | ![Tracy, HackerRank, CodeSignal show interview prices; HeyMilo, HireVue quote; TestGorilla, Spark Hire price tests or video.](/blog/images/testgorilla-alternatives-2.png) *Tracy AI Interviewer and six other tools by what each is built around and whether a price is public. Source: vendor pages, Nov 2026. Source: vendor pages, Nov 2026.* Pick by what each tool is built around and whether its price is public. ## How do you choose a TestGorilla alternative? 1. Tests or conversations? If scored skills tests across many roles are the point, stay with TestGorilla or look at CodeSignal. If the first round has to show how someone reasons, you need a tool that asks the second question: Tracy, HackerRank for engineering, or HireVue at a quote. See [one-way vs two-way video interviews](/blog/one-way-vs-two-way-video-interviews). 2. How will you pay? A yearly plan favours steady volume. A per-result price favours uneven volume and a trial on one role. See [AI interview cost](/blog/ai-interview-cost). 3. What must it connect to? Name your ATS and ask which plan includes the connector. Tracy connects through API and MCP and works beside Workable and Breezy HR; [Connect](/connect) lists the rest. Send us the interview your current tool handles worst and we will draft it in Tracy. Free starts: TestGorilla's free plan has 10 credits a month and limited access to AI interviews at 2 credits each, Tracy's first 3 results are free, and HackerRank offers a free trial. - [Technical interviews](/use-cases/technical-interviews): role-specific interviews with follow-ups. - [Volume hiring](/use-cases/volume-hiring): in-house teams, staffing firms and graduate programmes. - [Tracy's pricing](/pricing): Pay as you go and Enterprise to your volume. ## Common questions ### What is the best alternative to TestGorilla for AI interviews? It depends on what the first round must show. Tracy AI Interviewer is the pick for follow-up questions in the person's own field, with a recording and a timestamped transcript behind every result and a published price of $1 per credit. HackerRank and CodeSignal suit engineering teams that want an IDE. ### Is there a free alternative to TestGorilla? TestGorilla has its own free plan. Among the alternatives, Tracy's first 3 results are free with no card, and HackerRank offers a free trial of its AI interviewer. ### How does TestGorilla compare with HackerRank? TestGorilla is a library of 350+ skills tests across many roles, with an AI interview and a coding IDE. HackerRank is built for engineering hiring and includes its AI interviewer on every plan, from $79 a month. ### Does Tracy replace TestGorilla's skills tests? Tracy is an interviewer, not a test library. A coding task can sit inside the interview, with follow-ups on approach and trade-offs. Teams that need a library of tests keep one. ### Which TestGorilla competitors publish a price? Tracy ($1 per credit), HackerRank (from $79 a month), CodeSignal (from $79 a month billed annually) and Spark Hire (from $249 a month billed annually). HeyMilo and HireVue do not. --- Create one interview for the workflow your current tool misses: send us one role and we build the interview plan, customised in days, with a recording and report you can check. The first 3 results are free. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=testgorilla-alternatives&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=testgorilla-alternatives&utm_content=try-free) ## Sources - [TestGorilla, Pricing](https://www.testgorilla.com/pricing/), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Connect](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. # Tracy vs Talview (Nov 2026) > Talview AI interview vs Tracy for technical volume hiring: where Ivy leads on identity and proctoring, where Tracy leads on follow-ups, evidence and price. - URL: https://tracyinterview.ai/compare/tracy-vs-talview - Published: 2026-11-12 - Author: Tracy team > **In short** Tracy AI Interviewer is the stronger first round when a technical role needs follow-ups in the candidate's field and a report a reviewer can check: a recording of both sides, a timestamped transcript and a written assessment per question. A person decides. > > She costs $1 per credit, with the first 3 results free and no seats. > > Talview is the better pick for face, voice and device verification, proctored exams, Workday or Greenhouse connectors, or many languages. > > Next step: send us one role. This Talview AI interview vs Tracy comparison is for technical teams that hire at volume. Talview's AI interview is Ivy, one product in a suite that also covers live interview rooms, candidate verification and exam proctoring. Tracy AI Interviewer is a two-way video interview that asks follow-ups in the person's own field and leaves a recording and report a reviewer can check. Both describe an interviewer that listens and follows up. Talview adds identity and audit tooling around the interview. Tracy adds a recording and report you can open, and a published price. ## What does the Talview AI interview do well? Talview calls itself a "Security partner for high-stakes testing and hiring platforms", and the interview product shows that heritage. Ivy reads the role, builds a structured interview plan, runs the conversation and scores each answer against your rubric. It asks contextual follow-ups in real time by voice, video or live coding, and the report holds the full transcript, rubric scores and the evidence behind every score. Around Ivy sit the parts Tracy does not sell. Candidate verification checks face, voice and device before and during the interview, and a deepfake scan adds a 360-degree environment check. Interview scheduling handles one-on-one, panel and bulk bookings with calendar sync. The integrations page lists Workday, SuccessFactors, Greenhouse, iCIMS, SmartRecruiters and more. Talview lists SOC 2 Type II, ISO 27001, GDPR and WCAG 2.1 AA. It says every proctor and recruiter action is logged with a reason, and that a human makes the final call. ## Talview vs Tracy: buyer criteria side by side Talview column: Talview's own pages. Tracy column: approved Tracy product records. "Not stated" and "Not verified" mean the page does not say, not that the feature is absent. | Criterion | Tracy AI Interviewer | Talview | Who has the edge on what is published | | --- | --- | --- | --- | | Depth | Two-way video interview. She asks follow-ups in the person's own field and waits for the answer. Questions come from your job description or your own list, and a coding task can sit inside the interview. | Ivy builds a structured plan and asks contextual follow-ups in real time by voice, video or live coding, scored against your rubric. | Level on what is published. Ask each for a live run. | | Volume | Every applicant starts on click, at any hour, with no scheduling. Each interview is ready in under 10 seconds, and a dropped applicant can rejoin within 10 minutes (configurable per customer). | Says thousands of interviews at once with no scheduling; bulk and panel scheduling tools; built for BPOs, ITeS, contact centers and technology organizations. | Level. Both start applicants without a recruiter. | | Stack | API and MCP connections (ChatGPT, Claude, Mistral, Gemini, Copilot Studio), Slack, WhatsApp and Zapier, and use beside the Workable and Breezy HR connectors. | Workday, SuccessFactors, Greenhouse, SmartRecruiters, Taleo, iCIMS and others; Microsoft Teams; REST APIs. | Talview on ATS depth; Tracy on chat and MCP connections. | | Evidence | A recording of both sides, a timestamped transcript, a written assessment per question with its reasoning, and integrity signals with their evidence. A person decides. | Full transcript, rubric scores and the evidence behind every score; face, voice and device checks; deepfake scans. | Talview on identity evidence. Level on transcript and score evidence. Tracy shows a recording of both sides beside a per-question assessment; an Ivy recording is not stated. | | Price | $1 per credit, one credit per started 20 minutes. First 3 results free, no seats, no upfront cost, and an empty session is never charged. | Not public. A quote form; unit-based, with a typical minimum of about 150 to 200 candidates. | Tracy. | | Accessibility | Not stated. | Lists WCAG 2.1 AA and describes accommodations such as flexible timing and alternative formats. | Talview. | | Multilingual | A spoken-language check scores fluency, coherence, grammar, pronunciation and vocabulary from 0 to 100 beside the transcript. No language count published. | Says Ivy interviews in 42 languages. | Talview on published languages. | | Security and compliance | SOC 2 Type II attested, with the report on request at hello@tracyinterview.ai. | Lists SOC 2 Type II, ISO 27001 and GDPR, and a longer list on its Trust Center. | Talview lists more. Level on SOC 2 Type II. | | Validation, bias controls and human oversight | A person reviews the evidence report and decides, with the full history kept on the record. Tracy publishes no bias audit. | Says its AI supports human judgement, cites third-party audits and monitors interviewer bias. | Talview documents more. | Talview leads on identity verification, proctoring, ATS depth, published languages, accessibility and audit trail. Tracy leads on a published price and on chat and MCP connections. On follow-ups the published claims are level, and the live run is where they separate. For the cost terms, see [Tracy's pricing](/pricing). ## How the published workflows differ for the same technical role Take one invented role, a site reliability engineer on a payments team. Talview says Ivy builds a structured plan, asks contextual follow-ups in real time and scores each answer against your rubric. Tracy drafts the questions from your job description, then asks follow-ups in the person's field. What follows is an illustration of a Tracy interview, not a test of either product. ![Recreated Tracy interview screen: a follow-up asks what to check when queue lag rises after no recent deploy.](/blog/images/tracy-vs-talview-1.png) *Tracy AI Interviewer asks the next question from the candidate's last answer, so a reviewer sees how the person reasons. Illustration with sample data* Tracy asks the next question from the candidate's last answer. Here is a chain Tracy could run on that role, with an invented candidate. > Tracy: Your queue lag starts climbing at 3 a.m. What do you check first? > > Candidate: Whether a deploy went out in the last hour. > > Tracy: The last deploy was two days ago and lag is still rising. What do you look at next? The second question comes from the first answer, not from a list. The recording keeps both sides, and the transcript keeps the timestamps. A Director at a top-tier early-stage VC says "Tracy sometimes brings up exactly those questions I might have forgotten to ask, which has been a pleasant surprise." ## Can a reviewer check a Talview or a Tracy result? Both can, in different ways. Talview says Ivy's report carries the full transcript, rubric scores and the evidence behind every score, and that every proctor and recruiter action is logged with a mandatory reason. That audit trail is the stronger of the two. Tracy's result holds the recording of both sides, the timestamped transcript and a written assessment for every question with its reasoning. A recruiter decides: they can move a candidate forward or invite them to interview again, and the full history stays on the record. The package, with resume match, interview score, strengths and gaps, goes to a hiring manager by link or PDF. [AI interview reports](/blog/ai-interview-reports) shows what a report holds. ![Four steps of a Tracy result: recording, transcript, assessment per question, and a decision by your team.](/blog/images/tracy-vs-talview-2.png) *Tracy AI Interviewer gives a reviewer the recording, the timestamped transcript and a written assessment per question. Illustration of Tracy's output. Illustration of Tracy's output.* A Tracy result sits beside the recording and a timestamped transcript, and your team decides. ## How does Talview verify who is on screen, and what does Tracy show? Talview checks a selfie or ID document against a base image, matches face and voice continuously through the session, flags a second person in frame or an empty seat, flags device switches, and scans the environment for deepfake artifacts. Results go to a recruiter to confirm or flag. For a team that fights proxy candidates, this is the deciding section, and Talview wins it. Tracy captures tab changes, abnormal eye tracking, live face detection, delay before answers, leaving fullscreen and external screens, and shows the evidence to a person. Nothing is decided in the interview, candidates are told on the join screen that the conversation is recorded, and the decision stays with your team. The signals are for a reviewer to interpret and are not proof. ID document and voice checks are not stated for Tracy, so ask if you need them. [AI interview cheating detection](/blog/ai-interview-cheating-detection) covers what signals show and what they do not. ## Does Tracy hold up at volume? One customer cut its average time to hire from 44 days to 7, and Tracy did the work of 53 full-time recruiters in one month. Each applicant starts on click, so interviews need no recruiter's calendar. See [hiring at volume](/use-cases/volume-hiring) for how that works. Talview says Ivy runs thousands of interviews at once with no scheduling, and aims at BPOs, contact centers and technology organizations. ## How do you test both before you commit? Run one role, one team and two weeks. For the mechanics first, read [how AI interviews work](/blog/how-do-ai-interviews-work). 1. Pick one technical role with at least 20 expected applicants. 2. Ask Talview for a live Ivy run on that role, and ask for the quote in writing: unit, minimum and first-year total. 3. Run the same role in Tracy: send us the role and we will build the interview plan, then you get a recording and report you can check. 4. Have one reviewer open three results from each: a strong, a middling and a weak one. Ask how many minutes it takes to get from the score to the candidate's own words. 5. Compare the cost per finished interview and the reviewer's time. ## Which path fits you? - Technical hiring: [technical interviews](/use-cases/technical-interviews) covers role-specific interviews with follow-ups. - High-volume hiring: [volume hiring](/use-cases/volume-hiring) covers pilots for large applicant pools. - Expert vetting: [expert vetting](/use-cases/expert-vetting) covers interviews for specialists and contributors. - Integrations: [Connect](/connect) lists the connections, and [pricing](/pricing) explains Enterprise pricing to your volume. ## Common questions ### What is the Talview AI interview? It is Ivy, Talview's AI interviewer. Talview says Ivy reads the role, builds a structured interview plan, runs the conversation by voice, video or live coding, and scores answers against your rubric, with a human making the final call. ### How much does a Talview AI interview cost? Talview publishes no price. Its pricing page is a quote form, pricing is unit-based, and a minimum commitment of approximately 150 to 200 candidates is typically required. Tracy costs $1 per credit, one credit per started 20 minutes. ### Does Tracy verify identity? Tracy shows tab changes, gaze, live face detection, delay before answers, fullscreen exits and external screens as evidence for a person to review. ID document and voice checks are not stated for Tracy, so ask if you need them. --- Run one role in Tracy and compare the follow-ups and report. Send us the role and we will build the interview plan, customised in days, with a recording and report you can check. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=tracy-vs-talview&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=tracy-vs-talview&utm_content=try-free) ## Sources - [Talview, Pricing](https://www.talview.com/en/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Home](https://tracyinterview.ai/), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. # Tracy vs TestGorilla: A Test or a Conversation? (Nov 2026) > TestGorilla vs Tracy for technical hiring: where each wins on skills tests, follow-ups, evidence, integrations, validation and price. - URL: https://tracyinterview.ai/compare/tracy-vs-testgorilla - Published: 2026-11-12 - Author: Tracy team > **In short** Tracy AI Interviewer is the stronger first round when a technical role needs follow-ups in the person's own field and a report a reviewer can check. TestGorilla is the stronger skills-test platform. > > Tracy runs a two-way video interview, can put a coding task inside it, and leaves a recording and a timestamped transcript behind every result. Candidates rate Tracy 92 on NPS. > > Choose TestGorilla when scored skills tests across many roles, published validation and ATS sync are the center of your screen. > > Next step: run one role in Tracy and compare the follow-ups and the report. This TestGorilla vs Tracy comparison is for technical teams that hire at volume. TestGorilla is a test platform with an AI interview inside it. Tracy AI Interviewer is the interview, and it follows up on each answer. ## What does TestGorilla do well? Its library holds 350+ tests across cognitive ability, communication, technical, role-specific skills, software, language and personality. TestGorilla publishes a fact sheet with reliability and validity figures for each test, and monitors adverse impact. Its AI interviews sit beside the tests: an AI interviewer asks follow-up questions within 3 to 9 structured questions, scored against a rubric with reasoning, and the recruiter can override any score. Its Trust Layer adds an honesty agreement, behavior monitoring, AI-agent blocking and optional ID verification. TestGorilla also says that pairing interviews with skills tests gives a fuller picture, and a team that trusts its tests may only need a stronger interview round. ## TestGorilla vs Tracy: buyer criteria side by side TestGorilla column: TestGorilla's own pages, read 9 Nov 2026. Tracy column: approved Tracy product records. "Not verified" and "Not stated" mean the page does not say. | Criterion | Tracy AI Interviewer | TestGorilla | Who leads | | --- | --- | --- | --- | | Interview depth | Two-way video. She reads the CV when the CV check is on, asks in the person's own field, follows up and waits for the answer. | Follow-ups within 3 to 9 structured questions, scored against a rubric you can edit. | Tracy for open-ended follow-ups. TestGorilla for a fixed rubric beside a test library. | | Tests and coding | A coding task inside the same interview, with follow-ups on approach and trade-offs. | 350+ tests, a coding IDE with code playback, and custom coding challenges on Plus. | TestGorilla for breadth. Tracy for discussing the code. | | Volume and start | Each applicant starts on click, at any hour, with no scheduling. Ready in under 10 seconds; a dropped applicant can rejoin within 10 minutes (configurable per customer). | Candidates complete AI interviews on their own time. Start and rejoin rules: Not verified. | Level on starting without a recruiter. | | Evidence and integrity | A recording of both sides, a timestamped transcript, an assessment per question, and signals (tab changes, gaze, live face detection, answer delay, fullscreen exits, external screens) shown to a human reviewer. | Scores with reasoning; a log of tab switches and full-screen exits, AI-agent blocking and ID verification on Plus. Records answers; whether a full transcript and both sides are kept: Not verified. | Level on scores with reasoning. TestGorilla on identity checks and AI-agent blocking. Tracy shows the recording of both sides and a timestamped transcript. | | ATS and stack | API and MCP connections, Slack, WhatsApp and Zapier, and use beside the Workable and Breezy HR connectors. No write-back is stated. | Greenhouse, Workable, Lever, Workday, SAP SuccessFactors and more, plus Zapier, Make and an open API. Scores sync back. Plus plan. | TestGorilla on native ATS depth. | | Support and security | Enterprise: a dedicated customer manager, a dedicated support channel, a setup customised in days. SOC 2 Type II attested. | 24/7 live chat and email on every plan. Lists SOC 2 Type II and GDPR. | TestGorilla on chat. Level on SOC 2 Type II. | | Validation, bias and accessibility | A person reviews the evidence and decides. Tracy publishes no bias audit. A spoken-language check scores fluency, grammar and vocabulary. Accessibility: Not stated. | Fact sheets with reliability and validity figures, adverse impact monitoring, scoring on the content of what a candidate says. Accessibility: Not verified. | TestGorilla on published validation. | | Reporting and price | A report per interview with an assessment per question, shareable by link or PDF. $1 per credit, first 3 results free, no seats. | Dashboards and CSV or PDF export. Free: limited AI interviews. Plus from $520 a month billed annually: full AI interview library, 2 credits per candidate. Core from $215 lists no AI interviews. | TestGorilla on dashboards. Tracy on a price per interview you can compute today. | ## How the published workflows differ for the same technical role Take one invented role, a data engineer on a reporting team. TestGorilla's workflow is an assessment that can hold a skills test, a coding test and an AI interview. The Tracy column is an illustration with a fictional candidate, not a recording of a real interview. Illustration of the published workflows for one invented role. | Step | TestGorilla (from its pages) | Tracy (illustration) | | --- | --- | --- | | Setup | Pick tests and an AI interview from the library, or write your own questions and scoring criteria. A library AI interview runs about 26 minutes. | Describe the role and Tracy drafts the questions from the job description. Edit them or bring your own. | | Second question | The AI interviewer asks follow-ups in the moment, scored against the same rubric. | She asks about what the person just said, in their field, and waits until they finish. | | Coding | A separate coding test with a built-in IDE and code playback. | A coding task in the same interview, with questions on approach and trade-offs. | | What the reviewer opens | Scores with reasoning, a log of tab switches and full-screen exits, and recorded answers. Whether a full transcript and both sides are kept: Not verified. | A recording of both sides, a timestamped transcript and an assessment per question. | ![Transcript card: Tracy asks about late rows, the candidate answers, and Tracy follows up on the re-read window.](/blog/images/tracy-vs-testgorilla-1.png) *Tracy AI Interviewer asks the follow-up the answer calls for, so a reviewer sees reasoning, not only a score. Illustration with sample data. Illustration with sample data* A follow-up shows the reasoning behind an answer that a score on its own cannot show. ## Does a test or a conversation show more? A test makes candidates comparable on a validated scale and resists look-up answers. A conversation shows how someone handles the question after the question: the first answer in the illustration would earn the same note from someone who has built a late-data pipeline and someone who has read about one, and the second shows the failure path. A Director at a top-tier early-stage VC put it this way: "Tracy sometimes brings up exactly those questions I might have forgotten to ask, which has been a pleasant surprise." For engineers, use tests where a right answer exists and an interview where the value is the discussion. Tracy runs a coding task inside the interview, as [AI technical interviews](/blog/ai-technical-interviews) explains. ## Can a reviewer check a TestGorilla or a Tracy result? Both leave a person in charge. TestGorilla's AI suggests a score with its reasoning and the recruiter can override it. Tracy's assessment covers each question and its follow-up answers, with the recording and transcript to open. Integrity works differently: TestGorilla logs tab switches and full-screen exits and blocks known AI agents, while Tracy works inside the conversation and shows her signals to a person as evidence, not proof of cheating. Nothing is decided in the interview. See [AI interview cheating detection](/blog/ai-interview-cheating-detection) and [AI interview reports](/blog/ai-interview-reports). ![Four steps: recording of both sides, timestamped transcript, assessment per question, and your team decides.](/blog/images/tracy-vs-testgorilla-2.png) *What Tracy AI Interviewer gives a reviewer: the recording, a timestamped transcript and an assessment per question, with the decision left to your team. Illustration with sample data* A Tracy result gives a reviewer the recording, a timestamped transcript and an assessment per question before the team decides. ## What does Tracy cost? A Tracy credit costs $1 and covers each started 20 minutes of an interview that produced a result, so 100 interviews of up to 20 minutes cost $100. The first 3 results are free, there are no seats, and a session with no answers is never charged. TestGorilla's AI interview uses 2 credits per candidate, and the page does not state the credits in Core or Plus. See [Tracy's pricing](/pricing) and [hiring at volume](/use-cases/volume-hiring). ## How do you test both before you commit? Run one role for two weeks. 1. Pick one technical role with at least 20 expected applicants. 2. In TestGorilla, build an assessment with the skills tests you trust and an AI interview, and ask which plan and credits it needs. 3. In Tracy, describe the role or send us the role and we build the plan. The first 3 results are free. 4. Have one reviewer read the results from your next 20 applicants and count the follow-ups that landed, the applicants who finished and the cost per interview. Send us the role and we will build the interview plan, so you can compare the two reports side by side. ## Which path fits you? - [Technical interviews](/use-cases/technical-interviews): role-specific interviews with follow-ups. - [Expert vetting](/use-cases/expert-vetting): interviews for specialists and contributors. - [Connect](/connect): the integrations, and [pricing](/pricing) for Enterprise to your volume. ## Common questions ### Does TestGorilla have an AI interviewer? Yes. It asks dynamic follow-up questions in the moment, with 3 to 9 structured questions per interview and scoring against a rubric. The full library of 100+ AI interviews and custom interviews are on its Plus plan. ### Is TestGorilla better than Tracy? For validated skills tests across many roles, yes: 350+ tests with published fact sheets, native ATS connectors and 24/7 chat support. For a first-round interview that follows up in the person's own field and leaves a recording and transcript, Tracy is the better pick. ### How does TestGorilla compare with HackerRank? HackerRank offers an AI interviewer called Chakra and a free trial. TestGorilla is a skills-test library with an AI interview, so the two start from different places. ### Is Tracy secure enough for our candidate data? Tracy is SOC 2 Type II attested. The report is available on request at hello@tracyinterview.ai. --- Run one role in Tracy and compare the follow-ups and the report: send us the role and we build the interview plan, customised in days, with a recording and report you can check. The first 3 results are free. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=tracy-vs-testgorilla&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=tracy-vs-testgorilla&utm_content=try-free) ## Sources - [TestGorilla, Pricing](https://www.testgorilla.com/pricing/), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Connect](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. # HackerEarth Alternatives for AI Technical Interviews (Nov 2026) > HackerEarth alternatives for AI technical interviews: Tracy, HackerRank, CodeSignal, CoderPad and Karat on depth, volume, stack, evidence and price. - URL: https://tracyinterview.ai/compare/hackerearth-alternatives - Published: 2026-11-11 - Author: Tracy team > **In short** HackerEarth OnScreen adds government ID checks and proctoring; Tracy AI Interviewer asks own-field follow-ups, runs a coding task in the same interview and leaves a recording and an assessment per question. > > Her price is published: $1 per credit, the first 3 results free, no seats, so 100 interviews of up to 20 minutes cost $100. > > HackerRank suits assessments and ATS depth, CodeSignal one credit across assessments and interviews, CoderPad shared pads, Karat engineer-run interviews. Stay with HackerEarth if ID checks and native ATS decide. > > Next step: create one interview for the workflow your current tool misses. HackerEarth alternatives fall into two groups: tools that interview the candidate for you, and tools that give your engineers a better room to interview in. This page covers both for a technical team that screens at volume, puts Tracy AI Interviewer first, and says where each tool beats her. [AI technical interviews](/blog/ai-technical-interviews) explains how a coding task and follow-ups fit in one interview. ## Why do teams look for HackerEarth alternatives? HackerEarth's OnScreen checks identity with a government ID, face match and live presence detection, runs proctoring through the interview, and has an avatar ask about code written in a live IDE. That suits regulated or high-fraud hiring. OnScreen plans cap invites at 10 or 50 a month, assessments and FaceCode sell as a separate subscription, and the identity step names DigiLocker for India and document checks "across other regions". ## Which HackerEarth alternative fits which team? - **Tracy AI Interviewer:** a first round that shows how each applicant reasons, with a recording and report a reviewer can check, and a price per interview. - **HackerRank:** teams that run its assessments and want its AI interviewer on every plan (as of 3 Nov 2026), with a library of up to 7,500+ questions. - **CodeSignal:** one credit across an assessment, an AI interview or a live interview, with a short structured AI screen. - **CoderPad:** engineering teams that want a shared pad, with agentic interviews on higher tiers. - **Karat:** interviews run by trained engineers, priced by quote. ## Tracy AI Interviewer: the alternative for depth and evidence Tracy interviews every applicant on video, asks in their own field, follows up on what they said and waits for the answer. Questions come from the job description and can be edited or replaced with your own, and a coding task runs inside the same interview with follow-ups on approach and trade-offs. A Director at a top-tier early-stage VC says, "Tracy sometimes brings up exactly those questions I might have forgotten to ask, which has been a pleasant surprise." After each interview you get a recording of both sides, a timestamped transcript, a written assessment per question with its reasoning, and integrity signals for a reviewer: tab changes, gaze, live face detection, delay before answers, fullscreen exits and external screens. A person decides. See [AI interview reports](/blog/ai-interview-reports). Each applicant starts on click, at any hour. Connections include API and MCP, Slack, WhatsApp and Zapier, and use beside the Workable and Breezy HR connectors. **Pick Tracy if** how a candidate reasons matters more to you than verified identity and you want to pay per interview. **Pick another if** you need government ID checks, a live panel, or a library of benchmarked coding problems. ![Recreated Tracy result report for a fictional backend candidate: the recording beside a timestamped transcript, and three questions with what the candidate said and quoted evidence.](/blog/images/hackerearth-alternatives-2.png) *Tracy shows the recording and timestamped transcript, then each question with what the candidate said and a quote, so a reviewer can check a result against the recording. Illustration with sample data* Each question sits beside what the candidate said, with a quote, next to the recording and transcript. ## HackerRank, CodeSignal, CoderPad and Karat **HackerRank** lists its AI interviewer, Chakra, on every plan (as of 3 Nov 2026). It opens a live canvas, asks follow-ups on the work and scores per competency with a reason for each score. Starter is $79 a month billed annually for 60 candidate attempts a year, Pro is $419 for 360, extra attempts are $20, and Enterprise adds 22+ enterprise ATS integrations. **CodeSignal** spends one credit when a candidate starts an assessment, an AI interview or a live interview. Its AI Interviewer is a short structured interview, typically 15 minutes, that probes thin answers. Build is $79 a month billed annually for 60 credits a year, Grow is $479 for 420, and extra credits are $20. Every plan lists SOC 2 Type II, GDPR and CCPA. **CoderPad** offers real-work, AI-assisted pads in 40+ languages. Agentic interviews are available on Frontier Lite ($1,000 a month, annual billing) and Enterprise. Starter is $80 a month annual for 60 interviews a year, and extra interviews are $30. **Karat** sells interviews run by its own Interview Engineers, available 24/7 and on demand, with a report that carries a score, written feedback per question and video and code playback. Its price is not stated; its site offers a consult and a demo request. ## How do the five compare on the four pillars? Source: each vendor's own pages, read 3 to 9 Nov 2026, and approved Tracy product records. Not stated means the page does not say. | Tool | Depth | Volume | Stack | Evidence and integrity | Price | | --- | --- | --- | --- | --- | --- | | Tracy AI Interviewer | Follow-ups in the person's own field, coding task inside the interview | Starts on click, any hour | API, MCP, Slack, WhatsApp, Zapier, Workable, Breezy HR | Recording, transcript, assessment per question, signals for a reviewer | $1 per credit, first 3 free | | HackerRank | Chakra canvas with follow-ups on the work | 60 or 360 attempts a year | 22+ enterprise ATS, API and MCP on Enterprise | Scores per competency, flags for unauthorized apps | From $79 a month billed annually | | CodeSignal | Short AI interview that probes thin answers | 60 or 420 credits a year | Gem and Ashby; Workday, iCIMS, Oracle on enterprise | Suspicion Score, ID verification, scored report | From $79 a month billed annually | | CoderPad | Real-work pads; agentic interviews on higher tiers | 60 or 360 interviews a year | Not stated | AI summaries and transcripts on Frontier Lite | From $80 a month annual | | Karat | Interviews run by Interview Engineers | 24/7 on demand | Not stated | Score, written feedback, video and code playback | Not stated | ![Bar chart of first-year cost for 100 interviews: Tracy $300, OnScreen $1,290, HackerRank $1,748, CodeSignal $1,748, CoderPad $2,160.](/blog/images/hackerearth-alternatives-1.png) *Tracy's cost for 100 interviews a year beside each tool's starting plan plus extra attempts at the published rate, so you can size the budget. Karat: price not stated. Our arithmetic from published plan prices, Nov 2026. Source: vendors' published plan prices, Nov 2026.* At 100 interviews a year, Tracy at 45 minutes each costs a fraction of the other tools' starting plans plus extra attempts. ## What would 100 interviews a year cost on each? Our arithmetic from starting plans, assuming one attempt or credit per interview and no assessments. Tracy: $100 for interviews of up to 20 minutes, $300 at 45 minutes. HackerEarth OnScreen Growth: $1,290 for 120 invites. HackerRank Starter: $948 for 60 attempts plus 40 at $20, so $1,748. CodeSignal Build: the same $1,748. CoderPad Starter: $960 plus 40 at $30, so $2,160. Karat: price not stated. For the wider picture see [AI interview cost](/blog/ai-interview-cost). ## How do you choose between them? 1. **What must the first round show?** If how a person reasons, ask each vendor for a live run on your own role and count the follow-ups. If a verified identity above all, HackerEarth leads. 2. **Who runs it, and in what unit do you pay?** Convert credits, attempts or invites to the cost of one finished interview at your volume. 3. **Where do results go?** Tracy works beside the Workable and Breezy HR connectors; HackerRank, CodeSignal and HackerEarth publish deeper ATS lists. 4. **Security and support?** Tracy is SOC 2 Type II attested with the report on request; CodeSignal and HackerEarth list SOC 2 Type II, and HackerEarth lists five ISO standards too. Tracy publishes no bias audit; accessibility is Not stated. See [AI interview cheating detection](/blog/ai-interview-cheating-detection) for weighing integrity signals. Which path fits you: - Engineers and other technical roles: [technical interviews](/use-cases/technical-interviews). - Hundreds or thousands of applicants: [volume hiring](/use-cases/volume-hiring). - Connections and plans: [Connect](/connect) and [Tracy's pricing](/pricing). To see the difference on your own role, create one interview for the workflow your current tool misses: send us the role and we build it. ## Common questions ### What are the best HackerEarth alternatives for AI technical interviews? Tracy AI Interviewer for follow-ups and evidence, HackerRank with Chakra, CodeSignal for one credit across assessments and interviews, CoderPad for pads, and Karat for engineer-run interviews. ### Who are HackerEarth's competitors? HackerRank, CodeSignal, CoderPad and Karat compete for technical hiring, and Tracy competes for the AI-run first interview. Compare the cost of one finished interview, since each prices a different unit. ### Is there a cheaper alternative to HackerEarth OnScreen? Tracy's usage is $1 per credit, so 100 interviews of up to 20 minutes cost $100 against $1,290 for OnScreen Growth at its yearly price. HackerRank and CodeSignal start at $79 a month billed annually. ### Which alternative has the deepest follow-up questions? Ask each vendor to run your own role. Tracy asks follow-ups in the person's own field and waits for the answer, with a coding task in the same interview. ### Does Tracy replace HackerEarth's identity checks? Tracy shows integrity signals with evidence to a reviewer. Identity verification against a government ID is Not stated, so keep HackerEarth if that check decides your purchase. --- Create one interview for the workflow your current tool misses: send us one role and we build the plan, customised in days, with a recording and report you can check. The first 3 results are free, then $1 per credit. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=hackerearth-alternatives&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=hackerearth-alternatives&utm_content=try-free) ## Sources - [HackerEarth, OnScreen pricing](https://www.hackerearth.com/ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer, home](https://tracyinterview.ai/), retrieved 9 Nov 2026. # HackerEarth AI Interview Agent Pricing (Nov 2026) > HackerEarth pricing for its AI interviewer: OnScreen from $129 a month for 10 invites. Cost at 100, 500 and 1,000 interviews against Tracy's $1 per credit. - URL: https://tracyinterview.ai/compare/hackerearth-pricing - Published: 2026-11-11 - Author: Tracy team > **In short** Tracy AI Interviewer prices by completed interview: $1 per credit, one credit per started 20 minutes, so a 45-minute technical interview is $3. The first 3 results are free, with no seats. > > Each interview that produces a result costs credits and comes with own-field follow-ups, a recording and an assessment per question. > > At 100 interviews a year Tracy costs $300 against $1,290 for HackerEarth OnScreen; at 500, $1,500 against $4,990. HackerEarth wins when ID checks, proctoring and native ATS decide. > > Next step: try three real interviews free, then calculate your own cost. HackerEarth pricing for its AI interviewer is public. This page lists the plans, shows what 100, 500 and 1,000 interviews cost, and sets them beside Tracy AI Interviewer, which charges for each interview that produced a result. See [Tracy's pricing](/pricing) for the full price list. ## What does HackerEarth pricing look like for OnScreen? HackerEarth's agents page lists its AI Interview Agent at $20 an hour and does not define the hour; the OnScreen plans below are the ones on its pricing page. Plans are sold monthly or yearly by the number of interview invites. Growth adds a 30-day free trial, recorded interviews with playback and multi-language support. Scale adds custom roles and permissions, single sign on, premium support and realtime code evaluation. Enterprise adds custom question development, a dedicated account manager, enterprise security and SLAs. HackerEarth says pricing scales with the interview credits consumed, so the invite allowance is the number to size against your volume. Source: HackerEarth OnScreen pricing page, read 9 Nov 2026. | Plan | Monthly | Yearly | Includes | | --- | --- | --- | --- | | Growth | $129 a month, 10 invites | $1,290 a year, 120 invites | 30-day trial, AI technical interviews, recording and playback | | Scale | $499 a month, 50 invites | $4,990 a year, 600 invites | Growth plus roles, SSO, realtime code evaluation, premium support | | Enterprise | Custom | Custom | Scale plus custom questions, account manager, SLAs | ## What costs extra on HackerEarth? Coding assessments and FaceCode live panels are a separate subscription, at $99 a month on Growth and $399 on Scale, and the pricing page suggests buying the two together. SSO and realtime code evaluation start on Scale. You can cancel by emailing support, but amounts already charged are non-refundable. ## What does HackerEarth cost at 100, 500 and 1,000 interviews? This is our arithmetic from the plan prices, not a HackerEarth quote. The Tracy columns show 20 and 45 minutes. Tracy charges one credit for each started 20 minutes, so 45 minutes is 3 credits. At full use OnScreen works out to $10.75 per invite on Growth and $8.32 on Scale. HackerEarth's pricing page sells plans by invite allowance, while its FAQ says you pay for interviews candidates take, so these per-invite figures are the most it can cost per interview. HackerEarth figures from its OnScreen pricing page, read 9 Nov 2026. Tracy figures are $1 per credit before the first 3 free results. Bring your yearly volume and interview length and we will work out your own cost on a demo call. | Interviews in a year | OnScreen: cheapest yearly plan that covers it | Tracy, up to 20 minutes | Tracy, 41 to 60 minutes | | --- | --- | --- | --- | | 100 | Growth, $1,290 (120 invites) | $100 | $300 | | 500 | Scale, $4,990 (600 invites) | $500 | $1,500 | | 1,000 | Scale allows 600 invites a year, so 1,000 is an Enterprise quote. | $1,000 | $3,000 | ![Bar chart: OnScreen $1,290 and $4,990 against Tracy $300 and $1,500 for 100 and 500 interviews.](/blog/images/hackerearth-pricing-1.png) *Tracy's usage cost at 100 and 500 interviews of 45 minutes, beside the cheapest OnScreen yearly plan that covers each volume, so you can size the budget before you start. Plan prices, not a quote. Source: HackerEarth OnScreen pricing page and Tracy pricing page, Nov 2026.* At 100 and 500 interviews a year, Tracy's usage cost is a fraction of the cheapest OnScreen yearly plan that covers the volume. ## How does Tracy's pricing suit volume? Tracy has no seats, no plans to climb and no upfront cost. A credit is $1 and covers each started 20 minutes of an interview that produced a result. Credits never expire, and a session with no answers is never charged. Each applicant starts on a click, at any hour, with no scheduling, and a dropped applicant can rejoin within 10 minutes. One customer cut its average time to hire from 44 days to 7. See [hiring at volume](/use-cases/volume-hiring). ![Recreated Tracy interviews list: platform engineer interviews, each row with a status, the answers given and a View result link.](/blog/images/hackerearth-pricing-2.png) *Each finished interview is one row you can open. Illustration with sample data* Each finished interview is one row with its status, the answers given and a View result link. ## How do HackerEarth and Tracy compare beyond price? Rival column: HackerEarth's own pages, read 9 Nov 2026. Tracy column: approved Tracy product records. Not stated means the page does not say. | Criterion | Tracy AI Interviewer | HackerEarth OnScreen | Who leads | | --- | --- | --- | --- | | Depth | Follow-ups in the person's own field, a coding task inside the interview, questions from the job description or your own list. | Adaptive follow-ups in a fixed framework, with code written in a live IDE. | Level on coding; Tracy on own-field questions. | | Volume | Starts on click at any hour; ready in under 10 seconds. | Always on; plans cap invites per month. | Tracy, with no invite allowance to outgrow. | | Stack | API and MCP (ChatGPT, Claude, Mistral, Gemini, Copilot Studio), Slack, WhatsApp and Zapier, and use beside the Workable and Breezy HR connectors. | Native Workday, Greenhouse, Lever and Ashby connectors, plus an API. | HackerEarth on ATS depth. | | Evidence | Recording of both sides, timestamped transcript, assessment per question with reasoning, integrity signals for a reviewer. A person decides. | Scores on your rubric, transcript, code, integrity signals. The team decides. | Level. | | Identity and proctoring | Signals only: tab changes, gaze, live face detection, answer delays, fullscreen exits, external screens. Government ID check: Not stated. | Government ID, face match, live presence, proctoring stack. | HackerEarth. | | Security and bias controls | SOC 2 Type II attested, report on request at hello@tracyinterview.ai. No bias audit published. | Lists SOC 2 Type II and five ISO standards; says it runs adverse-impact studies. | HackerEarth documents more. | | Support | Enterprise: dedicated customer manager, dedicated support channel, setup customised in days. | Premium support on Scale; account manager and SLAs on Enterprise. | Level. | ## Is HackerEarth worth its price, and for whom? If you need government ID checks and a full proctoring stack on every interview, native ATS connectors, and one vendor for assessments and live panels, HackerEarth is built for it, and its 30-day trial lets you test first. If you want the first round to show how each applicant reasons and the cost to follow the number of interviews, at the plan prices above Tracy costs less at 100 and 500 interviews a year, for interviews up to 60 minutes: the first 3 results are free, there is no seat to buy, and the next 100 interviews of up to 20 minutes cost $100. Enterprise is priced to how many interviews you run. Which path fits you: - Engineers and other technical roles: [technical interviews](/use-cases/technical-interviews). - Hundreds or thousands of applicants: [volume hiring](/use-cases/volume-hiring). - A first-round budget: [AI interview cost](/blog/ai-interview-cost). ## What should you ask HackerEarth before you buy? 1. What is the first-year total at our volume and at twice our volume, with SSO and realtime code evaluation? 2. What happens to unused invites at the end of a month or year? 3. What does the eleventh interview on Growth cost, and when do we move to Enterprise? 4. Is the assessment subscription needed to run OnScreen? 5. Which ATS connector is on our plan, and does it write results back? Tracy answers the money questions on her pricing page, so a demo call can go to fit. For other vendors, see [HireVue pricing](/compare/hirevue-pricing). ## Common questions ### How much does HackerEarth cost? OnScreen is $129 a month for 10 interview invites and $499 a month for 50, or $1,290 and $4,990 a year, and Enterprise is custom. Coding assessments are a separate subscription from $99 a month. ### Does HackerEarth have a free trial? Its Growth plan lists 30 days free. Tracy's first 3 results are free, whatever their length, and no card is needed. ### Is Tracy cheaper than HackerEarth? At the published plan prices, yes at 100 and 500 interviews a year: $300 and $1,500 for 45-minute interviews against $1,290 and $4,990. Scale allows 600 invites a year, so 1,000 is an Enterprise quote. ### Does Tracy charge per seat or per month? No. There are no seats, no subscription and no upfront cost, and a session with no answers is never charged. ### What does Tracy Enterprise cost? Enterprise is priced to how many interviews you run, with a dedicated customer manager and a setup customised to your workflow. Book a demo for a figure. --- Try three real interviews and calculate your completed-interview cost: bring your yearly volume, interview length and stack, and we build the plan, customised in days, with a recording and report you can check. The first 3 results are free, then $1 per credit. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=hackerearth-pricing&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=hackerearth-pricing&utm_content=try-free) ## Sources - [HackerEarth, OnScreen pricing](https://www.hackerearth.com/ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. # HackerRank Pricing: Cost vs Tracy (Nov 2026) > HackerRank pricing as of 3 Nov 2026: Starter $79 a month, Pro $419, extra attempts $20. Cost at 100, 500 and 1,000 interviews against Tracy. - URL: https://tracyinterview.ai/compare/hackerrank-pricing - Published: 2026-11-11 - Author: Tracy team > **In short** Tracy AI Interviewer charges $1 per credit, one credit per started 20 minutes, so a 45-minute technical interview is $3. The first 3 results are free, with no seats. > > She follows up on approach and trade-offs inside the coding task and leaves a recording and an assessment per question. > > At 100 interviews a year Tracy costs $300 against $1,748 on HackerRank's listed plans, which also include tests and a library. HackerRank is the better pick for timed coding tests and its question library. > > Next step: try three real interviews free, then work out your own cost. HackerRank pricing is public: Starter is $79 a month billed annually and Pro is $419 a month billed annually, as listed on 3 Nov 2026. This page prices a technical screen at 100, 500 and 1,000 candidates and sets it beside Tracy AI Interviewer, which charges for each interview that produced a result. See [Tracy's pricing](/pricing) for the full list. Choose HackerRank when: - You want timed coding tests from a large question library, with live pair programming. - An annual plan with ATS and calendar integrations covers your volume. Choose Tracy when: - You want the first round to be a conversation that follows up on how the candidate reasoned, in their own field. - Your volume swings and you would rather pay for each interview than for a yearly allowance of attempts. ## What does HackerRank pricing look like? HackerRank sells Starter and Pro by the year, with unlimited users and a fixed number of candidate attempts. Starter is $948 a year for 60 attempts and Pro is $5,028 a year for 360. Pro adds ATS and calendar integrations and a larger library, 4,000+ questions against 2,000+. Enterprise is custom, with custom users and attempts and 7,500+ questions. Billed monthly, the same prices carry 5 attempts a month on Starter and 30 on Pro. Chakra, HackerRank's AI interviewer, and real-time follow-ups are listed as included on every plan. Source: HackerRank pricing page, as of 3 Nov 2026. | Plan | Billed annually | Attempts | Adds | | --- | --- | --- | --- | | Starter | $79 a month, $948 a year | 60 a year | Unlimited users, 2,000+ questions | | Pro | $419 a month, $5,028 a year | 360 a year | ATS and calendar integrations, 4,000+ questions | | Enterprise | Custom | Custom | 7,500+ questions, 22+ enterprise ATS integrations | ## What costs extra on HackerRank? Extra attempts cost $20 each on Starter and Pro. Annual Pro customers can pre-purchase extra attempts at $15 each. Enterprise has volume pricing, with no figure listed. The plan price covers more than interviews: it includes take-home assessments, pair programming interviews, proctoring features, an identity match between screen and interview, and the library. ## What does HackerRank cost at 100, 500 and 1,000 interviews? This is our arithmetic from the listed prices, not a HackerRank quote. Starter plus 40 extra attempts at $20 covers 100 attempts. Annual Pro plus pre-purchased attempts at $15 covers 500 and 1,000. Tracy charges one credit for each started 20 minutes, so a 45-minute interview is 3 credits. HackerRank counts a candidate attempt, which can be a test or an interview, so its figure buys more than the interview alone; the comparison is cost per candidate screened, not like for like. HackerRank figures from its pricing page, as of 3 Nov 2026. Tracy figures are $1 per credit before the first 3 free results. | Interviews in a year | HackerRank, cheapest listed route | Tracy, up to 20 minutes | Tracy, 41 to 60 minutes | | --- | --- | --- | --- | | 100 | Starter plus 40 extra attempts, $1,748 | $100 | $300 | | 500 | Pro plus 140 pre-purchased attempts, $7,128 | $500 | $1,500 | | 1,000 | Pro plus 640 pre-purchased attempts, $14,628. Enterprise volume pricing: no figure listed. | $1,000 | $3,000 | ![Bar chart: HackerRank $1,748, $7,128, $14,628 against Tracy $300, $1,500, $3,000 at 100, 500, 1,000 interviews.](/blog/images/hackerrank-pricing-1.png) *Tracy's usage cost for 45-minute interviews beside the cheapest listed HackerRank route at each volume, so you can size a technical screening budget. Plan prices, not a quote. Source: HackerRank pricing page (as of 3 Nov 2026) and Tracy pricing page, Nov 2026.* At 100, 500 and 1,000 interviews a year, Tracy's usage cost is a fraction of HackerRank's cheapest listed route. ## What does the interview itself look like? The price per interview is the first difference; the second is what the interview does. Tracy asks questions in the person's own field, follows up on what they said and waits for the answer. Inside a coding task she follows up on approach and trade-offs: why this structure, and what breaks when the input grows. Questions come from the job description or your own list, and you set the skills and their weights per role. HackerRank's Chakra also asks follow-ups: it opens a live canvas like a dev environment and asks about the candidate's work. Both do more than score a test. See [how Tracy works](/product). ![Sample interview: Tracy asks why a hash map, the candidate answers, Tracy asks what breaks at scale.](/blog/images/hackerrank-pricing-2.png) *A follow-up chain inside a coding task: Tracy asks about a trade-off the candidate has just made, so the reviewer sees reasoning, not only a result. Illustration with sample data* Tracy answers a coding choice with a follow-up on its trade-off. ## How do HackerRank and Tracy compare beyond price? HackerRank column from its pricing page, as of 3 Nov 2026. | Criterion | Tracy AI Interviewer | HackerRank | Who leads | | --- | --- | --- | --- | | Depth | Follow-ups in the person's own field, a coding task inside the interview, questions from the job description or your own. | Live canvas with follow-ups on the work, pair programming interviews, 4,000+ library questions on Pro. | HackerRank on coding environment and library; Tracy on own-field questions. | | Volume | Starts on click at any hour, no scheduling; ready in under 10 seconds. | Priced by attempts a year; 60 on Starter, 360 on Pro. | Tracy: no allowance to outgrow. | | Stack | API and MCP (ChatGPT, Claude, Mistral, Gemini, Copilot Studio), Slack, WhatsApp and Zapier, and use beside Workable and Breezy HR. | ATS and calendar integrations from Pro; 22+ enterprise ATS integrations on Enterprise. | HackerRank on ATS depth. | | Evidence | Recording of both sides, timestamped transcript, assessment per question with reasoning. A person decides. | Report with transcript, session replay and scores. | Level. | | Integrity | Signals for a human reviewer to interpret, with their evidence: tab changes, answer delays, fullscreen exits, external screens. Applicants are told on the join screen that the interview is recorded. No automated decision. ID check: Not stated. | Advanced proctoring features and a screen to interview identity match. | HackerRank. | | Security | SOC 2 Type II attested; report on request. No bias audit published. | SOC 2, ISO 27001 and GDPR listed as included. | HackerRank on listed standards. | | Price | $1 per credit; first 3 results free; no seats. | From $948 a year; extra attempts $20 each. | Tracy at the volumes above. | ## Is HackerRank worth its price, and for whom? If you run timed coding tests from a large library, want live pair programming, and need ATS connectors and proctoring on every candidate, HackerRank is built for it. If you want the first round to show how each applicant reasons, Tracy fits, at a cost that follows the number of interviews. At the volumes above, Tracy's interview costs less than HackerRank's cheapest listed route, though HackerRank's price also buys tests, a library and pair programming: the first 3 results are free, there is no seat to buy, and the next 100 interviews of up to 20 minutes cost $100. Which path fits you: - Engineers and other technical roles: [technical interviews](/use-cases/technical-interviews). - Hundreds or thousands of applicants: [volume hiring](/use-cases/volume-hiring). - A first-round budget: [AI interview cost](/blog/ai-interview-cost). ## What should you ask HackerRank before you buy? 1. What is the first-year total at our volume and at twice our volume, with extra attempts? 2. What happens to unused attempts at the end of the year? 3. Does a Chakra interview use one attempt, the same as a test? 4. What does Enterprise volume pricing come to at our volume? 5. Which ATS connector is on our plan, and does it write results back? Tracy answers the money questions on her pricing page. ## Common questions ### How much does HackerRank cost? As of 3 Nov 2026, Starter is $79 a month billed annually ($948) for 60 attempts a year and Pro is $419 a month billed annually ($5,028) for 360. Extra attempts are $20 each. Enterprise is custom. ### How much does HackerRank cost for companies? The plans above are the company plans. Enterprise, for custom users and attempts, lists no price. Tracy lists hers: $1 per credit, no seats. ### Does HackerRank have a free trial? Its Chakra page offers a free trial. Tracy's first 3 results are free and no card is needed. ### Is Tracy cheaper than HackerRank? At the listed prices, yes at 100, 500 and 1,000 interviews a year: $300, $1,500 and $3,000 for 45-minute interviews against $1,748, $7,128 and $14,628. HackerRank's price also buys tests and a library. ### Does Tracy charge per seat or per month? No. There are no seats, no subscription and no upfront cost, and a session with no answers is never charged. --- Try three real interviews free. Send us one technical role and we build the interview plan, customised in days, with a recording and report you can check. Then $1 per credit. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=hackerrank-pricing&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=hackerrank-pricing&utm_content=try-free) ## Sources - [HackerRank, pricing](https://www.hackerrank.com/pricing), retrieved 3 Nov 2026. - [Tracy AI Interviewer, Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. # LeetCode Pricing for Companies vs Tracy (Nov 2026) > LeetCode pricing for companies is not stated on its Premium page. What it lists, how to estimate a cost, and Tracy cost at 100 to 1,000 interviews. - URL: https://tracyinterview.ai/compare/leetcode-pricing-for-companies - Published: 2026-11-11 - Author: Tracy team > **In short** Tracy AI Interviewer runs the technical first round at $1 per credit, one credit per started 20 minutes: $3 for a 45-minute interview. The first 3 results are free, with no seats and no upfront cost. > > She follows up on approach and trade-offs inside the coding task and leaves a recording and an assessment per question. > > LeetCode's Premium page lists $35 a month or $159 a year for individuals and no company price. LeetCode is the better pick for practice and its library of about 4,100 problems. > > Next step: try three real interviews free. LeetCode does not state a company price on its Premium page, read on 9 Nov 2026. What the page lists is Premium for individuals: $35 a month or $159 a year. This page is for the engineering manager who asked what it costs to use LeetCode to screen candidates. It sets out what LeetCode lists, how to estimate a company cost, what to ask on a quote call, and what the same screen costs on Tracy AI Interviewer, which publishes its price. See [Tracy's pricing](/pricing) for the full list. Choose LeetCode when: - You or your candidates want a very large bank of coding problems to practise on. - You are an individual preparing for interviews, not a team screening applicants. Choose Tracy when: - You want the first technical round run for you, with follow-ups on how the candidate reasoned. - You want a price you can calculate before talking to anyone. ## What does LeetCode pricing for companies look like? LeetCode's Premium page lists two plans. Monthly is $35 a month, shown against $39, billed monthly. Yearly is $159, shown against $299, billed annually, which the page works out to $13.25 a month. The page carried a Black Friday sale banner on the day it was read, so list prices may differ outside a sale. Both plans list the same things: premium problems and solutions, company-specific interview questions, a coding agent that brainstorms solutions and debugs, and mock assessments. Source: LeetCode Premium page, read 9 Nov 2026. | Plan | Price | Billing | Built for | | --- | --- | --- | --- | | Premium monthly | $35 a month | Monthly | Interview practice, per the page | | Premium yearly | $159 a year | Annually | Interview practice, per the page | | Company plan | Not stated | Not stated | Not stated | ## How do you estimate what LeetCode would cost your team? The Premium page describes interview practice for candidates: mock assessments where you choose a company and get a timed question, and company-specific questions to prepare with. Whether Premium may be used to screen applicants, or whether LeetCode sells an employer plan, is not stated on the Premium page. This part is reasoning, not a LeetCode price. A subscription priced per account suits one person. If a team screens 500 candidates, the question is whether each candidate or each interviewer needs an account, and what happens to cost when volume doubles. Low volume under a per-seat or contract model usually costs more per interview than usage pricing. ## What should you ask on a quote call? If you ask LeetCode for a quote: 1. Is there an employer plan, and may we use Premium to screen applicants? 2. Who needs an account: each candidate, each interviewer, or only us? 3. What is the first-year total at our volume and at twice our volume? 4. Do unused invitations or seats roll over? 5. Which ATS connectors, reports and integrity checks come with it? Other vendors state their prices. CodeSignal lists Build at $79 a month billed annually with 60 credits (9 Nov), and HackerRank lists Starter at $79 a month billed annually with 60 candidate attempts (3 Nov). Extra credits and attempts are $20 each on both, so on their entry plans 100 candidates comes to $1,748 by our arithmetic. ## What does Tracy cost at 100, 500 and 1,000 interviews? Tracy charges one credit, $1, for each started 20 minutes of an interview that produced a result. A 45-minute interview is 3 credits. Credits never expire, there are no seats, and the first 3 results are free. This is our arithmetic from that rule, not a comparison with a LeetCode figure, because LeetCode lists none for companies. Tracy figures are $1 per credit, one credit per started 20 minutes, before the first 3 free results. | Interviews in a year | Up to 20 minutes | 21 to 40 minutes | 41 to 60 minutes | | --- | --- | --- | --- | | 100 | $100 | $200 | $300 | | 500 | $500 | $1,000 | $1,500 | | 1,000 | $1,000 | $2,000 | $3,000 | ![Bar chart of Tracy cost: $100, $500, $1,000 for 20-minute interviews and $300, $1,500, $3,000 for 45 minutes.](/blog/images/leetcode-pricing-for-companies-1.png) *Tracy's usage cost at 100, 500 and 1,000 interviews, for interviews up to 20 minutes and of 45 minutes, so you can budget a technical first round before talking to a vendor. Pricing rule, not a quote. Source: Tracy pricing page, Nov 2026.* A 45-minute technical interview costs $3 on Tracy, so 1,000 of them cost $3,000 with no seats. ## How does Tracy run the first technical round? A bank of problems tests whether a candidate can solve one. Tracy's interview asks how they got there. She reads the CV when the CV check is on, asks questions in the person's own field, follows up on what they said and waits for the answer. Inside a coding task she asks about approach and trade-offs. Each applicant starts on a click, at any hour, with no scheduling, and a dropped applicant can rejoin within 10 minutes. See [technical interviews](/use-cases/technical-interviews). ![Sample interview: Tracy asks about memory cost, the candidate answers, Tracy asks what changes at ten million rows.](/blog/images/leetcode-pricing-for-companies-2.png) *A follow-up chain inside a coding task: Tracy asks what a candidate's solution costs and what changes at scale, so the reviewer sees reasoning. Illustration with sample data* Tracy follows a working solution with a question about its cost. ## How do LeetCode and Tracy compare for employers? LeetCode column from its Premium page, read 9 Nov 2026. Not stated means the Premium page does not say. | Criterion | Tracy AI Interviewer | LeetCode | Who leads | | --- | --- | --- | --- | | Question library | Questions generated from the job description, editable, or your own list. | About 4,100 problems, by the count LeetCode shows on its site on 9 Nov 2026, with company-specific questions on Premium. | LeetCode. | | Interview with follow-ups | Follow-ups in the person's own field and on coding trade-offs. | Mock assessments with a timed question, for candidates. | Tracy, for screening applicants. | | Volume | Starts on click at any hour; no scheduling; ready in under 10 seconds. | Employer volume handling: not stated on the Premium page. | Not compared: LeetCode's Premium page is silent. | | Evidence | Recording of both sides, timestamped transcript, assessment per question. A person decides. | Employer reporting: not stated on the Premium page. | Not compared: LeetCode's Premium page is silent. | | Integrity | Signals for a human reviewer to interpret, with their evidence: tab changes, answer delays, fullscreen exits, external screens. Applicants are told on the join screen that the interview is recorded. No automated decision. | Not stated on the Premium page. | Not compared: LeetCode's Premium page is silent. | | Stack | API and MCP, Slack, WhatsApp and Zapier, and use beside Workable and Breezy HR. | Employer integrations: not stated on the Premium page. | Not compared: LeetCode's Premium page is silent. | | Price | $1 per credit; first 3 results free; no seats. | Individual plans $35 a month or $159 a year; no company price on the Premium page. | Tracy, for a price you can calculate. | ## Is LeetCode worth it for hiring, and for whom? LeetCode is the right tool for practice, and its library is larger than anything you would write yourself. If your job is to screen applicants, you need an interview that runs without you, at any hour, and leaves evidence a person can check. Tracy does that for a price you can calculate: the first 3 results are free, there is no seat to buy, and 100 interviews of up to 20 minutes cost $100. Enterprise is priced to how many interviews you run. Which path fits you: - Engineers and other technical roles: [technical interviews](/use-cases/technical-interviews). - Hundreds or thousands of applicants: [volume hiring](/use-cases/volume-hiring). - A first-round budget: [AI interview cost](/blog/ai-interview-cost). ## Common questions ### How much does LeetCode cost for companies? The Premium page lists no company price; it lists Premium for individuals at $35 a month or $159 a year. Ask LeetCode whether an employer plan exists. ### Is LeetCode free for companies? Not stated on the Premium page, which lists only paid Premium for individuals. Tracy's first 3 results are free, then $1 per credit. ### What are LeetCode alternatives for hiring teams? Assessment platforms such as CodeSignal and HackerRank publish plan prices and run timed coding tests. Tracy runs the interview itself, with follow-ups on trade-offs, priced per completed interview. ### How much does Tracy cost for 500 technical interviews? $500 if they run up to 20 minutes and $1,500 if they run 45 minutes, before the first 3 free results. ### Does Tracy charge per seat or per month? No. There are no seats, no subscription and no upfront cost, and a session with no answers is never charged. --- Try three real interviews free. Send us one technical role and we build the interview plan, customised in days, with a recording and report you can check. Then $1 per credit. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=leetcode-pricing-for-companies&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=leetcode-pricing-for-companies&utm_content=try-free) ## Sources - [LeetCode, Premium](https://leetcode.com/subscribe/), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. # Mercor Reviews: What Experts Say and What Buyers Should Take From Them (Nov 2026) > Mercor reviews read for a buyer: 8 public Trustpilot reviews with dates and quotes, where Mercor fits, and how to vet your own experts instead. - URL: https://tracyinterview.ai/compare/mercor-reviews - Published: 2026-11-11 - Author: Tracy team > **In short** Tracy AI Interviewer interviews the experts you already have, in their own field, and leaves a recording and a timestamped transcript behind every result. > > Mercor's public Trustpilot page holds 8 reviews, 7 of them one-star and dated July to November 2026, mostly written by experts rather than buyers. > > Choose Mercor when you want vetted people supplied to you. > > Tracy's price is public: the first 3 results free, then $1 per credit, no seats. Mercor reviews are worth reading, but they answer a different question from the one a lab operations lead is asking. Nearly every public review of Mercor comes from someone working on the platform, not from a company that bought experts through it. If you are deciding whether to buy experts from Mercor or to [interview your own experts](/use-cases/expert-vetting), the reviews tell you how the supply side feels, and little about outcomes for buyers. Mercor is the better pick when you need a large pool of vetted professionals placed with your team quickly. [Tracy AI Interviewer](/product) is the better pick when the experts are already your applicants, referrals or contractors and you need each one interviewed in depth, with evidence you can check. ## What is Mercor, and who writes its reviews? Mercor says it offers "5M+ experts: physicians, lawyers, engineers, consultants" and describes its enterprise work as "Human evaluation at the scale frontier labs rely on." For buyers, its enterprise page says expert professionals are "found, vetted, and placed inside your team in days." Experts apply once to the Mercor talent network, take AI assessments and then receive offers. That shapes the reviews. The public reviews describe the expert's side of that arrangement: accepting an offer, doing a project, getting paid. ## Where do the counts come from? The corpus is Trustpilot's page for mercor.com: 8 reviews dated May 2024 to November 2026, with a TrustScore of 2.2. Seven are one-star and all seven are dated July to November 2026. One is five-star, from 2024, and reads as written by a customer's team. Counts below come from the opening line of each review. A review can raise more than one issue, and 8 is a very small sample. ## What do reviewers praise about Mercor? Only one of the eight is positive, and it dates from May 2024. It says "Mercor consistently followed up with our team to make sure we were having a good experience." (Trustpilot, May 2024). The wording points to a company that bought from Mercor, so it is the one review that speaks to the buyer's side. One review says nothing about how often this happens. Mercor's own pages make the case for the supply: its experts page lists an average contracted rate and open roles, and its enterprise page promises placement in days. Those are claims by the vendor, not reviews. ## What do reviewers complain about? Counts are out of the 8 Trustpilot reviews, with one quote each. Counts: opening line of each of the 8 Trustpilot reviews of mercor.com. A review can raise more than one issue. Quotes are word for word. | Issue raised | Reviews (of 8) | One quote | Reviewer and date | | --- | --- | --- | --- | | Project paused or ended | 1 | "Accepted an offer, bought the required computer, then the project was paused" | Trustpilot reviewer, Nov 2026 | | Trial work not yet paid | 1 | "I did some trial work for this company with no pay sent yet." | Trustpilot reviewer, Nov 2026 | | General disappointment or a scam label | 3 | "My experience working on a project through Mercor was extremely disappointing." | Trustpilot reviewer, Aug 2026 | | A problem with talent-facing operations | 1 | "...highlighted a significant issue within the platform’s talent‑facing operations." | Trustpilot reviewer, Jul 2026 | | Warning about an imitation site | 1 | "fake version of mercor.io" | Trustpilot reviewer, Oct 2026 | ![Bar chart of Mercor Trustpilot reviews: one five-star, no 2 to 4 stars, seven one-star.](/blog/images/mercor-reviews-1.png) *Reviews of Mercor by star rating on Trustpilot. Eight is a small sample, and the reviewers are mostly experts. Trustpilot, 8 reviews of mercor.com, May 2024 to Nov 2026.* Seven of the eight Trustpilot reviews are one-star, all from July to November 2026. ## What do the Mercor reviews mean for a buyer? The repeated theme is projects that stop: two reviews describe work that stopped or went unpaid. These reviewers are people doing work for Mercor's customers. If you buy experts from a marketplace, the project can pause or end on the client side, so ask how a pause is handled for the expert and for you. Eight reviews cannot tell you how often projects pause, or whether buyers see the same thing. Public buyer-side reviews of Mercor are thin, so a reference call with an existing customer will tell you more. ## Mercor or Tracy: which fits? Mercor wins on reach. A network the vendor puts at 5M+ experts, placement in days and a research arm built around human evaluation is a different offer from an interview tool. If you have no applicants and need people, a marketplace is the faster route. Tracy starts from the other end. You bring the applicants, from your own site, referrals or an existing pool, and Tracy interviews each one in their own field, follows up on what they said and waits for the answer. The reviewer gets a recording, a timestamped transcript and a written assessment per question, so a person decides from evidence. Nobody joins a marketplace, and nobody leaves your funnel. The table sets the two side by side. To see it on your own applicants, [try three real interviews free](/connect). Mercor column: its own pages, read 9 Nov 2026. Tracy column: Tracy's published capabilities. | | Mercor, as it describes itself | Tracy | | --- | --- | --- | | What you get | Vetted experts found and placed with your team (vendor claim) | An interview of each applicant you already have | | Who is vetted | Experts who joined the Mercor network and took its AI assessments | Your own applicants, referrals or contractors, in their own field | | Evidence you can check | Not stated on the pages read | A recording and timestamped transcript behind every result, plus a written assessment per question | | Price | No buyer price on the pages read; contact form | $1 per credit, one credit per started 20 minutes, first 3 results free, no seats | | How you start | Talk to the team | Send a browser link, no account, any hour | ![Tracy asks a fictional expert applicant what would make them hold off, then what a repeat review would have to show.](/blog/images/mercor-reviews-2.png) *Tracy asks a follow-up on what an expert said, so you see their reasoning before you rely on their field. Illustration with sample data* A follow-up built on the applicant's own answer shows how an expert reasons, which a profile does not. ## What does Mercor cost, in brief? Mercor publishes no price for buyers on the pages read, only a contact form and two ways of working: staffing and per-task pricing. Tracy publishes its price: $1 per credit, one credit for each started 20 minutes, the first 3 results free, no seats and no upfront cost. At 20 minutes each, 100 interviews cost $100. The [pricing page](/pricing) has worked examples, and [what an AI interview costs](/blog/ai-interview-cost) covers the wider picture. ## What are the alternatives to Mercor? Stay with Mercor if you need people placed with your team and a vendor to carry the sourcing. Look at Tracy if you want to qualify the experts you already have, and read [what an interview report contains](/blog/ai-interview-reports) first. For a wider shortlist, see [AI interview tools for expert vetting](/compare/ai-interview-tools-for-expert-vetting). Hold any vendor to one test: when you question an expert's result, can you see the words behind it, and can your team change the outcome? ## Common questions ### Are Mercor reviews good? The public Trustpilot reviews are mostly negative: 7 of 8 are one-star, all dated July to November 2026. They come mostly from experts, and 8 reviews is a small base. ### What do people complain about in Mercor reviews? Projects paused after an offer (1 of 8), unpaid trial work (1), general disappointment or a scam label (3), a talent-facing operations issue (1) and an imitation site (1). ### Are there Mercor reviews on Reddit? This page quotes Trustpilot only. Reddit threads are not part of it, so treat any complaint you read there as separate evidence. ### Does Mercor publish a price for buyers? Not on the pages read. Its enterprise pages send buyers to a contact form and name two models, expert staffing and task-based pricing, with no amounts. Tracy publishes $1 per credit. ### When is Tracy the better choice than Mercor? When the experts are already your applicants. Tracy interviews each one in their own field and leaves a recording, transcript and assessment per question. Choose Mercor when you need people supplied. --- Run one role in Tracy: send us a role and we build the interview plan in days, then compare the follow-ups and the report with what you see today. The first 3 results are free. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=mercor-reviews&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=mercor-reviews&utm_content=try-free) ## Sources - [Trustpilot: Mercor reviews](https://www.trustpilot.com/review/mercor.com), retrieved 9 Nov 2026. - [Tracy AI Interviewer: Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. # micro1 Reviews: What Experts Say and What Labs Should Take From Them (Nov 2026) > micro1 reviews read for a lab buyer: 4.7 on Trustpilot from 11,219 reviews, what experts praise and criticise, and how to vet your own experts. - URL: https://tracyinterview.ai/compare/micro1-reviews - Published: 2026-11-11 - Author: Tracy team > **In short** Tracy AI Interviewer vets the experts you already have: she interviews each in their own field and leaves a recording and a timestamped transcript behind every result. > > micro1 holds 4.7 on Trustpilot from 11,219 reviews, written by experts about being screened, not by labs about buying; on page one, 17 of 20 are five-star. > > Choose micro1 when you want expert data supplied. > > Tracy's price is public: the first 3 results free, then $1 per credit. The micro1 reviews are mostly warm, and they are about a specific moment: an expert's AI interview with Zara, micro1's AI recruiter. That makes them useful if you want to know how the interview feels to the person taking it, and less useful if you are a lab deciding how to vet your own experts. This page reads them for that second reader. micro1 is the better pick when you want expert human data and evaluations supplied to you. [Tracy AI Interviewer](/product) is the better pick when the experts are already your applicants and you need each one [interviewed in their field](/use-cases/expert-vetting), with evidence you can check. ## What is micro1, and who writes its reviews? micro1 describes itself as "building the infrastructure for advancing intelligence through expert human data, real-world training environments, and contextual evaluations." Its guide says every candidate takes "A conversational interview with our AI recruiter, Zara." The interview is part of how micro1 screens experts who apply to work for it. The pages read do not say whether companies can use that interview; the home page sends companies to a contact form. ## Where do the counts come from? The main corpus is Trustpilot's page for micro1.ai: 11,219 reviews (read 10 Nov) and a TrustScore of 4.7, split 83% five-star, 10% four, 3% three, under 1% two and 3% one-star. We read the 20 reviews on page one on 9 Nov: 17 five-star, 1 four-star and 2 one-star. Eighteen are dated November 2026 and two November 2025. Separately, micro1's own reviews page shows a live tracker of 4.34 out of 5 over 465,974 experiences, beside testimonials the company chose. A review can raise more than one point, and 20 is a small slice of 11,219. ## What do reviewers praise about micro1? Twelve of the 20 reviews write about the interview itself. Ten are positive, one is mixed and one is negative. The praise is specific: "The conversation was exceptional and well focused" (Trustpilot, Nov 2026); "The AI-led technical screening was clear, structured, and straightforward to navigate." (Trustpilot, Nov 2026). Others call it smooth, well structured or well focused. micro1's own page leans the same way. One expert it features says, "Zara, the AI Interviewer, was honestly better than any human who has ever interviewed me." That is a testimonial the vendor selected, so weigh it as marketing. The Trustpilot reviews are not selected by micro1. ## What do reviewers complain about? Criticism is rare on page one: three of 20 reviews carry any, with one quote each. Counts: the 20 Trustpilot reviews of micro1.ai on page one. A review can raise more than one point. Quotes are word for word. | Issue raised | Reviews (of 20) | One quote | Reviewer and date | | --- | --- | --- | --- | | Interview language did not match the task | 1 | "In tasks that explicitly require Portuguese fluency, the interview is conducted in English; I didn't pass because of that." | Trustpilot reviewer, 1 star, Nov 2026 | | Questions drifted into grey areas | 1 | "The conversation was exceptional and well focused, though I feel like I was driven into some grey areas" | Trustpilot reviewer, 4 stars, Nov 2026 | | Service, unspecified | 1 | "Poor service" | Trustpilot reviewer, 1 star, Nov 2025 | ![Bar chart of micro1 Trustpilot reviews: 83% five-star, 10% four, 3% three, 3% one-star.](/blog/images/micro1-reviews-1.png) *Share of micro1's Trustpilot reviews by star rating. The reviewers are mostly experts who took the AI interview. Trustpilot, 11,219 reviews of micro1.ai, read 10 Nov 2026; shares as of 9 Nov.* 83% of micro1's Trustpilot reviews are five-star and 3% are one-star. ## Why do the micro1 reviews say little about buying? The one criticism with a clear cause is language: a reviewer says an interview for a Portuguese-fluency task ran in English. That is one review, but it names a question to put to any interview tool: in which language will your applicants be interviewed, and is that the language the role needs? Ask every vendor. The larger point for a buyer is who the reviewers are. They are experts being screened for micro1's own work. They say little about whether micro1 delivers good experts to labs, how fast, or at what cost. Treat the reviews as evidence that the interview experience is smooth for candidates, which matters if you copy the format, and as no evidence about supply quality. If you do buy from a marketplace, ask for references from lab customers and for the same evidence you would want from your own interviews: the recording, the transcript and the reasoning behind each pass or fail. ## micro1 or Tracy: which fits? micro1 wins on supply and on the candidate experience at scale: thousands of reviews, mostly positive, and an interview built around its own certification criteria. If you want experts provided, ask micro1. Tracy answers a different need. She interviews your applicants in their field, follows up on what they said and waits for the answer, then leaves the reviewer a recording, a timestamped transcript and a written assessment per question. A person decides. You set the skills and their weights per role, and the questions can start from a job description. A Director at a top-tier early-stage VC says of Tracy: "Having a transcript to refer back to makes a real difference." To check it on your own roles, [try three real interviews free](/connect). micro1 column: its own pages, read 9 Nov 2026. Tracy column: Tracy's published capabilities. | | micro1, as it describes itself | Tracy | | --- | --- | --- | | What you get | Expert human data and evaluations for AI teams | An interview of each applicant you already have | | Interview format | Conversational interview with its AI recruiter, Zara; no two candidates get the same questions | Questions from your job description or your own, follow-ups in the applicant's field | | Output | A real-time result against micro1's certification criteria | Recording, timestamped transcript and a written assessment per question | | Price for buyers | Not stated on the pages read; contact form | $1 per credit, first 3 results free, no seats | ![Tracy asks a fictional expert applicant which clause they flagged first, then how they would cap the liability.](/blog/images/micro1-reviews-2.png) *Tracy asks a follow-up on what an expert said, so the reviewer reads their reasoning, not only their profile. Illustration with sample data* Tracy builds each follow-up from the applicant's own answer, so the reviewer sees how an expert reasons. ## What does micro1 cost, in brief? micro1 publishes no buyer price on the pages read. Tracy's is public: $1 per credit, one credit for each started 20 minutes, the first 3 results free, no seats and no upfront cost. At 20 minutes each, 100 interviews cost $100. See the [pricing page](/pricing) for worked examples and [what an AI interview costs](/blog/ai-interview-cost). ## What are the alternatives to micro1? Stay with micro1 if you want experts and data supplied. Look at Tracy if you want to qualify the experts you already have: read [what an interview report contains](/blog/ai-interview-reports), then compare [AI interview tools for expert vetting](/compare/ai-interview-tools-for-expert-vetting). Ask any vendor one thing before you buy: when a result is questioned, can your reviewer see the words behind it? ## Common questions ### Are micro1 reviews good? On Trustpilot, yes: 4.7 from 11,219 reviews, 83% five-star. micro1's own tracker shows 4.34 out of 5. The reviewers are mostly experts who took the AI interview. ### What are the main complaints in micro1 reviews? On page one, three of 20 carry any: an interview in English for a Portuguese-fluency task, questions that drifted into grey areas, and unspecified poor service. ### Is the micro1 Zara interview well reviewed? Ten of the 12 page-one reviews that discuss the interview are positive, one is mixed and one is negative. They call it conversational, structured and well focused. ### Are there micro1 reviews on Reddit? This page quotes Trustpilot and micro1's own reviews page only. Reddit threads are not part of it. ### When is Tracy a better choice than micro1? When the experts are already your applicants. Tracy interviews each one in their field and leaves a recording, transcript and assessment per question. Choose micro1 for supplied experts. --- Run one role in Tracy: send us a role and we build the interview plan in days, then compare the follow-ups and the report with your current process. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=micro1-reviews&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=micro1-reviews&utm_content=try-free) ## Sources - [Trustpilot: micro1 reviews](https://www.trustpilot.com/review/micro1.ai), retrieved 9 Nov 2026. - [Tracy AI Interviewer: Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. # Phenom AI Interviewer Alternatives (Nov 2026) > Phenom alternatives for technical hiring: Tracy, HireVue, HeyMilo, Paradox and Sapia.ai compared on depth, volume, stack, evidence and price, Nov 2026. - URL: https://tracyinterview.ai/compare/phenom-ai-interviewer-alternatives - Published: 2026-11-11 - Author: Tracy team > **In short** Among Phenom alternatives, Tracy runs the AI interview without a platform around it: follow-ups in the person's field and a recording behind every result. A person decides. > > She starts on click from a link or your own system, and her price is published: $1 per credit, the first 3 results free. > > Stay with Phenom for one platform with scheduling, fraud checks and an audit trail; HireVue, HeyMilo, Paradox and Sapia.ai suit other needs below. > > Next step: run one role in Tracy. Teams look for Phenom alternatives when the quote, the platform scope or the interview format stops fitting. This list puts Tracy first, then four tools for other needs, each with who it suits and where it falls short. Choose Tracy when you want the interview itself, a recording and transcript behind every result, and a price you can compute today. Choose Phenom when the interview is one part of a hiring platform you want from one vendor. ## Why do teams look for Phenom alternatives? Four reasons come from Phenom's own pages. The price is a form. Its pricing page is headed "Pricing & RFP Requests". Tracy shows hers: $1 per credit, on the [pricing page](/pricing). The interview lives inside a platform. Phenom says its AI Interviewer "runs on the same platform as the rest of Phenom's recruiter experience" and starts from a candidate's profile. The pages do not say whether it is sold alone. The interview is spoken. Phenom describes a real-time voice conversation; Tracy interviews by video. The suite is wide, from career site and talent CRM to onboarding and AI agents. That is the point for some buyers and unused weight for a team that wants a first-round interview. Some teams should stay: Phenom also schedules human interviewers and runs a fraud agent that compares identity across rounds. ## Tracy AI Interviewer: the interview without the platform Tracy interviews every applicant by two-way video. After each answer she asks about what was said, in the person's field, and waits for them to finish. Behind every result sit a recording, a timestamped transcript, a written assessment per question and integrity signals shown to a human reviewer. A person decides. A coding task can sit inside the interview for technical roles. A Director at a top-tier early-stage VC says: "Tracy sometimes brings up exactly those questions I might have forgotten to ask, which has been a pleasant surprise." She starts on click at any hour with no scheduling, from an invite link, an apply page or your own system, and connects through API and MCP (ChatGPT, Claude, Mistral, Gemini, Copilot Studio), Slack, WhatsApp and Zapier. Where others are better: Phenom schedules human interviewers; scheduling for human panels is not stated on Tracy's pages. Phenom documents an audit trail, its Interview Intelligence page says its AI recommendations carry bias audits, and Tracy publishes no bias audit. HeyMilo lists two-way sync with 20+ ATS partners, and write-back is not stated on Tracy's Connect page. HireVue publishes more governance material. ![Sample interview: Tracy asks about duplicate writes, the candidate answers, Tracy follows up on concurrent retries.](/blog/images/phenom-ai-interviewer-alternatives-3.png) *Tracy AI Interviewer asks a follow-up from the candidate's own answer, so a reviewer sees how they reason, not a recited answer. Illustration with sample data* Tracy asks the second question from what the candidate just said, in their field. ## Phenom AI Interviewer: the full platform around the interview The AI Interviewer builds an interview plan and an evaluation plan, has a separate review check the plan, then holds a real-time voice conversation. Scores tie to moments in the transcript, and interviews can be re-evaluated later. The Voice Screening Agent scores each answer 1 to 5 against your rubric, and the Fraud Detection Agent adds an Identity Match Score across rounds. The hiring decision stays with your team. It publishes no price. ## HireVue: the enterprise suite with governance documents HireVue publishes an AI interviewer with probing follow-ups, live and on-demand video, and a coding assessment library. Its governance material is the deepest here: annual independent audits, adverse impact testing before models go live, and it says it is FedRAMP certified. It suits enterprise teams that want one suite at high volume, and shows no price. See [HireVue alternatives](/compare/hirevue-alternatives). ## HeyMilo: AI voice and video screening with deep ATS sync HeyMilo screens over voice, video, SMS and forms. Recruiters set follow-ups, tied to the scoring rubric. It lists two-way sync with 20+ ATS partners, 20+ languages and monthly independent bias audits. It suits high-volume teams that want the record in their ATS. Its interview pricing page leads to a demo form and states no price. Coding interview: not stated. ## Paradox: chat scheduling and basic screening Paradox handles apply, screening and scheduling by text and chat, and Workday completed its acquisition on 1 Oct 2025. Workday's datasheet says it screens for basic qualifications, and Paradox claims interviews scheduled within 10 minutes. It suits frontline hiring where scheduling is the bottleneck. An AI-led video interview with scored follow-ups is not stated on its pages, and it has no pricing page. See [Paradox alternatives](/compare/paradox-alternatives). ## Sapia.ai: a text-only interview for large volumes Sapia.ai runs five structured questions in an untimed text chat, with no video or voice, and ranks every candidate with a fit score and evidence per item. It prices per hire with unlimited interviews, says it is built for companies hiring more than 500 people a year, and publishes no rate. Follow-up questions: not stated. ## Phenom and its alternatives compared on four pillars and price Source: vendor pages and Tracy product records, Nov 2026. | Tool | Depth | Volume | Stack | Evidence | Price | | --- | --- | --- | --- | --- | --- | | Tracy AI Interviewer | Two-way video, follow-ups in the person's field | Starts on click, any hour, no scheduling | API, MCP, Slack, WhatsApp, Zapier | Recording, transcript, assessment per question, integrity signals | $1 per credit, first 3 results free | | Phenom | Real-time voice conversation, follow-ups, plan checked beforehand | Day or night, no scheduling; scheduling for later rounds | Inside the Phenom platform; scheduling on Talent CRM | Scores tied to transcript moments, re-evaluation, identity match | Not public | | HireVue | AI interviewer with probing follow-ups; coding assessments | Not stated beyond high-volume positioning | Not verified | Reviewable transcripts, score explanations, flags for human review | Not public | | HeyMilo | Voice or video, recruiter-set follow-ups tied to the rubric | Invites sent automatically, drop-offs re-engaged | 20+ ATS partners, two-way sync | Per-question rationale, quotes, transcripts, recordings, cheat flags | Interview pricing page leads to a demo form | | Paradox | Chat screening for basic qualifications | Scheduling claimed within 10 minutes | Workday-owned; SMS, WhatsApp, Messenger | Not verified | Not public | | Sapia.ai | Five untimed text questions | Unlimited interviews, built for 500+ hires a year | Integrates with your ATS or HRIS | Ranked shortlist, scores traceable to the candidate's words | Per hire, no rate published | ## What do the four pillars leave out? Rival cells: vendors' own pages, read 3 to 9 Nov 2026. Phenom's bias-audit statement is from its Interview Intelligence page. | Criterion | Who leads | Why | | --- | --- | --- | | Audit trail and re-evaluation | Phenom | An audit trail across every interview the agent conducts; interviews can be re-evaluated later. | | Bias audits | HeyMilo, HireVue | HeyMilo says it runs monthly independent audits; HireVue says annual. Phenom's Interview Intelligence page says its AI recommendations carry bias audits. Tracy publishes none. | | Languages | HeyMilo, Phenom | HeyMilo lists 20+; Phenom localises by region. Tracy adds a spoken-language check scored 0 to 100. | | Accessibility | Sapia.ai | Untimed and screen reader compatible, by its own reply. Tracy and Phenom: not stated. | | Security | Level | SOC 2 Type II: Tracy attested, Phenom says certified. | ## How do you choose a Phenom alternative? Work down these questions. 1. Do you want a platform or an interview? A platform: stay with Phenom or look at HireVue. The interview alone: Tracy or HeyMilo. 2. What must the first round show? If a text answer or a chat screen is enough, Sapia.ai or Paradox fit. If you need to see how people reason in their field, you need a tool that asks the second question: Tracy, Phenom, HireVue or HeyMilo. See [how AI interviews work](/blog/how-do-ai-interviews-work). 3. How will you pay? Quote-only prices cannot be compared until you know the unit and the first-year total. Tracy's is $1 per credit. 4. What must it connect to? Name your ATS and ask each vendor for the connector. Which path fits you: - Engineers and other technical roles: [technical interviews](/use-cases/technical-interviews). - Hundreds or thousands of applicants: [volume hiring](/use-cases/volume-hiring). - Subject experts to vet: [expert vetting](/use-cases/expert-vetting). - Programme, fellowship or admissions applications: [applications](/use-cases/applications). - A first-round budget: [Tracy's pricing](/pricing). Then run 20 applicants through your top two and have one reviewer read both outputs. For wider lists, read [best AI interview software](/blog/best-ai-interview-software). ## Common questions ### What is the best Phenom alternative? It depends on the job. Tracy AI Interviewer suits technical and expert roles that need follow-ups and evidence at a published price. HireVue is the nearest enterprise suite, HeyMilo is an AI voice and video screener, Paradox handles chat scheduling and basic screening, and Sapia.ai runs a text-only interview. ### Who are Phenom's competitors? Among the tools on this page, HireVue competes for enterprise suites, HeyMilo for AI screening, Paradox for conversational scheduling and Sapia.ai for text-based volume hiring. Tracy competes where the first round needs follow-ups. ### Is there a Phenom alternative with public pricing? Yes. Tracy costs $1 per credit with the first 3 results free. Phenom, HireVue, Paradox and Sapia.ai publish no price, and HeyMilo's interview pricing page is a demo form. ### Is the Phenom AI Interviewer voice only? Phenom describes a real-time voice conversation. Tracy interviews by two-way video. ### Which alternative has the best integrations? HeyMilo lists two-way sync with 20+ ATS partners. Tracy connects through API, MCP, Slack, WhatsApp and Zapier. --- Create one interview for the workflow your current tool misses: send us one role and we build the interview plan, customised in days, with a recording and report you can check. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=phenom-ai-interviewer-alternatives&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=phenom-ai-interviewer-alternatives&utm_content=try-free) ## Sources - [Tracy AI Interviewer, Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Connect](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. # Phenom Pricing: AI Interviewer and Voice Screening Costs (Nov 2026) > Phenom pricing is quote-only. Tracy's is public: $1 per credit. Compare features, see the cost at 100 to 1,000 interviews and what to ask on a quote call. - URL: https://tracyinterview.ai/compare/phenom-pricing - Published: 2026-11-11 - Author: Tracy team > **In short** Tracy AI Interviewer publishes her price: $1 per credit, one credit per started 20 minutes, the first 3 results free, no seats. Phenom pricing is not public; its pricing page is a form headed "Pricing & RFP Requests". > > Each Tracy result comes with a recording, a timestamped transcript and an assessment per question. A person decides. > > Phenom's AI Interviewer pages lead on audit trail, re-evaluation, localisation and later-round scheduling. Choose Phenom if you want one platform and will buy through a quote. > > Next step: run three real interviews free and estimate your cost. Phenom pricing is only available through a form. Tracy AI Interviewer publishes hers on [Tracy's pricing](/pricing), so you can budget a first round before you talk to anyone. Choose Tracy when: - You want to start an interview from a link, an apply page or your own system, and connect it to your ATS and chat tools through API, MCP, Slack, WhatsApp and Zapier. - You want the price at 100, 500 and 1,000 interviews before a sales call. - The first round has to show how someone reasons in their own field, with a recording and transcript behind every result. Choose Phenom when: - You want one platform for career site, talent CRM, scheduling, interviews and fraud checks, and you will buy it through a quote. - You need scheduling and structured guides for human interviewers in later rounds. - Your buyers ask for the bias-audit statements on Phenom's Interview Intelligence page, or its SOC 2 Type II and GDPR and CCPA statements. ## What is public about Phenom pricing? Nothing that is a price. The pricing page says a Phenom expert will answer your pricing or RFP questions. Navigation lists products from Talent CRM to the Voice Screening Agent, with the AI Interviewer beside them. None of those pages shows a plan, a unit or a trial. What Phenom's pricing and product pages say about cost, with Phenom's own words. Source: Phenom pages, Nov 2026. | What the page covers | What Phenom says | Exact words | | --- | --- | --- | | Plans and prices | No plans and no prices. A contact form. | "Pricing & RFP Requests" | | How you get a number | A Phenom expert answers pricing or RFP questions. | "answer your pricing or RFP questions" | | Pricing unit | Not stated. No per-interview, per-seat or per-hire unit. | Not stated | | Free trial | Not stated. The product pages offer a demo. | "Book a demo" | ## What does Tracy cost at 100, 500 and 1,000 interviews? She charges one credit for each started 20 minutes of an interview that produced a result, and a credit costs $1. A 15-minute interview is $1 and a 34-minute interview is $2. Credits never expire, a session with no answers is never charged, and API and MCP access (see [Connect](/connect)) is included. The first 3 results are free, and the totals ignore them. Tracy cost by interview length. Phenom publishes no figure to put beside it. | Interviews | Tracy, up to 20 minutes each (1 credit) | Tracy, 21 to 40 minutes each (2 credits) | Phenom | | --- | --- | --- | --- | | 100 | $100 | $200 | Not public. Quote on request. | | 500 | $500 | $1,000 | Not public. Quote on request. | | 1,000 | $1,000 | $2,000 | Not public. Quote on request. | ![Three figures: $100 for 100 interviews, $500 for 500 and $1K for 1,000, each up to 20 minutes.](/blog/images/phenom-pricing-1.png) *Tracy AI Interviewer cost at 100, 500 and 1,000 interviews of up to 20 minutes each, with a recording and report behind every result. $1K is $1,000. Phenom publishes no price to put beside it. Source: Tracy pricing page, Nov 2026.* Tracy's cost is a number you can compute today: $1 per interview of up to 20 minutes at any volume. ## How do you estimate a Phenom quote? Our reasoning, not a Phenom price: ask whether the quote is one yearly number. Add the first-year licence, implementation and every module the AI Interviewer needs, then divide by the interviews you expect to complete. Each $1,000 of that total buys 1,000 Tracy interviews of up to 20 minutes, so if you will complete fewer interviews than your quote has dollars, Tracy costs less per interview. For the wider market, see [AI interview cost](/blog/ai-interview-cost). ## How do Phenom and Tracy compare on the four pillars? Compare the two on Depth, Volume, Stack and Evidence, then price. Phenom's AI Interviewer and Voice Screening Agent are two products, and the table says which one each cell describes. Phenom's AI Interviewer pages lead on localisation, audit trail and re-evaluation, and it schedules later rounds; its Interview Intelligence page says its AI recommendations carry bias audits. Neither states accessibility, and Phenom's support and implementation scope are not verified. Tracy's contract starts without a quote, and Enterprise includes a dedicated customer manager and a setup customised to your workflow. For the enterprise-suite view, see [Tracy vs HireVue](/compare/tracy-vs-hirevue). Phenom column: Phenom's own product pages, read 9 Nov 2026. Tracy column: approved Tracy product records. | Pillar | Tracy AI Interviewer | Phenom | What to ask | | --- | --- | --- | --- | | Depth | Two-way video interview. She asks follow-ups in the person's own field and waits for the answer. Questions come from your brief or your own list, and a coding task can sit inside the interview. | The AI Interviewer holds a real-time voice conversation, follows up when an answer needs more depth and checks its interview plan before any candidate hears it. The Voice Screening Agent scores each answer 1 to 5 against your rubric. | A live run on your role, and how many answers get a second question. | | Volume | Every applicant starts on click, at any hour, with no scheduling. Each interview is ready in under 10 seconds. A dropped applicant can rejoin within 10 minutes (configurable per customer). | Candidates complete the AI Interviewer day or night with no scheduling. The Voice Screening Agent runs 24/7 by browser or callback. Automated Interview Scheduling books later rounds. | What happens when 200 start in one day. | | Stack | Start from an invite link, an apply page or your own system. API and MCP (ChatGPT, Claude, Mistral, Gemini, Copilot Studio), Slack, WhatsApp and Zapier, and use beside the Workable and Breezy HR connectors. | The AI Interviewer starts from a candidate profile inside Phenom and runs on the same platform as its recruiter tools. Scheduling runs on Talent CRM, with bi-directional integration to your system of record. | Whether the AI Interviewer is sold without the CRM, and your ATS. | | Evidence | A recording of both sides, a timestamped transcript, a written assessment per question with its reasoning, and integrity signals with their evidence. A person decides. | Every score traces to a moment in the transcript, and interviews can be re-evaluated later. The Fraud Detection Agent adds an Identity Match Score, never an automatic decision. | One result, from the assessment to the sentence the candidate said. | | Price | $1 per credit, one credit per started 20 minutes. First 3 results free, no seats, no upfront cost, and an empty session is never charged. | Not public. The pricing page is a form for pricing and RFP requests. | The unit, and the first-year total at your volume and at twice your volume. | ## What should you ask on a Phenom quote call? A quote is only comparable once you know the unit and the total. Ask in writing: 1. Is the AI Interviewer sold on its own, or does it need Talent CRM and other modules? 2. What is the pricing unit: per interview, per seat, per hire or a flat licence? 3. What is the first-year total at your volume, and at twice your volume? 4. What do implementation, training and professional services cost? 5. Which connector does your ATS get, and does it write scores and recordings back? 6. What are the term length, renewal and exit terms? Then put the same role through Tracy. Her first 3 results are free. ## Is Phenom worth the quote, and for whom? If you want one platform for the career site, CRM, scheduling, interviews and fraud checks, Phenom is built for that and its SOC 2 Type II, GDPR and CCPA statements are broader than Tracy's published ones. The quote may be worth chasing, especially if you already use its CRM. For a technical team that wants the first round to show how each applicant reasons, Tracy is the easier thing to test: the first 3 results are free, there is no seat to buy and the next 100 interviews of up to 20 minutes cost $100. One customer cut its average time to hire from 44 days to 7. Enterprise is priced to how many interviews you run; book a demo for a figure. For an ATS bundle against a dedicated interviewer, read [ATS AI interviewer vs dedicated](/blog/ats-ai-interviewer-vs-dedicated). Which path fits you: - Engineers and other technical roles: [technical interviews](/use-cases/technical-interviews). - Hundreds or thousands of applicants: [volume hiring](/use-cases/volume-hiring). - Subject experts to vet: [expert vetting](/use-cases/expert-vetting). - Programme, fellowship or admissions applications: [applications](/use-cases/applications). ## Common questions ### How much does Phenom cost? Phenom does not publish a price. Its pricing page is a form for pricing and RFP requests, and the AI Interviewer and Voice Screening Agent pages offer a demo and no price. Ask whether they are priced alone or with the platform. ### Is Phenom cheaper than Tracy at low volume? Phenom publishes no price, so we cannot say. Each $1,000 of first-year quote equals 1,000 Tracy interviews of up to 20 minutes, so divide your quote by the interviews you will complete. ### Is there a free way to try Tracy? Yes. The first 3 results are free and no card is needed. There are no seats and no upfront cost, and a session with no answers is never charged. ### What does Tracy Enterprise cost? Enterprise is priced to how many interviews you run, and includes a dedicated customer manager and a setup customised to your workflow. Book a demo for a figure. --- Try three real interviews and calculate your completed-interview cost: bring your annual volume, interview length and stack, and send us one role. We build the interview plan, customised in days, with a recording and report you can check. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=phenom-pricing&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=phenom-pricing&utm_content=try-free) ## Sources - [Phenom, Pricing and RFP requests](https://www.phenom.com/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. # Tracy vs HackerEarth AI Interview Agent (Nov 2026) > HackerEarth AI interview vs Tracy AI Interviewer for technical hiring: follow-ups, ID checks, ATS, evidence and price, and where each one wins. - URL: https://tracyinterview.ai/compare/tracy-vs-hackerearth - Published: 2026-11-11 - Author: Tracy team > **In short** Verdict: for a technical first round, Tracy AI Interviewer shows how each applicant reasons. She asks own-field follow-ups, can run a coding task inside the interview, and leaves a recording and an assessment per question. > > HackerEarth wins when you need ID checks with face match, full proctoring, native ATS connectors or one vendor for assessments and live panels. > > Next: test both on one role. This comparison is for engineering and talent leads who screen technical people at volume, including platforms that vet specialists for AI work. HackerEarth's AI interviewer, OnScreen, interviews candidates through a lifelike avatar while they write code in a live editor. Tracy AI Interviewer runs a two-way video interview, asks the next question from what the person just said, and leaves evidence behind every result. ## What does the HackerEarth AI interview agent do well? OnScreen verifies identity before the interview with a government ID, face match and live presence detection, then runs behavioral monitoring, AI-answer detection and deepfake detection through it. It adapts its questions to each answer inside a fixed framework, so the same skills are probed and the same rubric applied to every candidate, and for technical roles the candidate writes real code in a live IDE while the avatar asks about it. HackerEarth says it does not analyze facial expressions, body language, tone or sentiment. Add native Workday, Greenhouse, Lever and Ashby connectors, SOC 2 Type II and five ISO standards on its Trust Center, and one customer interviewing 2,000+ candidates in a weekend, and it is a serious option for regulated and high-fraud hiring. ## HackerEarth vs Tracy: buyer criteria side by side Rival column: HackerEarth's own pages, read 9 Nov 2026. Tracy column: approved Tracy product records. Not stated and Not verified mean the page does not say, not that it is absent. | Criterion | Tracy AI Interviewer | HackerEarth OnScreen | Who leads | | --- | --- | --- | --- | | Interview depth | Follow-ups in the person's own field, waiting for the answer. Questions from the job description or your own list. A coding task (live coding, SQL, debugging) inside the interview with follow-ups on approach and trade-offs. | Adaptive follow-ups in a deterministic framework: same skills, same rubric. Real code in a live IDE while the avatar probes it. | Level on coding. Tracy on own-field questions; HackerEarth on a calibrated coding library. | | Volume and start | Each applicant starts on click, at any hour. Ready in under 10 seconds. Rejoin within 10 minutes, configurable per customer. | Candidates interview whenever ready. Plans cap invites at 10 or 50 a month. | Tracy, because pay as you go has no invite allowance to outgrow. | | Evidence per result | Recording of both sides, timestamped transcript, assessment per question with reasoning. A person decides; a record shareable by link or PDF. | Scores on your rubric, transcript with exchanges highlighted, the code, integrity signals, recording and playback. | Level on what is published. | | Identity and integrity | Signals for a reviewer: tab changes, gaze, live face detection, delay before answers, fullscreen exits and external screens. Identity checks against a government ID: Not stated. | Government ID, face match, live presence, deepfake detection and a full proctoring stack. | HackerEarth. | | Price | $1 per credit, one credit per started 20 minutes. First 3 results free. No seats, no upfront cost. | OnScreen from $129 a month for 10 invites, $499 for 50. Enterprise on request. | Tracy for a price that follows usage. | | ATS and stack | API and MCP (ChatGPT, Claude, Mistral, Gemini, Copilot Studio), Slack, WhatsApp and Zapier, and use beside the Workable and Breezy HR connectors. | Native Workday, Greenhouse, Lever and Ashby connectors, plus an API. | HackerEarth on ATS depth. Tracy on chat tools and automation. | | Security, support and implementation | SOC 2 Type II attested, report on request at hello@tracyinterview.ai. Enterprise: dedicated customer manager, support channel, setup customised in days. | Lists SOC 2 Type II, five ISO standards, and GDPR, CPRA and DPDPA. Premium support on Scale; account manager and SLAs on Enterprise. | HackerEarth lists more standards. Level on support. | | Accessibility and languages | Accessibility: Not stated. A spoken-language check scores fluency, coherence, grammar, pronunciation and vocabulary from 0 to 100. | Lists ADA among its standards and multi-language support. Details Not verified. | Tracy for the spoken-language check. | | Validation, bias controls and oversight | A person reviews the evidence and decides. Tracy publishes no bias audit. | Says it runs adverse-impact studies and validates scoring with industrial-organizational psychologists; the decision stays with your team. | HackerEarth documents more. | ## How the published workflows differ for the same technical role Take one invented role, a senior backend engineer on a payments team. The table sets HackerEarth's published description of an OnScreen interview beside an illustration of a Tracy interview with a fictional candidate. The Tracy column is not a recording of a real interview. How the published workflows differ for the same technical role. The Tracy column is an illustration. | Step | HackerEarth OnScreen (from its pages) | Tracy (illustration) | | --- | --- | --- | | Before the interview | The candidate confirms identity with a government ID, face match and live presence detection. | The candidate opens a link in a browser, with no account, and is told on the join screen that the conversation is recorded and transcribed. | | Interview | A lifelike avatar asks adaptive questions by voice and video while the candidate writes code in a live IDE. | Tracy opens a coding task inside the same interview and asks about the approach as the code appears. | | Follow-up | The avatar probes reasoning and edge cases inside a deterministic framework. | She asks about what the candidate just said, waits for the end of the thought, then asks the next question. | | What you receive | Scores, transcript, code, integrity signals and a recommendation, with the decision left to your team. | A recording of both sides, a timestamped transcript, an assessment per question with its reasoning and integrity signals with evidence. | ![Recreated Tracy result report: a recording, a transcript with each follow-up question after the answer, and four questions with what the candidate said and quoted evidence.](/blog/images/tracy-vs-hackerearth-1.png) *Tracy keeps the follow-up beside the answer it came from, with a recording and a timestamped transcript behind the result, so a reviewer can check what was said. Illustration with sample data* A recording and a timestamped transcript sit behind the result, and each follow-up question follows the answer it came from. ## Do Tracy's follow-ups show more than a fixed framework? Both tools say they follow up. HackerEarth's design makes answers comparable and scores against a library calibrated on 150M+ evaluated solutions, a benchmark not stated for Tracy. Tracy asks the questions you set for the role, then follows up on what the person just said. A Director at a top-tier early-stage VC says, "Tracy sometimes brings up exactly those questions I might have forgotten to ask, which has been a pleasant surprise." A chain Tracy could run on that role, with an invented candidate: > Tracy: Your limiter keeps every call in a list. What happens to memory for an idle client? > > Candidate: It grows, so I would drop calls older than the window when a new one arrives. > > Tracy: How do you drop them without scanning the whole list each time? The second question shows how the person thinks about cost. Run both tools on your own role and check whether the second question depends on the first answer. Read [AI technical interviews](/blog/ai-technical-interviews) for the mechanics. ![Three lines of a sample interview: Tracy asks about idle client memory, the candidate answers, Tracy asks a follow-up.](/blog/images/tracy-vs-hackerearth-2.png) *Tracy builds each follow-up on the candidate's last answer, so the transcript shows depth a fixed list would miss. Illustration with sample data* The second question builds on the first answer instead of moving to the next item on a list. ## Can a reviewer check a HackerEarth or a Tracy result? Both give a reviewer something to open, and a person decides in both. HackerEarth verifies identity before the interview and runs its proctoring stack through it. Tracy works inside the conversation: she captures tab changes, gaze, live face detection, delays before answers, fullscreen exits and external screens, and shows the evidence to a person. These are signals to interpret, not proof of cheating, and nothing is decided in the interview. Candidates are told on the join screen that the conversation is recorded and transcribed. See [AI interview reports](/blog/ai-interview-reports) and the [participant experience](/participant-experience). ## Does Tracy hold up at volume? One customer cut its average time to hire from 44 days to 7, and Tracy did the work of 53 full-time recruiters in one month. See [hiring at volume](/use-cases/volume-hiring). ## What do HackerEarth and Tracy cost? HackerEarth's OnScreen is $129 a month for 10 interview invites or $499 for 50, or $1,290 and $4,990 a year, with a 30-day free trial on the Growth plan and Enterprise on request. Coding assessments are a separate subscription. Tracy charges $1 per credit for each started 20 minutes of an interview that produced a result, so a 45-minute interview costs $3. The first 3 results are free, there are no seats, and a session with no answers is never charged. Over a year, 100 interviews of 45 minutes cost $300 on Tracy against $1,290 for the cheapest OnScreen plan that covers them. See [Tracy's pricing](/pricing) and [AI interview cost](/blog/ai-interview-cost). ## How do you test both before you commit? Run one role, one team and two weeks. Pick a technical role with at least 20 expected applicants. Ask HackerEarth for a live OnScreen run on it. In Tracy, describe the role and edit the drafted questions, or send us the role and we build the plan; the first 3 results are free, so run the role in Tracy and compare the follow-ups and the report. Have one reviewer read the reports from your next 20 applicants and count the follow-ups that landed, the results verified from the recording in under a minute, and the cost per interview. Which path fits you: - Technical hiring: [technical interviews](/use-cases/technical-interviews). - High-volume hiring: [volume hiring](/use-cases/volume-hiring). - Subject experts: [expert vetting](/use-cases/expert-vetting). - Your stack: [Connect](/connect) lists the integrations. ## Common questions ### Does the HackerEarth AI interviewer ask follow-up questions? Yes. HackerEarth says OnScreen adapts to each response inside a fixed framework and asks about the code written in a live IDE. Tracy also asks follow-ups in the candidate's own field and runs a coding task inside the interview. ### Does HackerEarth verify who the candidate is? Yes. OnScreen uses government ID verification, face match and live presence detection before the interview. Tracy shows integrity signals with evidence to a reviewer; identity checks against a government ID are Not stated. ### Is Tracy better than HackerEarth? HackerEarth is better for identity verification, proctoring depth and native ATS connectors. Tracy is better when you want questions that follow each answer, a recording of both sides behind every result, and a price per interview. ### How much does HackerEarth cost compared with Tracy? OnScreen plans start at $129 a month for 10 invites. Tracy is $1 per credit, so 100 interviews of up to 20 minutes cost $100, and the first 3 results are free. ### Is Tracy secure enough for our candidate data? Tracy is SOC 2 Type II attested, and the report is available on request at hello@tracyinterview.ai. HackerEarth lists more standards, including five ISO standards. --- Run one role in Tracy and compare the follow-ups and the report: send us the role and we build the interview plan, customised in days, with a recording and report you can check. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=tracy-vs-hackerearth&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=tracy-vs-hackerearth&utm_content=try-free) ## Sources - [HackerEarth, OnScreen pricing](https://www.hackerearth.com/ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Connect](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. - [Tracy AI Interviewer, home](https://tracyinterview.ai/), retrieved 9 Nov 2026. # Eightfold AI Interviewer Alternatives (Nov 2026) > Eightfold alternatives for technical hiring: Tracy, HireVue, HeyMilo, Alex and Paradox compared on depth, volume, stack, evidence and price, Nov 2026. - URL: https://tracyinterview.ai/compare/eightfold-ai-interviewer-alternatives - Published: 2026-11-10 - Author: Tracy team > **In short** Tracy AI Interviewer is first on this shortlist for a team that wants a first round without a sales cycle: own-field follow-ups and a recording, transcript and assessment per question. > > She starts on click and costs $1 per credit, with the first 3 results free and no seats. > > Choose Eightfold for Oracle fit or a published bias audit; HireVue for one suite. > > Next: run one technical role through your top two. This guide to Eightfold alternatives is for technical teams that hire at volume. Tracy AI Interviewer leads the list, then come HireVue, HeyMilo, Alex and Paradox, each judged on interview depth, volume, stack and evidence. Eightfold stays the benchmark, and where it is the better pick, we say so. ## Why do teams look for Eightfold alternatives? Eightfold publishes no price and routes buyers to a demo or the Oracle Cloud Marketplace; it says AI Interviewer is purpose-built for organisations with 1,000 to 5,000 employees; it starts from an invite link or your ATS and sits inside Oracle Recruiting; and its 360 Interview runs a voice-led session with coding and case sections. A team that wants a published price may look elsewhere; Eightfold's API and MCP options are not stated on its AI Interviewer page. ## Tracy AI Interviewer: follow-ups in the person's field Tracy runs a two-way video interview. She asks follow-ups in the person's own field, a coding task can sit inside the interview, and questions come from your brief or your own list. Every applicant starts on click, at any hour. The result is a recording of both sides, a timestamped transcript and an assessment per question with its reasoning, and integrity signals go to a person as evidence. A Director at a top-tier early-stage VC says Tracy "sometimes brings up exactly those questions I might have forgotten to ask, which has been a pleasant surprise." Tracy connects through API and MCP, Slack, WhatsApp and Zapier, and works beside the Workable and Breezy HR connectors. See [Connect](/connect). Price: $1 per credit, one credit per started 20 minutes, the first 3 results free, no seats, and 100 interviews of up to 20 minutes for $100. Pick another tool for a published bias audit, FedRAMP or Oracle-native delivery: Tracy publishes no bias audit. See [Tracy vs HireVue](/compare/tracy-vs-hirevue). ![Tracy asks how a retry is handled, hears the answer, then asks what stops the same rows loading twice.](/blog/images/eightfold-ai-interviewer-alternatives-1.png) *Tracy follows up on the candidate's answer in their own field, and a person reads the exchange in the recording and transcript. Illustration with sample data* Tracy asks about what the candidate just said, so a reviewer sees how they reason. ## HireVue: the enterprise suite with governance depth HireVue is an enterprise suite. Its material stresses governance depth, validated science, scale, live, on-demand and AI interviews in one suite, SOC 2 Type 2, FedRAMP and 20+ languages. Its pricing page lists no prices and no plans. Pick HireVue for one vendor from engagement to interview when procurement values audit trails and validation over a quick start. See [HireVue alternatives](/compare/hirevue-alternatives). ## HeyMilo: many channels and two-way ATS sync HeyMilo interviews over voice, video, SMS and forms. The recruiter sets how many follow-ups it asks per question, tied to the scoring rubric. It syncs two ways with 20+ ATS partners, supports 20+ languages and has an MCP server and public API. Its pricing path redirects to a book-a-demo form. Pick HeyMilo if your ATS is on its sync list and you want text and voice as well as video. See [Tracy vs HeyMilo](/compare/tracy-vs-heymilo). ## Alex: video, phone, SMS and WhatsApp with identity checks Alex, formerly Apriora, runs two-way interviews over video, phone, SMS and WhatsApp and asks follow-ups guided by an Interview Guide. It integrates with 33+ ATS, says it supports 30+ languages, and adds proctoring plus an identity check with a government ID and a live selfie. Its pricing path redirects to a demo form. Pick Alex if identity verification and a wide channel mix matter. See [Tracy vs Alex](/compare/tracy-vs-alex). ## Paradox: text screening and scheduling for hourly roles Paradox is a conversational hiring assistant for apply, screening and scheduling by text and chat, focused on frontline and hourly hiring. It handles multi-location scheduling and integrates with Workday and SAP SuccessFactors. Its pages describe qualification screening; a scored interview with follow-ups is not stated. Its pricing path redirects to the homepage. Pick Paradox for scheduling and screening hundreds of hourly applicants. See [Tracy vs Paradox](/compare/tracy-vs-paradox). ## Eightfold and its alternatives compared on four pillars and price Five tools and Eightfold on four pillars and published price, Nov 2026. | Tool | Depth | Volume | Stack | Evidence | Price | What to test | | --- | --- | --- | --- | --- | --- | --- | | Tracy AI Interviewer | Follow-ups in the person's field, coding task. | Starts on click, any hour. | API, MCP, Slack, WhatsApp, Zapier. | Recording, transcript, assessment per question. Integrity signals for a person. | $1 per credit, published. | One technical role. | | Eightfold AI Interviewer | Follow-up when an answer is incomplete; coding and case. | Over one million interviews per hour claimed. | Invite link or ATS; Oracle Recruiting. | Recordings, transcripts, reasoning per section. Integrity flags for a person; optional ID check. | Not public. | A live run on your role. | | HireVue | Live, on-demand and AI interviews. | Enterprise high volume. | ATS integrations. | AI explains each response against criteria. | Not public. | Governance fit. | | HeyMilo | Capped follow-ups tied to the rubric. | Voice, video, SMS, forms. | 20+ ATS, two-way sync; MCP, API. | Scores, quotes, transcripts, recordings. | Not public. | Your ATS sync. | | Alex | Follow-ups guided by an Interview Guide. | Video, phone, SMS, WhatsApp. | 33+ ATS. | Transcripts, scores, fraud flags. | Pricing page is a demo form; its blog says from $36K a year (staffing). | Identity check. | | Paradox | Qualification screening; no scored follow-up interview stated. | Frontline and hourly. | Workday, SAP; open API. | Results go to the ATS as qualified or not. | Not public. | Whether a technical role needs a real interview. | ## What do the four pillars leave out? Buyer criteria beyond the four pillars, Nov 2026. | Criterion | Tracy AI Interviewer | Eightfold AI Interviewer | Verified edge elsewhere | | --- | --- | --- | --- | | Implementation and support | Enterprise adds a dedicated customer manager and support channel, with setup customised in days. The free start needs no sales call. | Says teams are up and running in weeks, with a 1 to 2 week activation. Support terms: Not verified. | HeyMilo lists a named CSM and a forward deployed engineer. | | Security and compliance | SOC 2 Type II attested. | FedRAMP Moderate, SOC 2 Type II and four ISO certifications. | HireVue lists SOC 2 Type 2 and FedRAMP. | | Accessibility and languages | Spoken-language check beside the transcript. Accessibility accommodations: Not stated. | 24+ languages, with CEFR-based language assessment. Candidates with accessibility needs go to a human-led process. | Alex says 30+ languages; Paradox answers questions in 100+. | | Validation and bias | A person reviews the evidence and decides. Tracy publishes no bias audit. | Independent BABL AI audit under NYC Local Law 144, audited 29 June 2026. | HeyMilo and Alex say monthly audits by Warden AI. | | Contract and first-year cost | $1 per credit, one per started 20 minutes. First 3 results free, no seats. | No public price. Demo route, plus an Oracle Cloud Marketplace listing. | None of the other pricing pages lists a price. Alex's blog says annual agreements start at $36K for staffing firms. | ## How do you choose an Eightfold alternative? Start with the constraint that decides it. Governance, FedRAMP and audit detail point to Eightfold (published BABL AI audit) or HireVue (adverse impact testing, SOC 2 Type 2, FedRAMP). Hourly text screening points to Paradox. Voice, video and an ATS sync point to HeyMilo or Alex. A price you can read and a technical interview you can start free point to Tracy. Then run one technical role through each finalist and have a reviewer count the follow-ups that landed and the cost per interview. For the mechanics, read [how AI interviews work](/blog/how-do-ai-interviews-work). Which path fits you: - Technical roles: [technical interviews](/use-cases/technical-interviews). - Hundreds of applicants: [volume hiring](/use-cases/volume-hiring). - Integrations and price: [Connect](/connect) and [pricing](/pricing). ![Four tiles: Eightfold or HireVue for audit detail, Paradox for hourly text, HeyMilo or Alex for ATS sync, Tracy for price.](/blog/images/eightfold-ai-interviewer-alternatives-2.png) *A shortlist by deciding constraint, ending with Tracy for a published price and a free start. Source: vendor pages, Nov 2026.* Pick by the constraint that decides it, then test two or three tools on one role. ## Common questions ### What are the best Eightfold alternatives? For a technical team at volume: Tracy AI Interviewer, HireVue, HeyMilo, Alex and Paradox. Tracy fits follow-ups in the person's field with a published price, HireVue fits one enterprise suite, HeyMilo and Alex fit multi-channel interviews synced to an ATS, and Paradox fits hourly text screening. ### Who are Eightfold's competitors? In AI interviewing, buyers compare it with HireVue, HeyMilo, Alex, Paradox and Tracy. Each solves a different first-round job, so compare them on one role. ### Is there a cheaper alternative to Eightfold? Eightfold publishes no price, so we cannot say. Tracy's is public: $1 per credit, with the first 3 results free. Alex's blog states annual contracts from $36K for staffing firms. Ask each demo-led vendor for a first-year total and divide it by the interviews you will complete. ### Which Eightfold alternative has a bias audit? Eightfold publishes an audit by BABL AI under NYC Local Law 144. HeyMilo and Alex say they run monthly audits by Warden AI. Tracy publishes no bias audit; a person decides from her evidence report. --- Create one interview for the workflow your current tool misses. Send us the role and we build the plan, customised in days, so you can read the recording and report; the first 3 results are free. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=eightfold-ai-interviewer-alternatives&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=eightfold-ai-interviewer-alternatives&utm_content=try-free) ## Sources - [Eightfold AI, AI Interviewer](https://eightfold.ai/products/ai-interviewer/), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Connect](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. - [Tracy AI Interviewer, home page](https://tracyinterview.ai/), retrieved 9 Nov 2026. # Eightfold AI Pricing: What It Costs and What You Get (Nov 2026) > Eightfold pricing is not public. Tracy's is: $1 per credit. See a four-pillar comparison and what to ask on an Eightfold quote call, Nov 2026. - URL: https://tracyinterview.ai/compare/eightfold-pricing - Published: 2026-11-10 - Author: Tracy team > **In short** Tracy AI Interviewer lets a technical team budget a first round before any call: $1 per credit, one per started 20 minutes, the first 3 results free, no seats. > > Each interview that produces a result is billed in credits, with own-field follow-ups and a recording, transcript and assessment per question. > > Eightfold publishes no price. Choose it for Oracle fit, a published bias audit, FedRAMP or 24+ languages. > > Next step: try three real interviews free, then bring your volume. Eightfold pricing is not public. Tracy AI Interviewer's is, on [Tracy's pricing](/pricing): $1 per credit, the first 3 results free, no seats and no upfront cost. A public price means you can budget a first round before you talk to anyone. This page sets out what Eightfold's pages say about buying, what Tracy costs at 100, 500 and 1,000 interviews, how the two compare, and what to ask on a quote call. ## Is Eightfold pricing public? Nothing on the AI Interviewer pages is a price. The table lists what the pages say about buying. What Eightfold's pages say about buying AI Interviewer, Nov 2026. | What the page covers | What Eightfold says | Exact words | | --- | --- | --- | | Plans and prices | None. The pricing path returns a not-found page. | "Page not found" | | How you buy | A demo request. | "Start a live demo" | | Purchase route | Oracle Cloud Marketplace, payable with Oracle Universal Credits. No figure. | "available for purchase using Oracle Universal Credits" | | Sized for | Organisations with 1,000 to 5,000 employees. | "purpose-built for organizations with 1,000 to 5,000 employees" | | Free trial | A try-it demo. No trial terms stated. | "Try AI Interviewer now" | ## What does Tracy cost at 100, 500 and 1,000 interviews? One credit covers each started 20 minutes of an interview that produced a result, and a credit costs $1. A 15-minute interview is $1 and a 34-minute interview is $2. Credits never expire, a session with no answers is never charged, and API and MCP access (see [Connect](/connect)) is included. The first 3 results are free, and the totals below ignore them. Tracy cost at 100, 500 and 1,000 interviews; Eightfold publishes no price. | Interviews | Tracy, up to 20 minutes each (1 credit) | Tracy, 21 to 40 minutes each (2 credits) | Eightfold | | --- | --- | --- | --- | | 100 | $100 | $200 | Not public. Quote on request. | | 500 | $500 | $1,000 | Not public. Quote on request. | | 1,000 | $1,000 | $2,000 | Not public. Quote on request. | ![Four tiles: $1 per credit, 3 free results, $100 for 100 interviews, 0 seats.](/blog/images/eightfold-pricing-1.png) *Tracy's published price in four numbers, so a buyer can budget before a call: $1 per credit, 3 free results, $100 for 100 interviews of up to 20 minutes, 0 seats. Source: Tracy pricing page, Nov 2026.* Tracy's cost is a number you can compute before any call: $1 per started 20 minutes, no seats. ## Why can a quote cost more per interview at low volume? This is our reasoning, not an Eightfold price. A quote usually comes down to one yearly number: licence, implementation and integrations. Divide it by the interviews you complete. Each $1,000 of first-year quote equals 1,000 Tracy interviews of up to 20 minutes, so if you will complete fewer interviews of up to 20 minutes than your quote has dollars, pay as you go costs less per interview. Longer interviews use more credits. See [AI interview cost](/blog/ai-interview-cost) for the wider market. ## What should you ask on an Eightfold quote call? Bring the same list to every vendor you shortlist: 1. Is AI Interviewer sold alone or inside a larger contract? 2. What is the unit: per interview, per candidate, per hire, per employee or a flat fee? 3. What is the first-year total at our volume, and at twice our volume? 4. Is implementation included, and what are the term and renewal terms? 5. Are ATS connectors, the Oracle integration and an API in the price? 6. Can we run it live on our own role and see the recording and report? Tracy answers the money questions on her pricing page, so the call can go to fit. ![Four items returned per result: recording, timestamped transcript, assessment per question, integrity signals.](/blog/images/eightfold-pricing-2.png) *Tracy returns a recording, a timestamped transcript, a written assessment per question and integrity signals for each interview that produces a result, so a buyer pays for something a person can decide from. Source: Tracy product description, Nov 2026.* Each interview that produces a result returns four things a person can decide from. ## How do Eightfold and Tracy compare on the four pillars? Price is one line. The bigger question is what the interview gives you. [Tracy vs HireVue](/compare/tracy-vs-hirevue) compares another enterprise suite the same way. Eightfold and Tracy compared on four pillars and buyer criteria, Nov 2026. | Pillar | Tracy AI Interviewer | Eightfold AI Interviewer | What to ask | | --- | --- | --- | --- | | Depth | Two-way video interview. Follow-ups in the person's own field, a coding task inside the interview, questions from your brief or your own list. | Follow-up when an answer is incomplete. Questions generated from a job description you approve. 360 Interview adds coding, case and panel questions, by voice. | Ask for a live run on your role. | | Volume | Every applicant starts on click, at any hour, with no scheduling. Ready in under 10 seconds. | Starts when a candidate applies, at any hour. Claims over one million interviews per hour. | Ask what volume the quote assumes. | | Stack | API and MCP (ChatGPT, Claude, Mistral, Gemini, Copilot Studio), Slack, WhatsApp, Zapier; beside Workable and Breezy HR. | Invite link or ATS. Inside Oracle Recruiting. API or MCP route: Not verified. | Ask whether an API route is included. | | Evidence | Recording of both sides, timestamped transcript, assessment per question with reasoning. | Summaries, recordings, transcripts, replay, notes and override, with reasoning per section. | Open one result and find the sentence behind an assessment. | | Implementation and support | Enterprise adds a dedicated customer manager and support channel, with setup customised in days. The free start needs no sales call. | Says teams are up and running in weeks, with a 1 to 2 week activation. Support terms: Not verified. | Ask what support includes. | | Security and compliance | SOC 2 Type II attested. | FedRAMP Moderate, SOC 2 Type II and four ISO certifications. | Eightfold lists more. | | Accessibility and languages | Spoken-language check beside the transcript. Accessibility accommodations: Not stated. | 24+ languages, with CEFR-based language assessment. Candidates with accessibility needs go to a human-led process. | Ask which languages the quote covers. | | Validation and bias | A person reviews the evidence and decides. Tracy publishes no bias audit. | Independent BABL AI audit under NYC Local Law 144, audited 29 June 2026. | Eightfold publishes an audit. | | Contract and first-year cost | $1 per credit, one per started 20 minutes. First 3 results free, no seats. | No public price. Demo route, plus an Oracle Cloud Marketplace listing. | Ask for the first-year total and term. | ## Is Eightfold worth the quote, and for whom? For teams that run Oracle Recruiting or need a published bias audit, FedRAMP or 24+ languages, Eightfold is built for the job and its quote may be worth chasing. For a technical team that wants the first round to show how each applicant reasons, Tracy is the easier thing to test: the first 3 results are free, there is no seat to buy, and the next 100 cost $100. Tracy has completed 600K+ interviews for 300+ clients, with 99% interview completion. Enterprise is priced to how many interviews you run; book a demo for a figure. Which path fits you: - Technical roles: [technical interviews](/use-cases/technical-interviews). - Hundreds of applicants: [volume hiring](/use-cases/volume-hiring). - A first-round budget: [Tracy's pricing](/pricing). ## Common questions ### How much does Eightfold AI cost? Eightfold does not publish a price for AI Interviewer. Its product pages route to a demo. Treat any number on a review site as an estimate until a quote confirms it. ### Is Eightfold cheaper than Tracy at low volume? Eightfold publishes no price, so we cannot say. As arithmetic, each $1,000 of first-year quote equals 1,000 Tracy interviews of up to 20 minutes. ### Does Eightfold AI Interviewer have a free trial? Its product page has a try-it demo and states no trial terms. Tracy's first 3 results are free, with no card. ### Does Tracy charge per seat or per month? No. There are no seats, no subscription and no upfront cost, and a session with no answers is never charged. --- Try three real interviews free and calculate your completed-interview cost. Or bring your annual volume, interview length and stack, and we build the interview plan for one role, customised in days, with a recording and report you can check. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=eightfold-pricing&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=eightfold-pricing&utm_content=try-free) ## Sources - [Eightfold AI, AI Interviewer](https://eightfold.ai/products/ai-interviewer/), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Connect](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. # Fireflies.ai Alternatives for Interviews (Nov 2026) > Fireflies ai alternatives for interviews: Tracy runs the first-round interview and writes the report; Otter, Granola, Metaview and BrightHire take notes. - URL: https://tracyinterview.ai/compare/fireflies-ai-alternatives - Published: 2026-11-10 - Author: Tracy team > **In short** If you record interviews with a notetaker and want the report to come from the interview itself, Tracy AI Interviewer runs the first round by two-way video and writes an assessment per question, with the recording and a timestamped transcript behind every result. > > $1 per credit, the first 3 results free, no seats. > > Fireflies is the better pick when your team keeps running the calls: it has a recruiting page, ATS sync and 100+ transcription languages. > > Next: create one interview for the workflow your current tool misses. Recruiters look for Fireflies.ai alternatives when a recorded call is not the same as a finished interview report. Fireflies transcribes and summarizes interviews your team runs, and it does that well. Running the first round for every applicant is a different job, and Tracy AI Interviewer is the first tool to try for it. The list puts Tracy first, then four tools for other needs. ## Why do teams look for Fireflies.ai alternatives? These reasons come from Fireflies' own pages. The notes depend on the interviewer. Fireflies' recruiting page describes summaries, scorecards and candidate insights from recorded calls, so what you get back depends on the questions a person asked and the time they had. The bot is visible. Fireflies sends a note taker bot to the call, and recording without it is a Chrome extension feature on paid plans, so on the default setting a bot joins the call. Spend follows seats. Pro lists $10 and Business $19 per seat per month on annual billing ($18 and $29 on monthly billing), so cost tracks the number of interviewers, not interviews. The candidate-scoring skills assess what was said on a call. Asking the candidate the next question is not stated. Some teams should stay. If your interviewers run the calls and you want transcripts in 100+ languages, notes in Greenhouse or Lever, and a low seat price, Fireflies is built for that. ## Which Fireflies.ai alternative suits which team? - Tracy AI Interviewer: roles where the first round must show depth for every applicant, at a published price per result. - Otter.ai: teams that want live transcription of meetings at a low price, with no hiring focus. - Granola: interviewers who want a private notepad and no bot on the call. - Metaview: teams that want interview notes mapped into ATS scorecards. - BrightHire: teams that want a notetaker and an AI interviewer from one vendor, and accept a quote. The sections below give the detail and the limits of each. ## Tracy AI Interviewer: the first round, run and evidenced Fireflies records what your team asks; Tracy asks. Each applicant opens a link in a browser, with no account, and is told on the join screen that the conversation is recorded and transcribed. She interviews by two-way video, asks about what was said in the person's own field, and waits for the answer. A Director at a top-tier early-stage VC said Tracy "sometimes brings up exactly those questions I might have forgotten to ask." You can bring your own questions or describe the role and let Tracy draft them. Each result comes with a recording of both sides, a timestamped transcript, a written assessment per question with its reasoning and integrity signals for a human reviewer. A person keeps the last word. Tracy connects through API and MCP (ChatGPT, Claude, Mistral, Gemini, Copilot Studio), Slack, WhatsApp, Workable, Breezy HR and Zapier. The price is $1 per credit, one credit per started 20 minutes, the first 3 results free, no seats. See [Tracy's pricing](/pricing). Where Fireflies is better: it records and summarizes calls your team already runs, sends notes to Greenhouse and Lever, and transcribes in 100+ languages. Tracy is not a notetaker for your calls, and ATS write-back is not stated on her Connect page. ![Sample Tracy exchange: an invoice dispute question, the candidate's answer, then a follow-up at 03:08.](/blog/images/fireflies-ai-alternatives-1.png) *A person reviews the recording, transcript and written assessment per question. Illustration with sample data* Tracy asks the next question from the candidate's answer. A recording and a timestamped transcript sit behind every result, plus a written assessment per question. ## Otter.ai: live transcripts for any meeting Otter joins Zoom, Microsoft Teams and Google Meet, or records without a bot from a Mac or Windows desktop, Chrome or a phone. It offers live transcription in multiple languages with speaker recognition and summaries with action items. Its listed use cases are Sales, Education and Media, and a hiring use case is not stated. Pro lists $16.99 and Business $24 per user per month on monthly billing ($8.33 and $19.99 on annual billing). Pick it for general meetings with a wider mix than interviews. ## Granola: private notes with no bot Granola uses your computer's audio instead of a bot and works with Zoom, Google Meet and Teams, on macOS, Windows, iOS and Android. Notes stay private until you share them, and audio is deleted after transcription, so a recording of the call is not kept. Basic is free; Business lists $14 per user per month with Zapier, HubSpot and API access. Hiring-specific features are not stated. Pick it when no bot on the call matters most. ## Metaview: notes mapped into ATS scorecards If your reason for leaving Fireflies is ATS depth, Metaview maps interview answers into scorecards in Greenhouse, Ashby, Lever, Workday and more, and its Notetaker can work without a bot in the call. Notetaker Pro lists $60 per user per month, monthly; Screening, a separate agent, has no price on its pricing page and the all-agents plan is custom-priced. ## BrightHire: notes and an AI interviewer from one vendor If you want notes and an AI interviewer under one contract, BrightHire lists both, with auto-filled scorecards in Greenhouse, Lever, Ashby and Workday. Its pricing page lists no dollar amounts (9 Nov snapshot). ## Fireflies.ai and its alternatives compared on four pillars and price Source: each vendor's own pages and approved Tracy product records, Nov 2026. Prices are list prices; billing basis is shown per cell. | Tool | Depth | Volume | Stack | Evidence | Price | | --- | --- | --- | --- | --- | --- | | Tracy AI Interviewer | Two-way video, follow-ups in the person's field, coding task inside | Starts on click, any hour, no scheduling | API, MCP, Slack, WhatsApp, Workable, Zapier | Recording, timestamped transcript, assessment per question | $1 per credit, first 3 free | | Fireflies.ai | Notes, scorecards and candidate-scoring skills on recorded calls | Bot on the call, or Chrome extension without it | Greenhouse, Lever, CRMs, Slack, MCP | Transcript in 100+ languages, summaries | Pro $18 and Business $29 per seat per month on monthly billing; $10 and $19 on annual | | Otter.ai | Notes and summaries of calls your team runs | Joins Zoom, Teams and Meet, or records bot-free | Salesforce, HubSpot, Zapier, MCP; ATS not stated | Live transcript, speaker recognition, summaries | Pro $16.99 and Business $24 per user per month on monthly billing; $8.33 and $19.99 on annual | | Granola | Private notes on calls you run; hiring features not stated | No bot; uses your computer audio | Zapier, Attio, Notion, HubSpot, MCP, API | Transcript and notes; audio deleted after transcription | Free; Business $14 per user per month (billing basis not stated) | | Metaview | Notetaker on human interviews; Screening is a separate agent | Notetaker can work without a bot in the call | Greenhouse, Ashby, Lever, Workday and more | Notes mapped into ATS scorecards | Notetaker Pro $60 per user per month, monthly; Screening price not stated | | BrightHire | AI notes on live interviews; AI interviewer, two-way | Candidates can interview 24/7 | Greenhouse, Lever, Ashby, Workday, Slack | Auto-filled scorecards; rubric scoring | No dollar amounts on its pricing page (9 Nov snapshot) | ![Three questions: who asks, what do you read, how is it priced.](/blog/images/fireflies-ai-alternatives-2.png) *Illustration, no data: three questions to ask before you pick an interview tool. Illustration, no data.* Three questions sort the shortlist: who asks, what you read afterwards, and how the tool is priced. ## How do you choose a Fireflies.ai alternative? Start with who asks the questions. If your interviewers run the calls and you want notes and ATS sync, stay with Fireflies or compare Metaview and BrightHire's notes. If you want the first round run for every applicant, shortlist Tracy and BrightHire's AI interviewer. Then read what you get back. A call summary tells you what was discussed. A report built from an interview the AI ran gives you a written assessment per question and the recording behind it, so the reviewer can check the answer, not only the notes. Finally, compare the price model. Seats scale with interviewers; credits scale with interviews. At 100 interviews of 20 minutes Tracy is $100, and the notetakers are a separate cost for a separate job. Create one interview for your role, and read [what an AI interview report contains](/blog/ai-interview-reports) and [the best AI interview software](/blog/best-ai-interview-software) for the wider list. ## Which path fits you? - Technical hiring: [technical interviews](/use-cases/technical-interviews) covers role-specific interviews with follow-ups. - High-volume hiring: [volume hiring](/use-cases/volume-hiring) covers applicant waves. - Applicant experience: [participant experience](/participant-experience) shows what the person interviewed sees. ## Common questions ### What is the best Fireflies.ai alternative for interviews? It depends on the job. Tracy AI Interviewer suits teams that want the first round run for every applicant. Otter.ai, Granola, Metaview and BrightHire's notetaker suit teams whose interviewers keep running the calls. ### Is there a free Fireflies.ai alternative? For notes, Otter and Granola each list a free plan, as does Fireflies. None of the tools covered here lists an open-source version on the pages read: not stated. Tracy's first 3 results are free, then $1 per credit with no seats. ### Is there a Fireflies.ai alternative with no bot? Granola takes notes without a bot, and Otter and Metaview can record without one. With Tracy, the applicant opens a browser link and no bot joins a call. ### Does Tracy replace a notetaker? No. Tracy runs the first-round interview and writes the report. For notes on calls your team runs, keep a notetaker. --- Create one interview for the workflow your current tool misses: send us one role and we will build the interview plan, with a recording and report you can check. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=fireflies-ai-alternatives&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=fireflies-ai-alternatives&utm_content=try-free) ## Sources - [Fireflies.ai, Pricing](https://fireflies.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Connect](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. # Granola Alternatives for Interviews (Nov 2026) > Granola alternatives AI for interviews: Tracy runs the first round and writes the report; Otter, Fireflies, Metaview and BrightHire take notes. Compared. - URL: https://tracyinterview.ai/compare/granola-alternatives - Published: 2026-11-10 - Author: Tracy team > **In short** If you like bot-free notes but need a decision-ready interview report, Tracy AI Interviewer runs the first round by two-way video, with no bot and no call to join, and writes an assessment per question beside the recording and a timestamped transcript. > > $1 per credit, the first 3 results free, no seats. > > Granola is the better pick when you run the interviews yourself and want a private notepad. > > Next: create one interview for the workflow your current tool misses. People looking for Granola alternatives usually like what it does: quiet notes on a call, with no bot. The question is what produces a report a team can decide from. Granola keeps notes for the person who ran the call. Running the first round for every applicant is a different job, and Tracy AI Interviewer is the first tool to try for it. The list puts Tracy first, then four tools for other needs. ## Why do teams look for Granola alternatives? These reasons come from Granola's own pages. The notes are personal. They are private by default and shared only if you choose, which suits an individual. Hiring features are not stated. Its home and pricing pages describe meeting notes, templates, AI chat and integrations with Attio, Notion, HubSpot, Affinity and Zapier; interview scorecards and ATS sync are not stated. The call is not kept. Granola deletes audio after transcription, so a reviewer gets notes and a transcript but no recording to replay. Spend follows users. Business lists $14 per user per month and Enterprise $35, so cost tracks the number of interviewers. Some teams should stay. If your interviewers run the calls and want to stay present with no bot, Granola does that well, and an AI interviewer does not replace it. ## Which Granola alternative suits which team? - Tracy AI Interviewer: roles where the first round must show depth for every applicant, at a published price per result. - Otter.ai: teams that want live transcription of meetings, bot or bot-free. - Fireflies.ai: hiring teams that want notes, scorecards and ATS sync on their own interviews. - Metaview: teams that want interview notes mapped into ATS scorecards. - BrightHire: teams that want a notetaker and an AI interviewer from one vendor, and accept a quote. The sections below give the detail and the limits of each. ## Tracy AI Interviewer: the first round, run and evidenced Tracy needs no bot in the room, because there is no meeting to join. The applicant opens a browser link at any hour, with no account and no scheduling, and Tracy interviews them by two-way video. She asks about what was said, in the person's own field, and waits for the answer. A Director at a top-tier early-stage VC said Tracy "sometimes brings up exactly those questions I might have forgotten to ask." Each result comes with a recording of both sides, a timestamped transcript, a written assessment per question with its reasoning and integrity signals for a human reviewer. A person decides, and the record can be shared by link or PDF. Bring your own questions or describe the role and let Tracy draft them. Tracy connects through API and MCP (ChatGPT, Claude, Mistral, Gemini, Copilot Studio), Slack, WhatsApp, Workable, Breezy HR and Zapier. The price is $1 per credit, one credit per started 20 minutes, the first 3 results free, no seats. See [Tracy's pricing](/pricing). Where Granola is better: invisible capture on calls you run yourself, kept private to you. Tracy is not a notetaker for your calls. ![Sample Tracy exchange: a shift-planning question, the candidate's answer, then a follow-up at 07:02.](/blog/images/granola-alternatives-1.png) *A person reviews the recording, transcript and written assessment per question. Illustration with sample data* Tracy asks the next question from the candidate's answer. A recording and a timestamped transcript sit behind every result, plus a written assessment per question. ## Otter.ai: live transcripts for any meeting Otter joins Zoom, Microsoft Teams and Google Meet, or records bot-free from a Mac or Windows desktop, Chrome or a phone. It offers live transcription in multiple languages with speaker recognition and summaries with action items, and lists an MCP server on every plan. Its listed use cases are Sales, Education and Media. Pro lists $16.99 and Business $24 per user per month on monthly billing ($8.33 and $19.99 on annual billing). Pick it for live transcripts across a wide mix of meetings. ## Fireflies.ai: a notetaker with a recruiting page Fireflies sends a note taker bot to the call, with a Chrome extension option on paid plans. Its recruiting page covers interview notes, scorecards and ATS sync, and notes can go to Greenhouse and Lever. It transcribes in 100+ languages and lists SOC 2 Type II. Pro lists $10 and Business $19 per seat per month billed annually ($18 and $29 on monthly billing). Pick it when hiring-specific notes and ATS sync matter more than staying bot-free. ## Metaview: notes mapped into ATS scorecards For a Granola user who wants a bot-free option built for hiring, Metaview's Notetaker does not need a bot in the call and maps answers into ATS scorecards. Notetaker Pro lists $60 per user per month, monthly; Screening price is not stated. ## BrightHire: notes and an AI interviewer from one vendor BrightHire lists AI notes and a two-way AI interviewer; whether its notes work without a bot is not stated on the pages read. Its pricing page lists no dollar amounts (9 Nov snapshot). ## Granola and its alternatives compared on four pillars and price Source: each vendor's own pages and approved Tracy product records, Nov 2026. Prices are list prices; billing basis is shown per cell. | Tool | Depth | Volume | Stack | Evidence | Price | | --- | --- | --- | --- | --- | --- | | Tracy AI Interviewer | Two-way video, follow-ups in the person's field, coding task inside | Starts on click, any hour, no scheduling | API, MCP, Slack, WhatsApp, Workable, Zapier | Recording, timestamped transcript, assessment per question | $1 per credit, first 3 free | | Granola | Private notes on calls you run; hiring features not stated | No bot; uses your computer audio | Zapier, Attio, Notion, HubSpot, MCP, API | Transcript and notes; audio deleted after transcription | Free; Business $14 per user per month (billing basis not stated) | | Otter.ai | Notes and summaries of calls your team runs | Joins Zoom, Teams and Meet, or records bot-free | Salesforce, HubSpot, Zapier, MCP; ATS not stated | Live transcript, speaker recognition, summaries | Pro $16.99 and Business $24 per user per month on monthly billing; $8.33 and $19.99 on annual | | Fireflies.ai | Notes, scorecards and candidate-scoring skills on recorded calls | Bot on the call, or Chrome extension without it | Greenhouse, Lever, CRMs, Slack, MCP | Transcript in 100+ languages, summaries | Pro $18 and Business $29 per seat per month on monthly billing; $10 and $19 on annual | | Metaview | Notetaker on human interviews; Screening is a separate agent | Notetaker can work without a bot in the call | Greenhouse, Ashby, Lever, Workday and more | Notes mapped into ATS scorecards | Notetaker Pro $60 per user per month, monthly; Screening price not stated | | BrightHire | AI notes on live interviews; AI interviewer, two-way | Candidates can interview 24/7 | Greenhouse, Lever, Ashby, Workday, Slack | Auto-filled scorecards; rubric scoring | No dollar amounts on its pricing page (9 Nov snapshot) | ![Three steps: you run the call, Tracy runs it, a person decides from the report.](/blog/images/granola-alternatives-2.png) *Illustration, no data: a notepad when you run the call, Tracy when a first round must be run for you. Illustration, no data.* Two jobs, two tools: a notepad when you run the call, Tracy when the first round has to be run for you. ## How do you choose a Granola alternative? Start with who runs the interview. If your recruiters run it and you want quiet notes, stay with Granola or look at Otter, Fireflies.ai and Metaview. If you want it run for every applicant, shortlist Tracy and BrightHire's AI interviewer. Then decide what a reviewer needs. A private notepad serves the person who took the call. A shared record of the interview, with the recording, a transcript and an assessment per question, serves a panel that has to compare many candidates and agree. Finally, compare the price model. Users scale with interviewers; credits scale with interviews. 100 interviews of 20 minutes is $100 in Tracy. Create one interview for your role, and read [what an AI interview report contains](/blog/ai-interview-reports) and [the best AI interview software](/blog/best-ai-interview-software) for the wider list. ## Which path fits you? - Technical hiring: [technical interviews](/use-cases/technical-interviews) covers role-specific interviews with follow-ups. - High-volume hiring: [volume hiring](/use-cases/volume-hiring) covers applicant waves. - Applicant experience: [participant experience](/participant-experience) shows what the person interviewed sees. ## Common questions ### What are the best Granola alternatives for interviews? It depends on the job. Tracy AI Interviewer suits teams that want the first round run for every applicant. Otter.ai, Fireflies.ai and Metaview suit teams that keep running calls and want notes. ### Is there a Granola alternative that does not use a bot? Otter and Metaview can record without a bot. With Tracy there is no call to join: the applicant opens a browser link. ### Are there open source Granola alternatives? None of the tools covered here lists an open-source version on the pages read: not stated. Compare them on price, hiring features and the evidence a reviewer gets. ### Is Tracy a notepad like Granola? No. Tracy runs the interview and writes the report. For private notes on calls you run, use a notetaker. --- Create one interview for the workflow your current tool misses: send us one role and we will build the interview plan, with a recording and report you can check. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=granola-alternatives&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=granola-alternatives&utm_content=try-free) ## Sources - [Granola, Pricing](https://www.granola.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Connect](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. # Notta Pricing: Cost vs Tracy (Nov 2026) > Notta pricing in Nov 2026: Pro $8.17 and Business $16.67 per month on annual billing. Tracy's AI interview is $1 per credit, no seats. - URL: https://tracyinterview.ai/compare/notta-pricing - Published: 2026-11-10 - Author: Tracy team > **In short** Tracy AI Interviewer runs the first-round interview for each applicant and is priced per completed interview: $1 per credit, one credit per started 20 minutes, the first 3 results free, no seats. > > Each result comes with a recording, a timestamped transcript and a written assessment per question, and a person decides. > > Notta lists Pro at $8.17 per month on annual billing. If your own team keeps running the interviews, Notta is the cheaper pick. > > Next: try three interviews free and calculate your cost per completed interview. Notta pricing is public for Free, Pro and Business, with a custom Enterprise plan. Notta prices seats and transcription minutes: Pro includes 1,800 minutes a month and Business lists unlimited transcription. It is a notetaker: the price buys a transcript of calls your people hold, and the interview itself stays with them. **Choose Notta when** your recruiters keep running the interviews and you want transcripts and AI summaries of each call, from $8.17 per month on annual billing, or free for 120 minutes a month. **Choose Tracy AI Interviewer when** you want the first-round interview run for every applicant, priced per interview instead of per seat and per minute of quota: $1 per credit, one credit per started 20 minutes. She starts on click at any hour, with no scheduling and no recruiter present. ## Notta pricing table: every plan the page lists Source: Notta pricing page, 9 Nov 2026. Prices are Notta list prices in USD at the annual rate, which the page shows by default at 40% off; its Monthly toggle shows monthly billing. | Plan | Price per month | What it lists | Page words | | --- | --- | --- | --- | | Free | $0 | 1 seat; 120 transcription minutes a month, up to 3 minutes per conversation; 10 AI summaries | "120 transcription minutes/month" | | Pro | $8.17 ($97.99 billed annually) | 1 seat; 1,800 transcription minutes, no carryover; up to 5 hours per recording; 100 AI summaries | "Total $97.99 · Billed annually" | | Business | $16.67 ($199.99 billed annually) | Unlimited transcription; 200 AI summaries; web meeting video; usage report; CRM and Zapier integration | "Total $199.99 · billed annually" | | Enterprise | Custom | Starting from 51 seats; SAML SSO, audit logs, priority support | "Starting from 51 seats" | ## What costs extra on Notta Notta's add-ons are billed on top and are available to Pro, Business and Enterprise: monolingual translation at $6 a month on annual billing ($10 monthly), bilingual transcription and translation at $9 ($15 monthly), and Notta Brain at $93.59 a year with 8,000 AI credits a month. The page says the transcription quota has no carryover, so unused Pro minutes lapse each month. It lists a 50% education discount and a 7-day Business trial. ## What a typical team pays for Notta Take a team where five people run interviews. Five Pro seats at $8.17 are $40.85 a month on annual billing. Notta shows Business as "seat(s)" without a count in the text we read, so this assumes the listed price is per seat. On that assumption five Business seats at $16.67 are $83.35 a month, or about $1,000 a year ($199.99 each). This is our arithmetic from list prices, not a Notta quote. Minutes decide whether Pro is enough: 100 interviews of 20 minutes is 2,000 minutes, above the 1,800 a Pro seat lists. ![Three steps: one credit per started 20 minutes, evidence per result, a person decides.](/blog/images/notta-pricing-2.png) *Illustration: what one Tracy credit buys. A recording and a timestamped transcript sit behind every result, with a written assessment per question, and a person decides. Illustration with sample data* One credit covers an interview with a recording, a transcript and a written assessment per question behind the result. ## How Tracy is priced A credit costs $1 and covers each started 20 minutes of an interview that produced a result. The first 3 results are free with no card, and there are no seats and no minute quota. The full model is on [Tracy's pricing page](/pricing), and [AI interview cost](/blog/ai-interview-cost) works through more examples. Enterprise is priced to your volume and adds a dedicated customer manager, a dedicated support channel and a setup customised to your workflow in days. ## Cost at 100, 500 and 1,000 interviews a month: Tracy next to Notta Interviews of up to 20 minutes on Tracy. The columns buy different work: Tracy runs the interview, Notta transcribes one a person runs. Seat counts and minutes are our arithmetic at 20 minutes per interview, annual billing, with one seat per 1,800 minutes. | Interviews per month | Tracy runs the interview | Notta Business, 5 seats | Notta Pro seats to cover the minutes | | --- | --- | --- | --- | | 100 | $100 | $83.35 a month, unlimited transcription | 2 seats, $16.34 a month (2,000 minutes) | | 500 | $500 | $83.35 a month, unlimited transcription | 6 seats, $49.02 a month (10,000 minutes) | | 1,000 | $1,000 | $83.35 a month, unlimited transcription | 12 seats, $98.04 a month (20,000 minutes) | ![Tracy cost tiles: $100, $500 and $1,000 for 100, 500 and 1,000 interviews.](/blog/images/notta-pricing-1.png) *Illustration: what Tracy costs at three volumes, from $1 per credit with no seats. Illustration with sample data* 100, 500 and 1,000 interviews of up to 20 minutes cost $100, $500 and $1,000 on Tracy, with no seats. ## Is a per-minute price cheaper than an interview price? Notta is the lower number at every volume, because a transcription plan does not conduct the interview. Per transcribed minute Pro is about 0.45 cents ($8.17 over 1,800 minutes), our arithmetic, against $1 per started 20 minutes on Tracy. Tracy's credit pays for the interview, the follow-ups and the assessment; a Notta minute pays for the words of a call someone else ran. Our reasoning, not a Notta figure: seat and quota pricing wins when a fixed team runs a steady number of calls. Per-interview pricing wins when applicant volume swings and you do not want to staff it. Tracy bills nothing in a month with no interviews. To see your own cost per completed interview, run three interviews free and read the report. ## What to ask Notta before you buy If you are weighing Notta for interview transcripts, ask these before you sign: 1. Are the quotas of 1,800 minutes and 100 or 200 AI summaries per seat or per account? 2. What does a month look like when 500 interviews arrive and quotas do not carry over? 3. Does an applicant tracking system connect? The pricing page names CRM tools and Zapier, not an ATS. 4. What is the monthly-billing price in your currency, since the page opens on annual? 5. Which plan do you need for video of web meetings and audit logs? ## Beyond price: how the published workflows differ Source: Notta pricing page and Tracy product documentation. Where the pricing page is silent the cell says so. | Criterion | Tracy AI Interviewer | Notta | Who leads | | --- | --- | --- | --- | | Depth | Two-way video interview. After each answer she asks about what was said, in the person's own field, and a coding task can sit inside the interview. | Transcribes and summarises meetings your people run. Interviewing the candidate is not stated. | Different jobs: Tracy asks, Notta transcribes. | | Volume | Starts on click at any hour, no scheduling, no recruiter present. | Seats with monthly minute quotas; Business lists unlimited transcription and 200 AI summaries. | Tracy for applicant waves, Notta for steady team calls. | | Stack | API and MCP (ChatGPT, Claude, Mistral, Gemini, Copilot Studio), Slack, WhatsApp, Workable, Breezy HR, Zapier. | Zoom, Teams, Meet and Webex; Chrome extension; Salesforce, HubSpot, Zapier, Slack, Notion. ATS not named. | Even. Check your ATS with both. | | Evidence | Recording of both sides, timestamped transcript, written assessment per question, integrity signals shown to a person. | Time-stamped notes, speaker identification and AI summaries. | Tracy for interview evidence. | | Security | SOC 2 Type II attested; report on request. | Says it is SOC2 Type II and ISO 27001 certified; SAML SSO and audit logs on Enterprise. | Notta on the standards it lists. | | Multilingual | A spoken-language check scores fluency, coherence, grammar, pronunciation and vocabulary. | Transcript translation on Pro; translation add-ons. | Notta for translation of transcripts. | | Accessibility | Not stated. | Not stated on the pricing page. | Neither. Ask both. | | Validation and bias | Tracy publishes no bias audit. A person decides from the evidence report. | Not verified. | Not verified. | ## Is Notta worth it, and for whom? At volume the cost that matters is staffing: in one month Tracy did the work of 53 full-time recruiters for one customer. The verdict for a recruiter or agency: if people stay on the call, price Notta and take the low number. If the first round should run without anyone on the call, set Tracy's per-interview price beside it, at your own volume. ## Which path fits you? - Technical hiring: [technical interviews](/use-cases/technical-interviews) covers role-specific interviews with follow-ups. - High-volume hiring: [volume hiring](/use-cases/volume-hiring) covers applicant waves. - Integrations: [Connect](/connect) lists what Tracy connects to. - The report: [AI interview reports](/blog/ai-interview-reports) shows what a reviewer gets. ## Common questions ### How much does Notta cost? Notta lists Free at $0, Pro at $8.17 a month ($97.99 billed annually), Business at $16.67 a month ($199.99 billed annually) and a custom Enterprise plan from 51 seats. ### Does Notta charge per minute? Notta prices seats with monthly transcription minute quotas: 120 on Free and 1,800 on Pro, with unlimited transcription listed on Business. The quota does not carry over. ### How does Tracy's pricing compare with Notta's? Tracy charges $1 per credit, one credit per started 20 minutes, with the first 3 results free and no seats. The two buy different work: Tracy runs the interview, and Notta transcribes calls your team runs. ### Is there an alternative to Notta for interviews? If you want the first round run for you instead of transcribed, Tracy interviews each applicant and gives a reviewer the recording, transcript and assessment. --- Try three real interviews and calculate your cost per completed interview: send us one role, your volume, interview length and stack, and we will build the interview plan, with a recording and report you can check. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=notta-pricing&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=notta-pricing&utm_content=try-free) ## Sources - [Notta, Pricing](https://www.notta.ai/en/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Connect](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. # Otter.ai Alternatives for Interviews (Nov 2026) > Otter ai alternatives for interviews: Tracy runs the first round, while Fireflies, Granola, Metaview and BrightHire take notes. Compared on four pillars. - URL: https://tracyinterview.ai/compare/otter-ai-alternatives - Published: 2026-11-10 - Author: Tracy team > **In short** When the first round has to be run, not only recorded, Tracy AI Interviewer runs it by two-way video, follows up in the person's own field and gives a reviewer a recording, transcript and written assessment per question. > > $1 per credit, the first 3 results free, no seats. > > Otter is the better pick when your own interviewers keep running the calls and you want live transcripts and summaries at a low seat price. > > Next: create one interview for the workflow your current tool misses. Teams look for Otter.ai alternatives when a general meeting notetaker stops matching the interview job. Otter records and summarizes calls your interviewers run. Running the first round for every applicant is a different job, and Tracy AI Interviewer is the first tool to try for it. The list puts Tracy first, then four tools for other needs, each with who it suits and where it falls short. ## Why do teams look for Otter.ai alternatives? These reasons come from Otter's own pages. The use cases on its home page are Sales, Education and Media. A hiring use case is not stated there or on the pricing page, so a hiring team adapts a general tool. The output is a summary of a call. Notes are only as deep as the questions your interviewer asked, and Otter's pages describe notes, summaries and action items; an AI that interviews the candidate is not stated. Spend follows seats. Pro lists $16.99 per user per month and Business $24 on monthly billing ($8.33 and $19.99 on annual billing), so cost tracks how many interviewers you have, not how many interviews you run. The stack is built for sales and meetings: Salesforce, HubSpot and Zapier on Pro, an MCP server, and an API on Enterprise. An ATS integration is not stated on those pages. Some teams should stay. If your interviewers run the calls and you want live transcription, speaker recognition and summaries, Otter does that, and an AI interviewer does not replace it. ## Which Otter.ai alternative suits which team? - Tracy AI Interviewer: roles where the first round must show depth for every applicant, at a published price per result. - Fireflies.ai: hiring teams that want notes, scorecards and ATS sync on interviews their own people run. - Granola: interviewers who want a private notepad and no bot on the call. - Metaview: teams that want interview notes mapped into ATS scorecards. - BrightHire: teams that want a notetaker and an AI interviewer from one vendor, and accept a quote. The sections below give the detail and the limits of each. ## Tracy AI Interviewer: the first round, run and evidenced Tracy interviews every applicant by two-way video. After each answer she asks about what was said, in the person's own field, and waits for the answer. A Director at a top-tier early-stage VC said Tracy "sometimes brings up exactly those questions I might have forgotten to ask." You can bring your own questions or describe the role and let Tracy draft them, and a coding task can sit inside the interview for technical roles (see [AI technical interviews](/blog/ai-technical-interviews)). Behind every result sit a recording of both sides, a timestamped transcript, a written assessment per question with its reasoning, and integrity signals shown to a human reviewer. A person decides. Each applicant starts on click at any hour with no scheduling, and Tracy connects through API and MCP (ChatGPT, Claude, Mistral, Gemini, Copilot Studio), Slack, WhatsApp, Workable, Breezy HR and Zapier. The price is $1 per credit, one credit per started 20 minutes, the first 3 results free, no seats. See [Tracy's pricing](/pricing). Where Otter is better: it transcribes and summarizes the calls your team runs, live, at a low per-user price. Tracy is not a notetaker for your calls. ![Sample Tracy exchange: a missed-quota question, the candidate's answer, then a follow-up at 05:01.](/blog/images/otter-ai-alternatives-1.png) *A person reviews the recording, transcript and written assessment per question. Illustration with sample data* Tracy asks the next question from the candidate's answer. A recording and a timestamped transcript sit behind every result, plus a written assessment per question. ## Fireflies.ai: a notetaker with a recruiting page Fireflies sends a note taker bot to Zoom, Google Meet, Microsoft Teams and more, and paid plans can record through a Chrome extension without the bot. Its recruiting page covers interview and screening-call notes, scorecards and ATS sync, and notes can go to Greenhouse, Lever and others. It transcribes in 100+ languages. Pro lists $10 and Business $19 per seat per month billed annually ($18 and $29 on monthly billing). Pick it when hiring-specific notes on your own interviews are the need. ## Granola: private notes with no bot Granola takes notes without a meeting bot, using your computer's audio, on macOS, Windows, iOS and Android, and works with Zoom, Google Meet and Teams. Notes are private by default and audio is deleted after transcription. Basic is free and Business lists $14 per user per month. Hiring-specific features are not stated on its home or pricing pages. Pick it when one interviewer wants a quiet notepad, not a team workflow. ## Metaview: notes mapped into ATS scorecards Coming from Otter for interviews, Metaview's Notetaker is the interview-specific notetaker: it captures video, phone and in-person conversations and maps answers into scorecards in Greenhouse, Ashby, Lever and Workday. Notetaker Pro lists $60 per user per month on monthly billing; Screening has no price on its pricing page, and the all-agents plan is custom-priced. ## BrightHire: notes and an AI interviewer from one vendor BrightHire's own pages pair AI notes on live interviews with a two-way AI interviewer candidates can take 24/7, so a team leaving Otter can keep both under one vendor. Its pricing page lists no dollar amounts (9 Nov snapshot). ## Otter.ai and its alternatives compared on four pillars and price Source: each vendor's own pages and approved Tracy product records, Nov 2026. Prices are list prices; billing basis is shown per cell. | Tool | Depth | Volume | Stack | Evidence | Price | | --- | --- | --- | --- | --- | --- | | Tracy AI Interviewer | Two-way video, follow-ups in the person's field, coding task inside | Starts on click, any hour, no scheduling | API, MCP, Slack, WhatsApp, Workable, Zapier | Recording, timestamped transcript, assessment per question | $1 per credit, first 3 free | | Otter.ai | Notes and summaries of calls your team runs | Joins Zoom, Teams and Meet, or records bot-free | Salesforce, HubSpot, Zapier, MCP; ATS not stated | Live transcript, speaker recognition, summaries | Pro $16.99 and Business $24 per user per month on monthly billing; $8.33 and $19.99 on annual | | Fireflies.ai | Notes, scorecards and candidate-scoring skills on recorded calls | Bot on the call, or Chrome extension without it | Greenhouse, Lever, CRMs, Slack, MCP | Transcript in 100+ languages, summaries | Pro $18 and Business $29 per seat per month on monthly billing; $10 and $19 on annual | | Granola | Private notes on calls you run; hiring features not stated | No bot; uses your computer audio | Zapier, Attio, Notion, HubSpot, MCP, API | Transcript and notes; audio deleted after transcription | Free; Business $14 per user per month (billing basis not stated) | | Metaview | Notetaker on human interviews; Screening is a separate agent | Notetaker can work without a bot in the call | Greenhouse, Ashby, Lever, Workday and more | Notes mapped into ATS scorecards | Notetaker Pro $60 per user per month, monthly; Screening price not stated | | BrightHire | AI notes on live interviews; AI interviewer, two-way | Candidates can interview 24/7 | Greenhouse, Lever, Ashby, Workday, Slack | Auto-filled scorecards; rubric scoring | No dollar amounts on its pricing page (9 Nov snapshot) | ![Three options: your team interviews, Tracy interviews, or both, with notes on later rounds.](/blog/images/otter-ai-alternatives-2.png) *Illustration, no data: who runs the interview decides the kind of tool. Illustration, no data.* One question sorts the shortlist: does your team run the interview, or does an AI interviewer run it? ## How do you choose an Otter.ai alternative? Start with who asks the questions. If your interviewers run the calls and you want transcripts and summaries, stay with Otter or look at Fireflies.ai, Granola or Metaview. If you want the first round run for every applicant, shortlist Tracy and BrightHire's AI interviewer. Both: keep Otter for later rounds and let Tracy run round one. Then compare the cost model. Seats scale with interviewers, so ten interviewers on a $16.99 monthly plan cost the same in a quiet month and a busy one. Tracy's credits scale with interviews: 100 interviews of 20 minutes is $100. That is a different job priced differently, so compare it against the recruiter hours the first round takes, not against a seat. Finally, run one real role. Read the second question the AI asks and open the recording behind one result. Create one interview for your role, and read [what an AI interview report contains](/blog/ai-interview-reports) and [the best AI interview software](/blog/best-ai-interview-software) for the wider list. ## Which path fits you? - Technical hiring: [technical interviews](/use-cases/technical-interviews) covers role-specific interviews with follow-ups. - High-volume hiring: [volume hiring](/use-cases/volume-hiring) covers applicant waves. - Integrations: [Connect](/connect) lists the integrations. ## Common questions ### What is the best Otter.ai alternative for interviews? It depends on the job. Tracy AI Interviewer suits teams that want the first round run for every applicant, with a recording and report to check. Fireflies.ai and Metaview fit teams that keep running interviews and want hiring notes with ATS links. ### Is there a free Otter.ai alternative? Yes, for notes. Fireflies.ai and Granola each list a free plan, as does Otter. Tracy's first 3 results are free, then $1 per credit. ### Is Tracy a notetaker like Otter? No. Tracy runs the interview and writes the report. For notes on calls your team runs, use a notetaker. ### Which alternative skips the bot on the call? Granola takes notes without a bot, and Otter and Metaview can record without one. With Tracy there is no call to join: the applicant opens a link in a browser. --- Create one interview for the workflow your current tool misses: send us one role and we will build the interview plan, with a recording and report you can check. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=otter-ai-alternatives&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=otter-ai-alternatives&utm_content=try-free) ## Sources - [Otter.ai, Pricing](https://otter.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Connect](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. # Read AI Pricing: Cost vs Tracy (Nov 2026) > Read AI pricing in Nov 2026: Pro is $15 per user per month on annual billing. Tracy's AI interview is $1 per credit with no seats. - URL: https://tracyinterview.ai/compare/read-ai-pricing - Published: 2026-11-10 - Author: Tracy team > **In short** Tracy AI Interviewer runs the first-round interview for each applicant and is priced per completed interview: $1 per credit, one credit per started 20 minutes, the first 3 results free, no seats. > > Each result comes with a recording, a timestamped transcript and a written assessment per question, and a person decides. > > Read AI lists Pro at $15 per user per month on annual billing. If your own team keeps running the interviews, Read is the cheaper pick. > > Next: try three interviews free and calculate your cost per completed interview. Read AI pricing is public for every self-serve plan. Free, Pro, Enterprise and Enterprise+ are priced per user per month, and the page lists an annual and a monthly rate for each. Read AI is a meeting notetaker, so the price buys notes on calls your people hold. Running the candidate interview is not stated on its pricing page. **Choose Read AI when** your recruiters and hiring managers keep running the interviews and you want a summary, transcript and action items from each call, at $15 per user per month on annual billing, or free for five meetings a month. **Choose Tracy AI Interviewer when** you want the first-round interview run for every applicant, at a price you can compute before any call: $1 per credit, no seats. She starts on click at any hour, with no scheduling and no recruiter present, and the cost follows interviews, not headcount. ## Read AI pricing table: every plan the page lists Source: Read AI pricing page, 9 Nov 2026. Prices are Read AI list prices in USD, the annual rate first. Enterprise is the name Read AI gives its second paid plan. | Plan | Price per user per month (annual / monthly) | What it lists | Page words | | --- | --- | --- | --- | | Free | $0 | 5 meeting transcripts per month; recordings up to 1 hour | "5 meeting transcripts per month" | | Pro | $15 billed annually, $19.75 monthly | Unlimited meeting transcripts; premium integrations; recordings up to 4 hours | "Unlimited meeting transcripts" | | Enterprise | $22.50 billed annually, $29.75 monthly | Pro plus audio and video playback, video highlights, premium support | "Audio & video playback" | | Enterprise+ | $29.75 billed annually, $39.75 monthly | Enterprise plus HIPAA compliance, SAML and SCIM, custom data retention, EU and CA hosting; recordings up to 8 hours Requires 5 or more licenses. | "REQUIRES 5+ LICENSES" | ## What limits and extras apply on Read AI Licenses are individual: a workspace needs one per invited person, and every license in it must be on the same plan, so a mixed team of Pro and Enterprise seats is not possible. Read AI discounts by annual license count: 10% at 100 or more, 15% at 500 or more, 20% at 1,000 or more. It says it does not offer refunds, though you can cancel before the renewal date. Under every plan the page also lists "30 meetings measured per month" and does not say what is measured. The same plans list unlimited transcripts and reports, so ask what the figure limits before you rely on it for a busy interview season. ## What a typical team pays for Read AI Take a team where five people run interviews and all five need Read. Five Pro licenses at $15 are $75 a month on annual billing, or $900 over twelve months. On monthly billing at $19.75 it is $98.75 a month. A team of 20 interviewers is $300 a month on annual billing. This is our arithmetic from list prices, not a Read AI quote. Because Pro lists unlimited transcripts, 100 calls and 1,000 calls cost the same for those five seats. ![Three steps: applicant clicks, Tracy interviews, you review the result.](/blog/images/read-ai-pricing-2.png) *Illustration: how Tracy takes the first round off your team. The applicant starts on click, she follows up on their answers, and a reviewer gets the recording, transcript and a written assessment per question. Illustration with sample data* Tracy runs the first round, and a person reviews the result and decides. ## What Tracy costs at 100, 500 and 1,000 interviews A credit costs $1 and covers each started 20 minutes of an interview that produced a result. The first 3 results are free with no card, and there are no seats. The full model is on [Tracy's pricing page](/pricing), and [AI interview cost](/blog/ai-interview-cost) works through more examples. Enterprise is priced to your volume and adds a dedicated customer manager, a dedicated support channel and a setup customised to your workflow in days. ## Cost at 100, 500 and 1,000 interviews a month: Tracy next to Read AI Pro Interviews of up to 20 minutes on Tracy; one over 20 minutes costs one more credit per started 20 minutes. Read AI seats are our assumption, rising with volume. The columns buy different work: Tracy runs the interview, Read AI records one a person runs. Hours are our arithmetic at 20 minutes per interview. | Interviews per month | Tracy runs the interview | Read AI Pro, people run the calls | Interviewer time if people run the calls | | --- | --- | --- | --- | | 100 | $100 | $30 a month, 2 interviewers on annual billing | about 33 hours | | 500 | $500 | $75 a month, 5 interviewers on annual billing | about 167 hours | | 1,000 | $1,000 | $150 a month, 10 interviewers on annual billing | about 333 hours | ![Tracy cost tiles: $100, $500 and $1,000 for 100, 500 and 1,000 interviews, plus first 3 results free.](/blog/images/read-ai-pricing-1.png) *Illustration: what Tracy costs at three volumes, from $1 per credit with no seats. Illustration with sample data* 100, 500 and 1,000 interviews of up to 20 minutes cost $100, $500 and $1,000 on Tracy. ## Is a seat price cheaper than an interview price? Read AI's number is lower at every volume in the table, and that is the honest read: a seat price is low because the page describes notes, not an interviewer. What the table leaves out is people time. 1,000 interviews of 20 minutes is about 333 hours of interviewer time, our arithmetic, before scheduling and note review. Our reasoning, not a Read AI figure: seat pricing wins when the same few people run a steady number of calls. Per-interview pricing wins when applicants arrive in waves and you do not want to staff them. Tracy bills nothing in a month with no interviews, because there are no seats and no upfront cost. To see your own cost per completed interview, run three interviews free and read the report. ## What to ask Read AI before you buy If you are weighing Read AI for interview notes, ask these before you sign: 1. What does "30 meetings measured per month" limit, and what happens after it? 2. Which plan do you need to keep audio and video playback of interviews? 3. Does an applicant tracking system connect? The pricing page lists Notion, Salesforce, HubSpot, Jira, Confluence, Zapier and webhooks. 4. What are the annual terms, given that refunds are not offered? 5. At what license count does the volume discount start for your team? ## Beyond price: how the published workflows differ Source: Read AI pricing page and Tracy product documentation. Where the pricing page is silent the cell says so. | Criterion | Tracy AI Interviewer | Read AI | Who leads | | --- | --- | --- | --- | | Depth | Two-way video interview. After each answer she asks about what was said, in the person's own field, and a coding task can sit inside the interview. | Notes on meetings your people run: summary, topics, action items and key questions. Interviewing the candidate is not stated. | Different jobs: Tracy asks, Read records. | | Volume | Starts on click at any hour, no scheduling, no recruiter present. | One license per person; unlimited transcripts on paid plans; recordings up to 4 hours on Pro. | Tracy for applicant waves, Read for steady team calls. | | Stack | API and MCP (ChatGPT, Claude, Mistral, Gemini, Copilot Studio), Slack, WhatsApp, Workable, Breezy HR, Zapier. | Zoom, Teams and Meet; Notion, Salesforce, HubSpot, Jira, Confluence, Zapier, webhooks; announces a Claude connector, ChatGPT app and MCP server. ATS not named. | Even. Check your ATS with both. | | Evidence | Recording of both sides, timestamped transcript, written assessment per question, integrity signals shown to a person. | Transcript by speaker and summary; audio and video playback on Enterprise and Enterprise+. | Tracy for interview evidence, Read for meeting playback. | | Security | SOC 2 Type II attested; report on request. | HIPAA, SAML and SCIM on Enterprise+. SOC 2 not stated on the pricing page. | Read on the controls listed. | | Multilingual | A spoken-language check scores fluency, coherence, grammar, pronunciation and vocabulary. | Reports in 16+ languages. | Not comparable: different outputs. | | Accessibility | Not stated. | Not stated on the pricing page. | Neither. Ask both. | | Validation and bias | Tracy publishes no bias audit. A person decides from the evidence report. | Not verified. | Not verified. | ## Is Read AI worth it, and for whom? On speed, not price: one customer cut its average time to hire from 44 days to 7, and Tracy did the work of 53 full-time recruiters in one month. The verdict for a budget owner: if the plan is to keep people on the call, price Read AI by seat and enjoy the low number. If the plan is to have the first round run for you, set Tracy's per-interview price beside it, at your own volume. ## Which path fits you? - Technical hiring: [technical interviews](/use-cases/technical-interviews) covers role-specific interviews with follow-ups. - High-volume hiring: [volume hiring](/use-cases/volume-hiring) covers applicant waves. - Integrations: [Connect](/connect) lists what Tracy connects to. - The report: [AI interview reports](/blog/ai-interview-reports) shows what a reviewer gets. ## Common questions ### How much does Read AI cost? Read AI lists Free at $0 for 5 meeting transcripts a month, Pro at $15 per user per month on annual billing ($19.75 monthly), Enterprise at $22.50 ($29.75 monthly) and Enterprise+ at $29.75 ($39.75 monthly). ### Is Read AI free? Yes, up to 5 meeting transcripts a month on the Free plan, with no credit card. ### How does Tracy's pricing compare with Read AI's? Tracy charges $1 per credit, one credit per started 20 minutes, with the first 3 results free and no seats. The two buy different work: Tracy runs the interview, and Read AI takes notes on calls your team runs. ### Is Read AI an alternative to an AI interviewer? The pricing page describes meeting notes, transcripts and summaries. Running the candidate interview is not stated. If you want the first round run for you, that is what Tracy does. --- Try three real interviews and calculate your cost per completed interview: send us one role, your volume, interview length and stack, and we will build the interview plan, with a recording and report you can check. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=read-ai-pricing&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=read-ai-pricing&utm_content=try-free) ## Sources - [Read AI, Plans and pricing](https://www.read.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Connect](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. # Tracy vs Eightfold AI Interviewer (Nov 2026) > Eightfold AI Interviewer vs Tracy for technical hiring: where each wins on depth, volume, stack and evidence, plus governance, languages and price. - URL: https://tracyinterview.ai/compare/tracy-vs-eightfold - Published: 2026-11-10 - Author: Tracy team > **In short** Tracy AI Interviewer gives a technical team a first round with follow-ups in the person's own field, an optional coding task, and a recording, transcript and assessment per question for a person to decide from. > > She fits your stack through API, MCP, Slack and Zapier, starts free (first 3 results), then costs $1 per credit, no seats. > > Eightfold AI Interviewer is the stronger enterprise program: published bias audit, FedRAMP, 24+ languages, Oracle Recruiting. Choose it when you need those. > > Next: test both on one role. This Eightfold AI Interviewer comparison is for technical teams that hire at volume. Eightfold calls AI Interviewer an AI-native digital worker that conducts "bias-conscious interviews at scale". Tracy AI Interviewer runs a two-way video interview, follows up on what the person says and returns a recording, a timestamped transcript and a written assessment per question. Eightfold wins on governance, languages and Oracle fit. Tracy wins on a published price, a free start and connections to the tools a technical team already uses. ## What does Eightfold AI Interviewer do well? Eightfold publishes an independent audit of AI Interviewer under New York City Local Law 144, lists FedRAMP Moderate, SOC 2 Type II and four ISO certifications, and says it evaluates candidates in 24+ languages. Its 360 Interview runs a first call, a coding exercise, a case and panel questions in one session. It flags tab switching, camera manipulation and duplicate voices for human review, offers an optional ID check through ID.me or CLEAR, and says no application is auto-rejected. It works beside an existing ATS and is integrated inside Oracle Recruiting. ## Eightfold AI Interviewer vs Tracy: buyer criteria side by side Buyer criteria for Eightfold AI Interviewer and Tracy, Nov 2026. Not verified means a point to put to the vendor. | Criterion | Tracy AI Interviewer | Eightfold AI Interviewer | Who leads | | --- | --- | --- | --- | | Interview depth | Two-way video interview. Follow-ups in the person's own field, a coding task inside the interview, questions from your brief or your own list. | Follow-up when an answer is incomplete. Questions generated from a job description you approve. 360 Interview adds coding, case and panel questions, by voice. | Level on follow-ups. Eightfold on one session with coding and case. | | Volume and start | Every applicant starts on click, at any hour, with no scheduling. Ready in under 10 seconds. | Starts when a candidate applies, at any hour. Claims over one million interviews per hour. | Level at the start. Eightfold on the larger throughput claim. | | ATS and stack | API and MCP (ChatGPT, Claude, Mistral, Gemini, Copilot Studio), Slack, WhatsApp, Zapier; beside Workable and Breezy HR. | Invite link or ATS. Inside Oracle Recruiting. API or MCP route: Not verified. | Eightfold for Oracle. Tracy for API, MCP and chat tools. | | Evidence per result | Recording of both sides, timestamped transcript, assessment per question with reasoning. | Summaries, recordings, transcripts, replay, notes and override, with reasoning per section. | Level. | | Integrity signals | Tab changes, gaze, face detection, answer delay, fullscreen exits and external screens, as evidence for a human to review. | Tab switching, camera manipulation and duplicate voices flagged for review. Optional ID.me or CLEAR check. | Eightfold on ID-check partners. | | Implementation and support | Enterprise adds a dedicated customer manager and support channel, with setup customised in days. The free start needs no sales call. | Says teams are up and running in weeks, with a 1 to 2 week activation. Support terms: Not verified. | Level on speed. | | Security and compliance | SOC 2 Type II attested. | FedRAMP Moderate, SOC 2 Type II and four ISO certifications. | Eightfold. | | Accessibility and languages | Spoken-language check beside the transcript. Accessibility accommodations: Not stated. | 24+ languages, with CEFR-based language assessment. Candidates with accessibility needs go to a human-led process. | Eightfold. | | Validation and bias | A person reviews the evidence and decides. Tracy publishes no bias audit. | Independent BABL AI audit under NYC Local Law 144, audited 29 June 2026. | Eightfold. | | Reporting and audit trail | A report per interview: recording of both sides, transcript and an assessment per question with reasoning. Package by link or PDF. | Section-by-section readouts with reasoning; recruiters add notes and override. | Level. Both give a reviewer the evidence and leave the call to a person. | | Price and first-year cost | $1 per credit, one per started 20 minutes. First 3 results free, no seats. | No public price. Demo route, plus an Oracle Cloud Marketplace listing. | Tracy for a published price. | ## How does Eightfold AI Interviewer work? How the published workflows differ Eightfold describes three steps: invite candidates, interview at scale, make your decision. For the same technical role the published workflows differ as shown. The Tracy column is an illustration. Tracy asks about what the person just said, in their field. A Director at a top-tier early-stage VC says Tracy "sometimes brings up exactly those questions I might have forgotten to ask, which has been a pleasant surprise." ![Recreated Tracy screen: a follow-up asks how to stop a retry from charging a customer twice.](/blog/images/tracy-vs-eightfold-1.png) *Tracy asks a follow-up on the detail the candidate skipped, so a reviewer sees how they reason. Illustration with sample data* At 11:20 Tracy asks about the detail the candidate skipped, and the answer is recorded with a timestamp. How the published workflows differ for the same technical role. | Step | Eightfold AI Interviewer (from its pages) | Tracy (illustration) | | --- | --- | --- | | Setup and invite | Questions and criteria are generated from a job description for you to approve. Send an invite link or connect your ATS. | Describe the role and Tracy drafts the questions, or bring your own. Share one link; each person starts in the browser with no account. | | Interview | A conversational interview with a follow-up when an answer is incomplete. Engineering interviews include a coding round. | She asks about what the person just said, in their field, and waits for the answer. A coding task can sit inside the interview. | | Review | Summaries and recordings, with replay, transcripts, notes and override. | Recording of both sides, transcript and an assessment per question with reasoning. | | Decision | Eightfold says AI surfaces the signal and people make the call. | A person decides: they can advance a candidate or invite them again. | ## Can a reviewer check an Eightfold or a Tracy result? Both tools hand a reviewer a recording and reasoning behind each assessment. Integrity signals work the same way in both: evidence for a person, never a verdict, with Tracy telling the person on the join screen that the conversation is recorded. See [AI interview reports](/blog/ai-interview-reports) and [how cheating detection works](/blog/ai-interview-cheating-detection). ## Does Eightfold or Tracy fit your stack? Tracy connects through API and MCP, Slack, WhatsApp and Zapier, and works beside the Workable and Breezy HR connectors, with no write-back stated. An interview starts from an invite link, an apply page or your own system, and Enterprise teams get a setup customised to their workflow. The [Connect page](/connect) lists each guide, and [ATS AI interviewer vs dedicated](/blog/ats-ai-interviewer-vs-dedicated) covers when a standalone interviewer wins. Eightfold sits inside Oracle Recruiting; an Oracle connector is not stated on Tracy's Connect page. An API or MCP route for AI Interviewer is not stated on Eightfold's pages, so ask in the demo. ## What do Eightfold and Tracy cost? Eightfold publishes no price. Its product pages route to a demo, and it is listed on the Oracle Cloud Marketplace. Tracy's price is on [her pricing page](/pricing): $1 per credit, one credit per started 20 minutes, the first 3 results free, no seats, and 100 interviews of up to 20 minutes for $100. Our reasoning, not an Eightfold price: each $1,000 of first-year quote equals 1,000 Tracy interviews of up to 20 minutes. ![Four points: Eightfold for Oracle, languages and a bias audit; Tracy for price, a free start and connections.](/blog/images/tracy-vs-eightfold-2.png) *Where each tool fits a first round: Tracy for a published price, a free start and connections to your tools; Eightfold for Oracle, languages and a bias audit. Source: vendor pages, Nov 2026.* Eightfold fits Oracle, language and audit needs. Tracy fits a published price, a free start and connections to your own tools. ## How do you test both before you commit? Run one technical role for two weeks: 1. Ask Eightfold for a live run on that role and a written first-year total. 2. In Tracy, describe the role and edit the drafted questions. The first 3 results are free. 3. Have one reviewer read the next 20 reports and count the follow-ups that landed and the cost per interview. For the mechanics first, read [how AI interviews work](/blog/how-do-ai-interviews-work). ## Which path fits you? - Technical roles: [technical interviews](/use-cases/technical-interviews). - Hundreds of applicants: [volume hiring](/use-cases/volume-hiring). - Integrations and price: [Connect](/connect) and [pricing](/pricing). ## Common questions ### How does Eightfold AI work? Eightfold's AI Interviewer runs on its Talent Intelligence engine. A recruiter sends an invite link or connects an ATS, the candidate completes a conversational interview, and the recruiter gets summaries, recordings and scored evaluations. Eightfold says people make the call. ### Does Eightfold AI Interviewer ask follow-up questions? Yes. Eightfold says the model asks a follow-up when a candidate did not fully answer a question. ### Is Eightfold AI Interviewer audited for bias? Eightfold publishes an audit by BABL AI under NYC Local Law 144. Tracy publishes no bias audit; a person decides from her evidence report. ### Can I try Tracy before talking to sales? Yes. The first 3 results are free and no card is needed. --- Send us one role and we build the interview plan, with a recording and report you can check. First 3 results are free. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=tracy-vs-eightfold&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=tracy-vs-eightfold&utm_content=try-free) ## Sources - [Eightfold AI, AI Interviewer](https://eightfold.ai/products/ai-interviewer/), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Connect](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. - [Tracy AI Interviewer, home page](https://tracyinterview.ai/), retrieved 9 Nov 2026. # AI Interview Tools for Accelerators and Fellowships (Nov 2026) > Accelerator application software compared: Tracy, Interviewer.AI, Willo, Spark Hire and VideoAsk on follow-ups, volume, stack, evidence and price. - URL: https://tracyinterview.ai/compare/ai-interview-tools-for-accelerators - Published: 2026-11-09 - Author: Tracy team > **In short** Tracy AI Interviewer is the pick when partner hours cap how many founders you meet. Every applicant gets a first interview in the browser, she asks the follow-up a partner would ask, and you read a recording, a transcript and an assessment per question instead of a form alone. > > One link goes in your application form and you can bring your own diligence questions. Pricing is public: $1 per credit, the first 3 results free, no seats. A partner still decides who moves on. > > Interviewer.AI is the closest tool built for admissions and programmes. Willo charges per live role, and Spark Hire and VideoAsk record answers, with no AI follow-up described. A short form is enough for a small cohort. A programme opens applications on a Monday, and by Friday the inbox holds more founders than the partners have hours to meet. The ones with a polished form get the calls. Accelerator application software has to fix that step, and a form cannot ask a founder to explain a number. Tracy AI Interviewer gives every applicant a first interview in the browser, follows up on what the founder says, and hands the partner a recording, a transcript and a written assessment per question. This page ranks five tools on what their own pages say. Our [applications page](/use-cases/applications) shows the flow from form to report. ## What does accelerator application software need to do? The people who decide are a programme lead, a selection partner or a managing partner. Partner hours cap how many founders are met, and a form cannot show what a founder is building toward. Six needs follow, and they are the ranking criteria: 1. Every founder gets a first conversation. 2. The interview asks a follow-up on what the founder said. 3. It runs from your form with one link and no login. 4. It takes your own diligence questions. 5. A partner can read the result in minutes and see the words behind it. 6. It costs little in a rush and nothing when quiet. For the interview side, see our guide to the [accelerator interview](/blog/accelerator-interview). ## How do the five tools rank, and why in this order? We ranked on how much of the six needs each tool covers, with a published price as the tie-breaker. Tracy is first because she covers all six. She is not first on everything: Interviewer.AI names admissions outright and has a branded portal, and for one open call of up to 150 candidates Willo's £49 per live role is lower than 150 Tracy interviews at $1 each. Where partners can meet every founder in a small cohort, keep the form and skip the tool. ## Which accelerator application software is the better pick, and when? Tracy gives every applicant a two-way video interview that starts when they click the link, at any hour, in the browser, with no account. After each answer she asks about what was said, and the founder has room to describe what they are building and to ask their own questions. You put one link in your form ([product page](/product)), and bring your own diligence questions or let Tracy draft them. The partner gets a recording of both sides, a timestamped transcript and a written assessment per question with its reasoning. Partners spend their hours on the strongest founders, because the first conversation already happened ([how reports read](/blog/ai-interview-reports)). Tracy costs $1 per credit, one credit per started 20 minutes, with the first 3 results free, no seats and no upfront cost ([pricing](/pricing)). Her public pages describe the interview and its report, not an application portal or reviewer panels, so keep your form and add the link. Tracy has completed 600K+ interviews for 300+ clients, with 99% interview completion. ![Recreated Tracy application page: a founder recording beside Tracy's question on pilot renewals and a short company form.](/blog/images/ai-interview-tools-for-accelerators-1.png) *The founder answers a question about pilot renewals while a short form sits beside the recording. Source: Tracy visual library, founder-pitch screen, fixture ai-interview-tools-for-accelerators-1. Illustration with sample data* A founder pitches to Tracy from the application page, and she asks about what they said. ## What did a residency programme change after switching? A director at an accelerator residency programme describes the before and after in their own words. They started with 30-minute one-on-one Zoom interviews, once scheduled eight back-to-back in a day: "I was effectively spending around six hours a day on screening." An earlier AI tool "still took about an hour to go through their answers." With Tracy: "I can now review and assess those candidates in approximately 1.5 hours. That's a time saving of more than 50%, and in practice, closer to 75%." For borderline candidates, Tracy makes it "much easier to share their responses and assessment results with other team members for further discussion." Their advice: "spend time designing the depth and structure of your interview questions during the initial setup." ### 2. Interviewer.AI: admissions and programmes in one product Interviewer.AI says universities, bootcamps and training programs use it for admissions. It lists a branded application portal, a shortlist to share with a panel, and fraud checks that flag more than one voice, lip-sync anomalies and screen or prompt assistance. Essential is $199 a month with 50 credits and one-way video; Professional is $399 a month with 120 credits and conversational AI avatar interviews with adaptive follow-ups. One credit buys a one-way interview and two buy a conversational one, so Professional covers up to 60 a month. A 14-day trial is free. Pick it when you want the portal, shortlist and admissions panel in one product on a monthly plan. ### 3. Willo: pay per live role Willo's Lite plan is £49 per live role with no contract, including 150 assessed candidates per live role. Questions can be video, audio, written, multiple choice or identity checks. Enterprise starts at £2,999 a year. It lists follow-up questions without saying how they work. Pick it for a programme with one open call a cohort. ### 4. Spark Hire: recorded and live video at a published price Spark Hire starts at $249 a month billed annually, with unlimited live and one-way video interviews and users, and 40+ ATS integrations. Founders record answers to set questions and a person reviews them, with no follow-up described. Pick it for a flat price and a human reading every submission. ### 5. VideoAsk: a video question inside a form flow VideoAsk, from Typeform, offers video forms and asynchronous interviews, starts free with no credit card, and connects to 1,500+ apps through Zapier plus an API. It describes no AI follow-up. Pick it when you want a recorded video question added to a form at no cost to start. ## How do the five tools compare on depth, volume, stack and evidence? | Tool | Depth | Volume | Stack | Evidence | Price | | --- | --- | --- | --- | --- | --- | | 1. Tracy AI Interviewer | Two-way video, follow-ups on what the founder says | Every applicant starts on click, any hour | One link in your form; API, MCP, Zapier | Recording, transcript, assessment per question | $1 per credit, first 3 results free | | 2. Interviewer.AI | Adaptive follow-ups on Professional | Credits: 50 a month on Essential, 120 on Professional | Branded application portal; shortlist to share | Transcripts and summaries; fraud flags for review | $199 or $399 a month; 2 credits per conversational interview | | 3. Willo | Structured questions; follow-up questions listed | 150 assessed candidates per live role on Lite | 50+ native integrations | Custom scorecards | £49 per live role; Enterprise from £2,999 a year | | 4. Spark Hire | Recorded and live video; no follow-up described | Unlimited interviews and users | 40+ ATS integrations | Recorded answers and AI scoring | From $249 a month, billed annually | | 5. VideoAsk | Video forms, async interviews; owner can reply by video | Not stated | 1,500+ apps through Zapier; API | Not stated | Free to start | ## What do the six needs leave out? Rival facts: vendors' own pages, read 9 Nov 2026. Tracy column: approved Tracy product records. "Not stated" and "Not verified" mean the page does not say or we have not checked, not that it is absent. | Criterion | What we verified | Where a rival leads | | --- | --- | --- | | Implementation and service | Tracy: one link in your form, and Enterprise adds a dedicated customer manager, a setup customised to your workflow and a team walkthrough, customised in days. | Interviewer.AI on the portal; Tracy on speed of customisation. | | Security and compliance | Tracy is SOC 2 Type II attested, with the report on request. Willo lists ISO 27001. The rest: not verified here. | Level between Tracy and Willo, on different standards. | | ATS integration depth | Spark Hire lists 40+ ATS integrations and Willo 50+ native integrations. VideoAsk reaches 1,500+ apps through Zapier. | Willo and Spark Hire. | ## What does a first interview have to capture that a form cannot? Programmes say what they look for. Techstars says its staff personally review every application and that in the interview it wants to understand a company well enough to write an investment memo; its Managing Director tells founders to be ready on the customer and the pain, speed and execution, market size, and what is different. YC runs 10-minute interviews with 2 to 3 partners who have read the application. A form cannot ask a founder to explain a number that looks odd. That follow-up is the job of the first interview, and Tracy takes your own questions, so you can load Techstars' four as they stand. See [one-way versus two-way video interviews](/blog/one-way-vs-two-way-video-interviews). ![Two panels. A form or recorded answers cannot ask a follow-up; a two-way first interview does, for every founder.](/blog/images/ai-interview-tools-for-accelerators-2.png) *What changes when the first step is a two-way interview instead of a form or recorded answers. Source: Tracy product documentation.* A form or recorded answers cannot ask a follow-up, while a two-way first interview does and reaches every founder. ## Will founders find an AI interview impersonal? Some partners say founders should be met by a person. In Tracy's design she is the first conversation for everyone, so partners spend their hours on the best founders. She asks the follow-up a partner would ask and lets founders ask their own questions. Founders are told on the join screen that the conversation is recorded and transcribed. Candidates rate Tracy 92 on NPS, a rating from candidates, not a founder survey. A person decides who moves on, and the report links each point to the founder's words. ## How do you pilot one for your next cohort? Run one cohort window with one partner as owner, and decide the stop condition before you start. 1. Put the link in your form for one programme or track, and load your own questions. 2. Have the owner read the reports for a sample of founders and compare them with who the partners would have picked from the form alone. 3. Stop if the owner would not change the shortlist after reading five reports. 4. Read founders who finished, partner minutes per applicant, and founders no partner would have met. At 20 minutes each, 100 interviews cost $100 on Tracy, with the first 3 results free; try one on the [Connect page](/connect). For the wider market, read the [best AI interview software](/blog/best-ai-interview-software) guide. ## How we chose these tools We looked for what a programme lead needs: a first conversation for every founder, a follow-up, a link that runs from your form, your own questions, a result a partner reads in minutes, and a price that stays low in a rush. "Not stated" in the table means the vendor does not say. Prices are as the vendors display them. Selection facts come from Techstars and YC pages. Which path fits you: a programme with a yearly open call fits the [applications page](/use-cases/applications); a very large pool fits [volume hiring](/use-cases/volume-hiring); investor-style founder interviews fit [expert vetting](/use-cases/expert-vetting); and [people teams](/use-cases/people-teams) can reuse the same link for internal check-ins. Budget a cohort on [Tracy's pricing](/pricing). ## Common questions ### What is the best AI interview tool for an accelerator? Tracy AI Interviewer is the pick when every founder should get a first interview that asks follow-ups. Interviewer.AI fits admissions with a branded portal, Willo fits a single open call, and Spark Hire and VideoAsk fit recorded answers. ### Can it run from our application form and ask our own questions? Yes. Put one link in your form and every applicant takes the interview in their browser. Bring your own diligence questions, or describe what you want to learn and Tracy drafts them. ### How much does it cost for a cohort? Tracy costs $1 per credit, one credit per started 20 minutes, with the first 3 results free, no seats and no upfront cost. 100 founders at 20 minutes each is $100. Interviewer.AI is $199 or $399 a month, Willo is £49 per live role, and Spark Hire starts at $249 a month billed annually. ### How do accelerators select startups? Programmes publish their own steps. Techstars reviews every application and interviews for the customer and pain, speed and execution, market size and differentiation. YC runs 10-minute interviews with 2 to 3 partners who have read the application. --- Send us your application form and we build the first-interview plan in days, with one link in your form and a recording and report a partner can read. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=ai-interview-tools-for-accelerators&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=ai-interview-tools-for-accelerators&utm_content=try-free) ## Sources - [Tracy AI Interviewer, Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Applications](https://tracyinterview.ai/use-cases/applications), retrieved 9 Nov 2026. # Best AI Interview Tools for Expert Vetting (Nov 2026) > Expert vetting software compared: Tracy, Braintrust AIR, HackerRank, CodeSignal, Karat, Mercor and micro1 on depth, volume, stack, evidence and price. - URL: https://tracyinterview.ai/compare/ai-interview-tools-for-expert-vetting - Published: 2026-11-09 - Author: Tracy team > **In short** Tracy AI Interviewer is the pick when you already have the applicants and need to see which ones know the field. She interviews each one by video, asks follow-ups in the expert's own field, and leaves a recording, a timestamped transcript and integrity signals for a person to review. > > Every applicant can interview the day they apply, at any hour, with no scheduling. Pricing is public: $1 per credit, the first 3 results free, no seats. Tracy is not a marketplace and does not take your contracts. > > Braintrust AIR is the closest enterprise interview tool. HackerRank, CodeSignal and Karat fit engineering experts. Mercor and micro1 are the better pick when you want vetted people delivered instead of a tool. A lab project lands on a Monday, and the experts you vetted last quarter do not cover it. The new applicants all list the right degrees, and the few people who can tell which of them know the field are already booked. Expert vetting software interviews applicants before a person spends an hour on any of them. Seven tools are ranked below, each with a case where it is the better pick. Our [expert vetting page](/use-cases/expert-vetting) shows how Tracy runs it from a brief. ## What does expert vetting software have to do? The buyer is usually a head of expert supply, a head of operations or a founder. Five needs follow, and they are the ranking criteria. 1. Depth in the field: follow-ups in the expert's own field, not only a test. 2. Capacity when a project lands: the interview starts when the applicant clicks. 3. Proof that the expert who passed does the work: integrity signals shown to a human reviewer, as evidence and not proof, plus follow-ups that make a rehearsed answer hard to keep up. 4. A reason you can show the lab client: each approval links to the expert's own answers and the recording. 5. Whose people they are: a tool interviews your applicants, while a network supplies its own and may compete for your contracts. For question banks by field, see [AI trainer interview questions](/blog/ai-trainer-interview-questions). ## How do the seven tools rank, and why in this order? We ranked on how much of the vetting job, for your own applicants in any expert field, each tool covers, with a published price as the tie-breaker. Tracy is first because she covers depth, capacity, evidence and stack for any field at a public price. She is not first on every line: Braintrust AIR lists more ATS connectors and languages, HackerRank has a coding environment, and Karat puts a human engineer in the interview. Mercor and micro1 are last because they do not vet your own applicants. If you want vetted experts delivered, they move to the top. A tool vets people you already have; a network supplies people it has vetted, and its criteria decide who passes. See how [AI interviews work](/blog/how-do-ai-interviews-work). ## Which expert vetting software is the better pick, and when? Tracy interviews each applicant by two-way video. After every answer she asks about what was said, in the expert's own field, and waits for them to finish. You describe the project or bring your own questions, and with the CV check on she can read the CV first. A director at an accelerator residency programme said Tracy "sometimes brings up exactly those questions I might have forgotten to ask, which has been a pleasant surprise." The [product page](/product) shows the flow. A reviewer gets a recording, a timestamped transcript, a written assessment per question and integrity signals with their evidence: tab changes, gaze, face detection, delay before answers, fullscreen exits and external screens. The signals go to a human reviewer to interpret, are not proof of anything, and the team decides. We give no detection rate. Every expert's evidence sits in one place and one shape, which makes a large pool reviewable ([how reports read](/blog/ai-interview-reports)). One customer cut its average time to hire from 44 days to 7, and Tracy did the work of 53 full-time recruiters in one month. Tracy has completed 600K+ interviews for 300+ clients, with 99% interview completion. Pricing is $1 per credit, one credit per started 20 minutes, the first 3 results free, no seats and no upfront cost ([pricing](/pricing)). Choose another tool when you need a supply of experts, a proctored coding library, or native ATS connectors: Tracy does not supply experts, and her public Connect page describes no ATS write-back. ![Recreated Tracy signals panel: tab changes and delayed answers flagged with timestamps. A note says signals are reported, not scored.](/blog/images/ai-interview-tools-for-expert-vetting-1.png) *Two signals are flagged with the moments to check in the recording. Source: Tracy visual library, integrity-signals screen, fixture ai-interview-tools-for-expert-vetting-1. Illustration with sample data* Signals arrive with timestamps for a person to review, and none of them is a verdict. ### 2. Braintrust AIR: enterprise voice screening that syncs to your ATS Braintrust says AIR runs the screen, scores answers against a role-specific rubric and sends a ranked evidence pack to humans and the ATS. Follow-ups adapt to each answer, interviews run in 16+ languages, and it lists native Greenhouse, Lever, Workday, iCIMS and SmartRecruiters integrations. Pricing is volume-based with scoped pilots. Pick it for multilingual voice screening in an enterprise ATS team. ### 3. HackerRank: the engineering pick with a coding environment HackerRank's Chakra interviewer opens a live canvas with a real-world problem, asks follow-ups on the work, and flags unauthorized apps. Starter is $79 a month ($948 billed annually) with 60 candidate attempts a year, then $20 per extra attempt. Pick it for software engineers who need a coding environment ([AI technical interviews](/blog/ai-technical-interviews)); a physician or lawyer is outside what it describes. ### 4. CodeSignal: validated assessments plus a short AI interview CodeSignal starts at $79 a month billed annually, with $20 per extra credit. Its AI interviewer probes thin answers in about 15 minutes, and it lists a Suspicion Score and AI proctoring. Pick it for assessments validated by I/O psychologists alongside a short interview. ### 5. Karat: human engineers on demand Karat sells interviews run by human Interview Engineers, available same day and around the clock, with a report of a recommendation, a score and feedback per question. It shows no price. Pick it when you want a human engineer in the room and will pay a quoted price. ### 6. Mercor: vetted people delivered Mercor says it has 5M+ experts and sells human evaluation to frontier labs. It shows no buyer price and names no interview tool for your own applicants. Pick it when you want vetted experts supplied. ### 7. micro1: a data lab with its own AI recruiter micro1 describes expert human data and evaluations. Every expert applicant has an on-demand interview with its AI recruiter, Zara, against its certification criteria. The interview is not offered as software to buy. Pick it when you want micro1's expert supply. ## How do the seven tools compare on depth, volume, stack and evidence? | Tool | Depth | Volume | Stack | Evidence | Price | | --- | --- | --- | --- | --- | --- | | 1. Tracy AI Interviewer | Two-way video, follow-ups in the expert's field | Starts on click, any hour, no scheduling | API, MCP, Slack, WhatsApp, Zapier | Recording, transcript, assessment per question, integrity signals | $1 per credit, first 3 results free | | 2. Braintrust AIR | Adaptive follow-ups, voice | Interviews on the candidate's schedule | Greenhouse, Lever, Workday, iCIMS, SmartRecruiters; 16+ languages | Ranked evidence pack; session-integrity signals | Volume-based, no public price | | 3. HackerRank | Chakra follow-ups on work in a live canvas | 60 attempts a year on Starter | ATS and calendar on Pro; API and MCP on Enterprise | Score per competency with a reason; flags unauthorized apps | From $79 a month ($948 annual); $20 per extra attempt | | 4. CodeSignal | AI interviewer probes thin answers; about 15 minutes | AI phone screens without waiting for a live call | Gem, Ashby; Workday, iCIMS, Oracle on enterprise | Score report; Suspicion Score, AI proctoring | From $79 a month annual; $20 per extra credit | | 5. Karat | Human Interview Engineers with a rubric | On demand, same day, evenings and weekends | Not stated | Recommendation, score, written feedback per question | No price shown; request a demo | | 6. Mercor | AI assessments for expert applicants | 5M+ experts (vendor figure) | Not stated | Not stated | No public price | | 7. micro1 | Conversational interview with Zara; coding challenge for technical roles | On demand | Not stated | Result against its certification criteria | Not stated | ## What do the five needs leave out? Rival facts: vendors' own pages, read 9 Nov 2026. Tracy: approved product records. "Not verified" means we have not checked, not that it is absent. | Criterion | What we verified | Where a rival leads | | --- | --- | --- | | Implementation and service | Tracy Enterprise: a dedicated customer manager, a dedicated support channel, a setup customised to your workflow, a team walkthrough, customised in days. | Tracy for a tool set up in days; Karat for a human in the room. | | Multilingual | Braintrust AIR lists interviews in 16+ languages. | Braintrust AIR, on published languages. | | Security and compliance | Tracy is SOC 2 Type II attested, with the report on request. The rest: not verified here. | Tracy, on what we verified. | | ATS integration depth | Braintrust AIR lists native Greenhouse, Lever, Workday, iCIMS and SmartRecruiters connectors. CodeSignal lists Gem and Ashby. | Braintrust AIR. | | Support | Tracy Enterprise has a dedicated support channel and a dedicated customer manager. The rest: not verified. | Tracy, on what we verified. | ![Three steps: one place for every expert, one report per applicant, and volume that scales because interviews start on click.](/blog/images/ai-interview-tools-for-expert-vetting-2.png) *One platform to centralize and scale expert interviews with Tracy: the same follow-up interview for each expert, one report per applicant, and no scheduling. Source: Tracy Product and Value records.* Every expert applicant gets the same follow-up interview and one report, in one place, and volume scales because interviews start on click. ## How do you run a one-week bake-off? Rank tools on your own experts, because a demo shows the vendor's best case. In one week: 1. Pick one field where you can judge answers, and one owner who reads every result. 2. Invite ten people to each tool with the same brief: five experts you trust and five you have not met. 3. Decide the stop condition first: drop a tool that cannot separate the trusted five from the rest. 4. Score follow-up depth, whether a reviewer could decide from the recording in minutes, whether integrity evidence came with timestamps, hours from invite to result, and cost at your volume. On Tracy, 100 interviews of 20 minutes cost $100, with the first 3 results free (see [pricing](/pricing)), and you can invite and read results in ChatGPT or Claude through the [Connect page](/connect). ## How we chose these tools We looked for what a head of expert supply needs: depth in the field, capacity when a project lands, evidence a reviewer can check, a stack that fits and a price you can read. "Not stated" in the table means the vendor does not say. Prices are as the vendors display them, in different units. Which path fits you: [expert vetting](/use-cases/expert-vetting) for your own applicants in any field, [technical interviews](/use-cases/technical-interviews) for engineers, [volume hiring](/use-cases/volume-hiring) when thousands apply, [customer research](/use-cases/customer-research) for interview panels beyond hiring. For a wider list, see the [best AI interview software](/blog/best-ai-interview-software) guide. ## Common questions ### What is expert vetting software? It checks whether an applicant really knows a field before a person spends time on them, with an interview, an assessment or both. Tracy AI Interviewer interviews each applicant by video and leaves a recording, a transcript and signals for a reviewer. ### Can AI vet experts in niche fields? Tracy can. She asks follow-ups in the expert's own field, so you see how each one reasons. Other vendors here describe engineering or general enterprise roles, so test any tool on your own field first. ### How do we know the expert who passed is the one who does the work? Each Tracy interview captures tab changes, gaze, face detection, delay before answers, fullscreen exits and external screens, and the evidence goes to a human reviewer. A signal is not proof of anything, and your team decides. Follow-ups also make a rehearsed answer hard to keep up. We give no detection rate. ### Can you show a lab client why each expert was approved? Yes. Each Tracy report links to the expert's own answers, with the recording. ### How much does expert vetting software cost? Tracy costs $1 per credit, and a credit covers each started 20 minutes. The first 3 results are free, with no seats and no upfront cost. HackerRank Starter and CodeSignal start at $79 a month billed annually. Braintrust AIR, Karat, Mercor and micro1 show no public price. --- Send us one expert brief and we build the interview plan, customised in days, with a recording and report you can check. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=ai-interview-tools-for-expert-vetting&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=ai-interview-tools-for-expert-vetting&utm_content=try-free) ## Sources - [Tracy AI Interviewer, Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Expert vetting](https://tracyinterview.ai/use-cases/expert-vetting), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Connect](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. # Sapia.ai Pricing (Nov 2026) > Sapia AI pricing is not public: it bills per hire. See how to estimate it, what you get at that price, and Tracy's cost at 100 to 1,000 interviews. - URL: https://tracyinterview.ai/compare/sapia-pricing - Published: 2026-11-09 - Author: Tracy team > **In short** Sapia.ai does not disclose its price. Its pricing page says the answer depends on your hiring needs and offers a session with its team. It bills per hire. Tracy AI Interviewer is transparent: $1 per credit, one credit per started 20 minutes, first 3 results free, no seats, no upfront cost. > > At low hiring volume, a per-hire model with a minimum or a contract is likely to cost more per hire than pay-as-you-go. That is our reasoning, not a Sapia.ai fact. > > At the price, Tracy gives follow-ups in the applicant's own field (Depth), start on click at any hour (Volume), API, MCP and chat-tool connections (Stack), and a recording and timestamped transcript behind every result (Evidence). > > Tracy costs $100 for 100 interviews, $500 for 500 and $1,000 for 1,000, at 20 minutes or less each. Sapia AI pricing is not public. The pricing page lists no plan, rate or billing period; it says the answer depends on the hiring needs of your organization and offers a session to build a pricing guide. Tracy AI Interviewer puts the whole price on [one pricing page](/pricing): $1 per credit, one credit per started 20 minutes, first 3 results free, no seats, no upfront cost. Low, transparent pricing is part of what Tracy is. This page shows what Sapia.ai says about how it charges, how to estimate your cost, and what each tool gives you for the money. ## What does Sapia.ai say about its pricing? Sapia.ai's pricing page states: "The answer depends on the hiring needs and challenges of your organization." The rest describes the shape of the deal, not a figure. Source: Sapia.ai pricing page, Nov 2026. | Item | What the page says | | --- | --- | | Plans and list prices | None listed. No plan names and no figures. | | Billing unit | Per hire. You pay for the people you end up hiring, and you can interview as many people as you want. | | Who it is designed for | Companies that hire more than 500 people a year. | | Extra costs | Custom integrations may carry fixed costs, told in advance. | | How to get a price | A session with its team. No free trial or self-serve sign-up is mentioned. | ## How would per-hire pricing work out for a small team? Per-hire pricing moves the cost from the interview to the outcome: you interview as many applicants as you like and pay when someone is hired. For a company that hires hundreds a year, that shape is attractive, and Sapia.ai's page says its technology is designed for companies above 500 hires a year. Reasoning for a smaller team, not a Sapia.ai fact: Sapia.ai states no minimum, contract term or annual fee, but a vendor that tailors a rate to your yearly hiring and may add fixed integration costs has some fixed part to its deal. A fixed part is divided by your hires: small per hire across 600 hires, ten times larger across 60 specialist hires. Pay-as-you-go has no fixed part. So at low volume a per-hire model with a minimum or a contract is likely to cost more per hire. With no minimum the effect disappears, and at very high volume a low per-hire rate can beat any per-interview price. ## How can you estimate what Sapia.ai might cost you? You cannot price it from the outside, and we will not guess a figure. Build the model from your own numbers so the quote slots in: - Count the hires you expect in the next twelve months for the roles you would screen. - Ask for the per-hire rate at that count, and any fixed fees, minimum or annual term. - Divide the fixed fees by your hires and add the per-hire rate. For Tracy the sum needs no quote: interviews per hire times credits per interview times $1. At 20 interviews per hire of 20 minutes or less, that is $20 per hire, or $1,200 for 60 hires. A 34-minute interview uses two credits, $2. Our [AI interview cost guide](/blog/ai-interview-cost) covers the full cost of a first round. ## What does Tracy cost at 100, 500 and 1,000 interviews? One credit is $1 and covers each started 20 minutes of an interview that produced a result. Credits never expire, there are no seats and no upfront cost, a session with no answers is never charged, and your first 3 results are free with no card. See the [pricing page](/pricing). Tracy: tracyinterview.ai/pricing, $1 per credit, before the 3 free results. Sapia.ai: pricing page, Nov 2026. | Interviews run | Tracy, up to 20 minutes each | Tracy, 21 to 40 minutes each | Sapia.ai | | --- | --- | --- | --- | | 100 | $100 | $200 | Per-hire rate on request | | 500 | $500 | $1,000 | Per-hire rate on request | | 1,000 | $1,000 | $2,000 | Per-hire rate on request | ![Three figures: $100 for 100 interviews, $500 for 500 and $1K for 1,000, each up to 20 minutes.](/blog/images/sapia-pricing-2.png) *Tracy's cost at three volumes, for interviews of up to 20 minutes each. Sapia.ai bills per hire and does not publish a rate. Source: Tracy pricing page and Sapia.ai pricing page, Nov 2026.* Tracy's cost at 100, 500 and 1,000 interviews is a sum you can do yourself, and Sapia.ai's is a rate you must ask for. ## What do you get at each price: depth, volume, stack and evidence? The two tools buy different things. The table sets them side by side on four pillars. Tracy: product documentation and pricing page. Sapia.ai: its Chat Interview and pricing pages, Nov 2026. | Pillar | What you get with Tracy at $1 per credit | What Sapia.ai's pages describe | | --- | --- | --- | | Depth | Follow-ups in the applicant's own field, waiting for the end of the thought. Your own questions or ones drafted from the role. | Five structured questions in an untimed text chat, the same for every applicant. Follow-ups are not mentioned. | | Volume | Starts on click at any hour, ready in under 10 seconds, no account. Billed per interview, so cost scales with use. | Text on a phone, untimed, pause and return. Unlimited interviews, billed per hire; designed for companies hiring more than 500 a year. | | Stack | API and MCP, chat tools (ChatGPT, Claude, Slack, WhatsApp), Workable and Breezy HR beside the ATS connector. No seat fee. | Integrates with the ATS or HRIS you already have; custom integrations may carry fixed costs. | | Evidence | Recording and timestamped transcript behind every result, a written assessment per question, integrity signals with evidence. A person decides. | Every score traceable to the candidate's own words; ranked shortlist with the evidence beneath it; people decide. | ![Two panels on what each price buys across the four pillars: Sapia.ai as published, and Tracy at $1 per credit.](/blog/images/sapia-pricing-3.png) *What each price buys across the four pillars, as published. Sapia.ai's rows record what its pages do and do not say. Source: Tracy product documentation, Sapia.ai Chat Interview and pricing pages, Nov 2026.* Tracy's published price buys follow-ups, instant start, open connections and a recording behind every result. ## Depth: what happens when an answer is thin? Take a data engineer who says they restart a failed batch from the top. In Tracy, that becomes the next question: how does the retry avoid loading the same rows twice, and what if two loads run at once? Sapia.ai's published chat asks five structured questions, the same for every applicant, and its pages do not mention follow-ups. If depth in the applicant's own field is what you pay for, ask to see a live set-up. [How AI interviews work](/blog/how-do-ai-interviews-work) walks through a Tracy session. ## Volume: what does a large applicant pool cost in time and money? Tracy starts each interview on click, in a browser with no account, and you pay only for interviews that produced a result. [Hiring at volume](/use-cases/volume-hiring) shows how teams run it. Sapia.ai's text chat is untimed, works on a phone and has no interview cap, a good fit for a team above 500 hires a year whose applicants lack a quiet room or camera. Tracy's two-way video asks more of an applicant's setup, and that is a real limit. ## Stack: what does it cost to connect to your tools? Tracy connects through API and MCP, chat tools and beside the Workable and Breezy HR connectors, with no seat fee; the [Connect page](/connect) lists each one. Sapia.ai says it integrates with the ATS or HRIS you already have, and that custom integrations may carry fixed costs, so put that line in your budget. ## Evidence: can a reviewer check a result? Sapia.ai states that every score is traceable to the candidate's own words and that people decide. Tracy keeps a recording of both sides and a timestamped transcript behind every result, plus integrity signals (tab changes, abnormal eye tracking, face detection, delay before answers, leaving fullscreen) for a person to review. We give no detection rate. See [AI interview reports](/blog/ai-interview-reports). ![Recreated Tracy interviews table: fictional data engineer applicants, each with a status, the answers given and a View result link.](/blog/images/sapia-pricing-1.png) *Illustration, not a screenshot. A recreated Tracy interviews list for a data engineer role, with fictional applicants. Illustration with sample data* Every finished interview is one line with a result to open, so you can see what each credit bought. ## What should you ask Sapia.ai for a quote? Ask for each answer in writing: - the per-hire rate at your annual hires, and what counts as a hire; - any minimum, annual fee or contract term; - fixed costs for your ATS integration and custom work; - whether a pilot is possible, and its price. Our [HireVue pricing](/compare/hirevue-pricing) and [Paradox pricing](/compare/paradox-pricing) pages show how others charge. ## Common questions ### How much does Sapia.ai cost? Sapia.ai does not publish a price. Its pricing page describes a per-hire model with unlimited interviews and offers a tailored cost-per-hire rate after a session with its team. ### Does Sapia.ai have a free trial? The pricing page mentions a session with the team and does not mention a free trial or a free tier. Tracy gives your first 3 results free with no card. ### Is per-hire pricing cheaper than pay-as-you-go? It depends on volume and contract. At very high volume a low per-hire rate can win. At low volume a per-hire deal with a minimum or term is likely to cost more per hire, because the fixed part is spread over few hires. This is reasoning, not a Sapia.ai fact. ### What does Tracy cost? $1 per credit, with one credit for each started 20 minutes of an interview that produced a result. The first 3 results are free with no card. There are no seats and no upfront cost, and credits never expire. It is published on the pricing page. ### What does a Tracy interview cost if it runs long? Each started 20 minutes is one credit. A 15-minute interview is one credit, $1, and a 34-minute interview is two credits, $2. --- Send us your annual hires and applicant count, and we build the interview plan and work out your Tracy cost on the call. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=sapia-pricing&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=sapia-pricing&utm_content=try-free) ## Sources - [Sapia.ai: Pricing](https://sapia.ai/pricing/), retrieved 9 Nov 2026. - [Tracy AI Interviewer: Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer: Connect](https://tracyinterview.ai/connect), retrieved 5 Nov 2026. # Sapia.ai Reviews: What Buyers and Candidates Say (Nov 2026) > Sapia.ai reviews counted, not picked: 19 read, with quotes, dates and sources. What buyers praise, what they complain about, and who it suits. - URL: https://tracyinterview.ai/compare/sapia-reviews - Published: 2026-11-09 - Author: Tracy team > **In short** The complaint reviewers repeat most is scoring that misses the answer: 7 of 19 Sapia.ai reviews on Software Finder. Tracy AI Interviewer leaves a recording, a timestamped transcript and a written assessment per question behind every result, so a reviewer can check each one. > > Reviewers also praise Sapia.ai: 10 of 19 mention ease of use or setup, and 9 say it helps hiring decisions. > > All 7 scoring complaints come from people who run the tool, dated December 2025 or January 2026. > > Tracy's price is public: $1 per credit, the first 3 results free, no seats. The review that matters most to a buyer is the one from the recruiter who has to defend a score to a hiring manager. Of the 19 Sapia.ai reviews we read on Software Finder, seven describe scores that do not match what the candidate wrote, and all seven come from people who run the tool. Reviewers also praise it. If scores you cannot check are your worry, [Tracy AI Interviewer](/product) asks follow-ups in the applicant's own field and keeps a recording, a timestamped transcript and a written assessment per question behind every result. ## Where do the counts come from? Software Finder's Sapia.ai page showed 19 reviews, dated January 2024 to January 2026. We read all 19 and counted the themes ourselves. A review can carry more than one theme and 19 is a small sample, so read each count as a direction. Twelve were written by people who run Sapia.ai at an employer, all dated November 2025 to January 2026. Seven read as written by someone who took the chat, or either, and date from 2024. ## What do buyers praise about Sapia.ai? Ease comes first. Ten of the 19 reviews say Sapia.ai is easy to use, set up or start. One mid-market reviewer wrote, "Getting started was very easy too and the team was fantastic to work with throughout" (Software Finder, December 2025). Nine say it helps hiring decisions: an enterprise reviewer wrote, "Its talent insights help us make better hiring decisions and make it easier to choose the right candidates" (Software Finder, December 2025). Four describe a better candidate experience. An employer with 11 to 50 employees liked that "Candidates can complete this on their own schedule since there is no pressure to finish it all at once" (Software Finder, December 2025), which matches Sapia.ai's untimed text chat. ## What do buyers complain about? Counts are out of the 19 Software Finder reviews, with one quote each. Counts: our reading of the 19 Software Finder reviews. A review can raise more than one. Quotes are word for word. Too much writing (2) is left out of the table. | Complaint | Reviews (of 19) | One quote | Reviewer and date | | --- | --- | --- | --- | | Scores do not match the answer | 7 | "accuracy and consistency can be hit or miss at times and the scoring does not always line up with the final recommendation" | 500+ employer, Dec 2025 | | Genuine answers flagged as AI | 3 | "Candidates who recently finished higher education get flagged as AI-generated because the tool mistakes their academic writing style for AI use." | 500+ employer, Dec 2025 | | Spelling, grammar or dyslexia marked down | 3 | "I have seen candidates pass overall but still get marked as unsuitable because of small things like grammar mistakes or the use of slang." | Staffing employer, 500+, Dec 2025 | | Failed interviews, logins and slow fixes | 3 | "it would be much better if we could quickly resend a video or chat interview without extra hassle" | 500+ employer, Dec 2025 | | Reporting and ATS data | 2 | "Detailed reporting is not as fast or good enough as it should be." | 500+ employer, Dec 2025 | | Feels scripted or AI-run | 3 | "the program's communication with users looks scripted and tends to sound the same every time" | Reads as a candidate, Dec 2024 | ![Bar chart of complaints in 19 Sapia.ai reviews: scores off 7, false AI flag 3, spelling 3, ATS data 2, fixing faults 3.](/blog/images/sapia-reviews-2.png) *Reviews that raise each complaint, of 19 read. A review can raise more than one. Software Finder, 19 Sapia.ai reviews, January 2024 to January 2026.* Scoring is the issue reviewers raise most: 7 of 19, against 3 or fewer for every other complaint. ## Why do reviewers say the scores miss the answer? Seven reviews say the scores do not fit the answers. One says the tool "ends up marking candidates down because of it even when their answers are detailed and thoughtful" (Software Finder, January 2026). Another says it "penalizes people too harshly for spelling mistakes" (Software Finder, December 2025). A third describes a response that was "thorough and well-written" and still marked as failed (Software Finder, December 2025). Three of the seven also raise the AI flag. The January 2026 review on G2 says, "sapia will say an answer is ai when it is clearly not ai". Our guide to [AI interview cheating detection](/blog/ai-interview-cheating-detection) covers what to ask any vendor about flags. Seven of 19 is a small base. Even within those seven, one reviewer calls "The scoring system and the questions it asks are excellent" (Software Finder, December 2025). Treat seven as a reason to ask for a live score review in the demo. ## What does Sapia.ai say back? Sapia.ai replied to that January 2026 G2 review. It said the chat has no time limits, uses free-text responses and works with screen readers and speech-to-text, and that candidates with dyslexia rate their experience 8.8 out of 10. On AI use, it said its system may prompt during interviews if AI usage is detected, which it called guidance and not penalisation. That is the vendor's account. ## What do candidates say about the chat? Seven of the 19 reviews read as written by someone who took the chat, or by someone who could be either. All seven are positive overall. One says "Using it was really quick and easy", and another says the questionnaire "gave me the space to answer honestly". Two say the chat asks for a lot of writing, and one says the replies "sound the same every time". These reviews date from 2024, so they say less about the product today than the employer reviews do. ## Who is Sapia.ai right for, and who is it not? Sapia.ai suits an employer that screens a large volume, wants a structured text screen that candidates can finish on a phone, and has people to review scores and handle edge cases. Sapia.ai says its pricing and technology are designed for companies that hire more than 500 people a year. It suits you less if your reviewers must defend scores line by line, if many applicants find writing hard, if you need follow-up questions in the interview (its pages do not mention them), or if you need hiring data to flow back to your ATS. The table sets each complaint against what Tracy does, and what she does not fix. Tracy column: Tracy's published capabilities. Counts: 19 Software Finder reviews. | Complaint raised in reviews | What Tracy does | What Tracy does not fix | | --- | --- | --- | | Scores miss the answer (7) | A recording and a timestamped transcript behind every result, with a written assessment per question. | Tracy gives no score to dispute. It writes an assessment per question with the evidence, and a person decides. | | Genuine answers flagged as AI (3) | Integrity signals come with evidence for a person to review. Nothing is decided in the interview. | A signal can be raised on an honest applicant, and never proves cheating. | | Spelling, grammar or dyslexia marked down (3) | The interview is spoken, so there is no typed answer to mark for spelling. | Speaking is harder for some people. Sapia.ai's text format is more accessible here. | | Failed interviews and slow fixes (3) | A dropped applicant can rejoin within 10 minutes. A session with no answers is never charged. | The rejoin window is configurable, so confirm yours in the demo. | | Reporting and ATS data (2) | API and MCP connections, chat tools, and Workable and Breezy HR beside the ATS connector. | Tracy works next to your ATS. Ask how a result reaches your ATS record. | | Scripted or AI-run, too much writing (5) | Tracy asks follow-ups in the person's own field and waits for the end of the thought. Candidates rate Tracy 92 on NPS. | NPS and Sapia.ai's 9/10 are different measures, so we do not compare them. | ![Recreated Tracy result report: recording, transcript with a follow-up question, each question with what was said and quoted, and fields marked Stated.](/blog/images/sapia-reviews-1.png) *Illustration, not a screenshot. A recreated Tracy result report shows the recording, a timestamped transcript with Tracy's follow-up, and stated fields with the quoted words. Recreated Tracy screen, fictional applicant.* Each complaint from the reviews has an answer you can see in a Tracy result: the person's words, the time, the follow-up and the recording. ## What does Sapia.ai cost, in brief? Sapia.ai publishes no price. Its pricing page describes a per-hire model with no cap on interviews and says the technology is designed for companies hiring more than 500 people a year. Tracy's price is public: $1 per credit, one credit for each started 20 minutes, the first 3 results free, no seats and no upfront cost. At 20 minutes each, 100 interviews cost $100. See the [pricing page](/pricing) for worked examples, and [what an AI interview costs](/blog/ai-interview-cost) for the wider picture. ## What are the alternatives to Sapia.ai? Stay with Sapia.ai if a structured text screen at 500 or more hires a year is what you want and your reviewers can live with the scoring issues above. Look at Tracy if you want a two-way interview with follow-ups and a result a reviewer can check against a recording. Look at an interviewer built into your ATS if one system matters most, and read [ATS interviewers vs a dedicated one](/blog/ats-ai-interviewer-vs-dedicated) first. For a wider shortlist, see [the best AI interview software](/blog/best-ai-interview-software). For another vendor's reviews counted the same way, read [HireVue reviews](/compare/hirevue-reviews). For what a Tracy report contains, see [AI interview reports](/blog/ai-interview-reports). Tracy's volume offer is on the [hiring at volume page](/use-cases/volume-hiring). ## What should you take from the reviews? A director at an accelerator residency programme says of Tracy: "Tracy sometimes brings up exactly those questions I might have forgotten to ask, which has been a pleasant surprise." Hold any vendor to one test: when a score is questioned, can a reviewer see the words behind it, and can your team change the outcome? Tracy's price is as open as her results: $1 per credit, the first 3 results free, no seats, published on the [pricing page](/pricing). You can [try Tracy free](/connect) on one of your own roles. ## Common questions ### Is Sapia.ai any good, according to reviews? Reviewers praise ease of use and setup (10 of the 19 we read) and say it helps hiring decisions (9 of 19). The complaint they repeat most is scoring that misses the answer (7 of 19). ### What are the main complaints about Sapia.ai? Scores that do not match the answer (7 of 19), genuine answers flagged as AI-written (3), spelling, grammar or dyslexia marked down (3), failed interviews and slow fixes (3), reporting and ATS data (2), and too much writing (2). ### Does Sapia.ai detect AI-written answers? In a reply on G2, Sapia.ai said its system may prompt during interviews if AI usage is detected, as guidance. Three reviewers say genuine answers were flagged. ### How much does Sapia.ai cost? Sapia.ai publishes no price. It bills per hire with no cap on interviews and gives a tailored rate after a session with its team. Tracy costs $1 per credit, with the first 3 results free and no seats. ### How does Tracy handle an assessment a reviewer doubts? A recording and a timestamped transcript sit behind every Tracy result, with a written assessment per question. The reviewer checks the assessment against the words, and a person decides. --- Send us one role. We build the interview plan in days, and you check an assessment against the recording yourself. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=sapia-reviews&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=sapia-reviews&utm_content=try-free) ## Sources - [Sapia.ai: Pricing](https://sapia.ai/pricing/). - [Tracy AI Interviewer: Pricing](https://tracyinterview.ai/pricing). # Tracy vs Sapia.ai: Chat Answers or a Two-Way Interview? (Nov 2026) > Tracy vs Sapia: a five-question text chat against a two-way interview with follow-ups. Where each one wins and what to test before you pick. - URL: https://tracyinterview.ai/compare/tracy-vs-sapia - Published: 2026-11-09 - Author: Tracy team > **In short** Tracy AI Interviewer asks follow-ups in the applicant's own field over two-way video, and keeps a recording and timestamped transcript behind every result. Sapia.ai runs a fixed text chat of five structured questions, with no video or voice. > > Pick Sapia.ai if you hire more than 500 people a year, many applicants have no camera or quiet room, and you want a published independent bias audit. > > Pick Tracy if the first round has to show how someone reasons in their field: each applicant starts on click in a browser with no account, and a person reviews the evidence. > > Tracy costs $1 per credit, published on her pricing page, with the first 3 results free, no seats and no upfront cost. Sapia.ai publishes no price. Candidates give Tracy a 92 NPS. A shortlist built from five chat answers looks tidy. The trouble shows up a week later, when a hiring manager asks how that applicant would handle a failed data load and the screen holds nothing on it, because nobody asked. That gap is the main difference between Sapia AI and Tracy AI Interviewer. Tracy asks follow-ups in the applicant's own field and keeps a recording and timestamped transcript behind every result. Sapia.ai sends every applicant the same five structured questions as text. Tracy also publishes her price: $1 per credit, no seats, no upfront cost, on the [pricing page](/pricing). ## Where is Sapia.ai the better pick? Sapia.ai's Chat Interview is text only: no camera, no bandwidth demand, no time limit, so an applicant can answer on a phone at 11 pm and come back later. Its pages list multiple languages, an independent bias audit it says is published openly, and ISO/IEC 42001 and SOC 2 Type 2. It says it has assessed more than 10 million candidates. It charges per hire with no limit on interviews and says its technology is designed for companies that hire more than 500 people a year. Pick Sapia.ai if you hire more than 500 people a year, many of your applicants have no quiet room or good camera, and your legal team wants a published bias audit before it signs. For frontline and graduate hiring at that scale, a text chat is a reasonable first round, and Tracy's video format asks more of an applicant's setup. We make no bias-audit claim for Tracy. What we can show is a SOC 2 Type II attestation, with the report on request at hello@tracyinterview.ai. ## How do Sapia.ai and Tracy compare on depth, volume, stack and evidence? One pillar per row. The last column is a check you can run yourself. Sapia.ai column: its own pages, Nov 2026. Tracy column: tracyinterview.ai pricing and Connect pages and product documentation. | Pillar | Tracy AI Interviewer | Sapia.ai (its own pages) | What to test | | --- | --- | --- | --- | | Depth | Follow-ups in the person's own field; waits for the end of the thought. Your own questions or ones drafted from the role; a coding task can sit inside. | Five structured questions in an untimed text chat, the same for every applicant. Follow-ups are not mentioned. | Give both tools one thin answer and see what each does next. | | Volume | Starts on click at any hour, ready in under 10 seconds, no account. A dropped applicant can rejoin within 10 minutes. | Text on a phone, untimed, pause and return. Built for companies hiring more than 500 people a year; interviews are not capped. | Measure how many of your own applicants finish, on their own devices. | | Stack | API and MCP, chat tools (ChatGPT, Claude, Slack, WhatsApp), Workable and Breezy HR beside the ATS connector. | Integrates with the ATS or HRIS you already have; a few days to go live; custom integrations may carry fixed costs. | Ask each to show a finished result arriving in your ATS record. | | Evidence | Recording and timestamped transcript behind every result, a written assessment per question, integrity signals with evidence. A person decides. SOC 2 Type II attested. | Every score traceable to the candidate's words; ranked shortlist with evidence; people decide; bias audit published; SOC 2 Type 2 and ISO/IEC 42001. | Pick a score or an assessment you disagree with and ask each to show the answer behind it. | | Price | $1 per credit, one credit per started 20 minutes, published on tracyinterview.ai/pricing. First 3 results free. No seats, no upfront cost. | Per hire, no rate published; a tailored rate comes from a session with its team. | Ask Sapia.ai for its per-hire rate and any fixed fees in writing. | ## What do the four pillars leave out? Rival columns: vendors' own pages, read 9 Nov 2026. Tracy column: approved Tracy product records. "Not stated" and "Not verified" mean the page does not say or we have not checked, not that it is absent. | Criterion | Tracy AI Interviewer | Sapia.ai | Who has the edge on what we verified | | --- | --- | --- | --- | | Implementation and service | Enterprise includes a dedicated customer manager, a setup customised to your workflow and a walkthrough for your team. Tracy is customised in days per customer need. | Says its integrations team can get you live in a few days. | Level: Sapia.ai goes live in a few days, Tracy is customised in days. | | Accessibility | Not stated. | No time limits, free text, screen reader and speech-to-text compatible (vendor reply to a review). | Sapia.ai, on what is published. | | Multilingual | A spoken-language check scores fluency, coherence, grammar, pronunciation and vocabulary from 0 to 100 beside the transcript. Tracy publishes no language count. | Multiple languages in one comparable shortlist; no count published. | Sapia.ai on published languages; Tracy adds a spoken-language check. | | Security and compliance | SOC 2 Type II attested, with the report on request at hello@tracyinterview.ai. | Lists ISO/IEC 42001, 27001, GDPR and SOC 2 Type 2. | Sapia.ai lists more standards; Tracy is SOC 2 Type II attested. | | Validation, bias controls and oversight | A person reviews the evidence report and decides, and the full history stays on the record. Tracy publishes no bias audit. | Says an independent bias audit is published openly; people decide. | Sapia.ai, with a published bias audit. Tracy publishes none. | | ATS integration depth | Workable and Breezy HR chat connectors beside your ATS, API, MCP and Zapier. Candidates stay in the ATS; no write-back is stated. | Integrates with the ATS or HRIS you already have; custom integrations may carry fixed costs. | Sapia.ai claims wider ATS coverage; Tracy works beside your ATS through chat tools, API and MCP. | | Reporting and audit trail | Recording of both sides, timestamped transcript and an assessment per question with reasoning; a record shareable by link or PDF. | Ranked shortlist with a fit score, summary and evidence; a Talent Intelligence Agent to explore responses. | Different formats: Tracy keeps the recording and a written assessment per question; Sapia.ai keeps the written answers. | | Contract structure | Pay as you go, no seats, no upfront cost. Enterprise is priced to how many interviews you run. | Per hire, tailored rate; no trial stated. | Tracy, if you want to start without a session. | | Support | Enterprise has a dedicated support channel and a dedicated customer manager. | Not verified. | Tracy, on what is published. | ## How do the published workflows differ on a thin answer? A fixed question set and a follow-up part ways on a thin answer. Take a data engineer who says they restart a failed batch from the top. In Tracy, that sentence becomes the next question: how does the retry avoid loading the same rows twice, and what happens if two loads run at once? Each answer sits in the transcript with a timestamp. A director at an accelerator residency programme describes the effect: "Tracy sometimes brings up exactly those questions I might have forgotten to ask." ![Two panels. Sapia.ai as published: five fixed questions, no follow-ups, text only. Tracy: follow-ups in their field, two-way video.](/blog/images/tracy-vs-sapia-2.png) *What each format does, as published. Sapia.ai's pages do not mention follow-up questions. Source: Sapia.ai Chat Interview page and Tracy product documentation, Nov 2026.* Sapia.ai's published chat is a fixed set of five questions, while Tracy builds her next question from the answer just given. ![Recreated Tracy interview screen: the applicant's answer beside Tracy's follow-up about loading the same rows twice.](/blog/images/tracy-vs-sapia-1.png) *Illustration, not a screenshot. A recreated Tracy interview screen with a fictional applicant: the follow-up quotes what the applicant just said. Illustration with sample data* Tracy's next question quotes the applicant's own words back to them. On Sapia.ai's published description, that third question has no place to come from: five structured questions, the same for every applicant, and no mention of follow-ups. Its home page describes a Talent Intelligence Agent that lets the recruiter explore every score and response, which is a recruiter questioning a transcript, not the applicant being asked more. If Sapia.ai asks follow-ups in some set-ups, the demo will show it. For the case for a two-way format, read [one-way vs two-way video interviews](/blog/one-way-vs-two-way-video-interviews); for the format itself, [how AI interviews work](/blog/how-do-ai-interviews-work). ## Can a reviewer check a score against what the person said? Both tools say yes, in their own formats. Sapia.ai states that every score is traceable to the candidate's own words and that people decide. In a text chat the written answer is itself the record. Tracy keeps a recording of both sides and a timestamped transcript behind every result, with a written assessment for each question. She also captures integrity signals for a person to review: tab changes, abnormal eye tracking, live face detection, delay before answers, leaving fullscreen and external screens. Nothing is decided in the interview, and we give no detection rate. Sapia.ai's product pages list no such signals; a vendor reply on G2 says its system may prompt if AI use is detected. See [AI interview reports](/blog/ai-interview-reports) and [AI interview cheating detection](/blog/ai-interview-cheating-detection). ## What does it cost to run 100, 500 or 1,000 interviews? Tracy's price is public, and low, transparent pricing is part of what she offers: $1 per credit, one credit per started 20 minutes of an interview that produced a result. At 20 minutes each, 100 interviews cost $100, 500 cost $500 and 1,000 cost $1,000, before the 3 free results. No seats, no upfront cost, and a session with no answers is never charged. Sapia.ai does not disclose its rate: its pricing page describes a per-hire model with unlimited interviews and a tailored rate from a session with its team. Because it bills hires and Tracy bills interviews, compare per hire: at 25 interviews per hire, Tracy's fees are $25 per hire. Ask Sapia.ai for its per-hire rate and any fixed fees in writing. See the [pricing page](/pricing) and [what an AI interview costs](/blog/ai-interview-cost). ## Which one fits the tools you already run? Sapia.ai says it integrates with the ATS or HRIS you already have and can go live in a few days. Tracy works next to your ATS through API and MCP, chat tools, and beside the Workable and Breezy HR connectors, with Zapier for the rest. The [Connect page](/connect) lists each one. In the demo, check how a finished result reaches your ATS record, for either tool. [ATS interviewers vs a dedicated one](/blog/ats-ai-interviewer-vs-dedicated) covers the bundled-tool question. ## How do you test both on the same 20 applicants? A demo shows the best case, so test on your own applicants. Take one role family and 20 real applicants whose outcome you already know, run them through both tools, and decide a stop condition first, for example fewer than 15 of 20 finishing. Measure how many finish on their own devices, reviewer minutes per result, how often the reviewer's pick matches the later human interview, and cost per hire. Tracy's first 3 results are free, so you can start with three of the twenty. More on the field in [AI interview tools for high-volume hiring](/blog/ai-interview-tools-for-high-volume-hiring) and our [volume hiring page](/use-cases/volume-hiring). ## Where does Tracy stop? Tracy runs a two-way video interview, so an applicant needs a browser and a setup that allows video; a text chat asks less. Tracy does not decide: a person reads the report. We make no bias-audit or language-count claim, so if those matter most, weigh Sapia.ai's published audit and languages heavily. ## Common questions ### Does Sapia.ai ask follow-up questions? Its pages describe five structured questions, the same for every applicant, and do not mention follow-ups. Ask to see a live set-up in the demo. ### How much does Sapia.ai cost? It does not disclose a rate. It prices per hire, with no cap on interviews, and gives a tailored rate after a session with its team. ### Does Tracy work with my ATS? Tracy works next to it, through API and MCP, chat tools and beside the Workable and Breezy HR connectors. --- Send us one role and 20 applicants. We build the interview plan in days, and you read the Tracy reports next to whatever you run today. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=tracy-vs-sapia&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=tracy-vs-sapia&utm_content=try-free) ## Sources - [Sapia.ai: Pricing](https://sapia.ai/pricing/), retrieved 9 Nov 2026. - [Tracy AI Interviewer: Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer: Connect](https://tracyinterview.ai/connect), retrieved 5 Nov 2026. # VidCruiter vs Tracy AI Interviewer > VidCruiter vs Tracy for technical hiring: where each one wins on live panels, follow-ups, integrations, security, support and price. - URL: https://tracyinterview.ai/compare/tracy-vs-vidcruiter - Published: 2026-11-09 - Author: Tracy team > **In short** Verdict: VidCruiter is the stronger full interview suite; Tracy is the stronger first round when a technical role needs follow-up questions and a report a reviewer can check. > > Choose VidCruiter when you need live panels, scheduling, native Workday or SAP SuccessFactors connectors, identical questions for every candidate, or a vendor-led rollout with support. > > Choose Tracy when the first round has to show who is deep and who is only polished, every applicant should be interviewed the day they apply, and you want a published price with the first 3 results free and a setup customised to your workflow in days. > > Run both on one role for two weeks when the role is technical and you expect at least 20 applicants. This VidCruiter vs Tracy comparison is for technical teams that hire at volume. A reviewer opens the fortieth recorded answer to the same question and still cannot tell whether the candidate runs this system or has only read about it. Tracy AI Interviewer runs the interview itself and asks the next question. VidCruiter is the stronger suite for live panels, scheduling and HR system connectors, and we say so below. ## What does VidCruiter do well? VidCruiter calls itself a platform for "enterprise HR, government, and healthcare teams requiring compliance and structured rubrics." It runs pre-recorded interviews, where candidates record answers to structured questions on their own time, and live interviews with a human interviewer, one on one, in a panel or in timed multiple mini interview stations. Its AI is assistive. VidCruiter says its AI tools are "assistive, not autonomous" and lists notes, summaries, rubric scoring, fraud flags and automated interview guides. AI screening writes a rationale for each score, and a hiring manager can override any result. Its home page navigation lists an AI Interviewer as coming soon. Its structured format has the interviewer ask "each candidate the same questions in the same order." If answers must compare cleanly across many hiring managers, that is a sound position. ## VidCruiter vs Tracy: buyer criteria side by side | Criterion | Tracy AI Interviewer | VidCruiter | Who leads | | --- | --- | --- | --- | | Interview depth | Two-way video interview with follow-ups in the person's own field. Questions come from your brief or your own list. | Pre-recorded or live with a human. The structured format asks the same questions in the same order. Its AI is described as assistive. | Tracy for adaptive follow-ups. VidCruiter for live panels. | | Evidence per result | Recording of both sides, timestamped transcript, written assessment per question. Integrity signals are shown to a reviewer. | Timestamped transcript, optional AI screening with a rationale, override by a reviewer. | Different. Open one result in each and go from the assessment to the sentence behind it. | | Price and first-year cost | $1 per credit, one credit per started 20 minutes. First 3 results free, no seats, no upfront cost. | No public price; a request form. It says annual terms come without multi-year lock-in. | Tracy for a published price. Ask VidCruiter for an all-in first-year number. | | ATS and stack | API and MCP connections, Slack, WhatsApp and Zapier, and use beside the Workable and Breezy HR connectors. | 75+ integrations, including Workday, SAP SuccessFactors, Oracle, ADP, UKG and Dayforce, plus an API. | VidCruiter on native HR connectors. | | Implementation and support | Enterprise includes a dedicated customer manager, a dedicated support channel, a setup customised to your workflow and a team walkthrough. Customisation takes days. | Says implementation is included with a dedicated specialist and a dedicated customer success manager. Live support answers within 1 minute, around the clock. | Level on dedicated service. VidCruiter on 24/7 support. | | Security and compliance | SOC 2 Type II attested. The report is available on request. | States SOC 2 Type II, GDPR and PIPEDA, ISO, and a choice of where data is hosted. | VidCruiter lists more standards. | | Accessibility and languages | A spoken-language check scores fluency, coherence, grammar, pronunciation and vocabulary beside the transcript. Accessibility accommodations: Not stated. | Says its platform meets WCAG 2.2 AA. Notes and transcripts in English and French; support in 8+ languages. | VidCruiter on accessibility. Tracy when spoken language is part of the role. | | Validation, bias controls and oversight | A person decides: they can move a candidate forward or invite them again. The full history stays on the record. Tracy publishes no bias audit. | Database of validated questions, rubrics built to I/O psychology standards, a rationale for each AI score, hiring manager override. | VidCruiter on published validation. Level on human oversight. | | Reporting and audit trail | A report per interview with an assessment per question and the recording behind it. The package ships by link or PDF, and the full history stays on the record. | Dashboards for time-to-hire and interviewer activity, DEI reporting, and audit trails for every interview, evaluation and system action. | VidCruiter on dashboards and audit reporting. Tracy on shareable per-interview reports. | ## How the published workflows differ for the same technical role Take one invented role, a backend engineer on a payments team. The figure illustrates how a Tracy interview would run. The candidate is fictional and nothing here is a recording of a real interview. ![Recreated Tracy screen: a follow-up asks where the event id is stored and what happens if two copies arrive at once.](/blog/images/tracy-vs-vidcruiter-1.png) *Illustration with sample data* At 07:48 Tracy asks about the detail the candidate just skipped, and the answer is recorded with a timestamp. Here is a chain Tracy could run on that role, with an invented candidate. > Tracy: Your service retries a failed webhook. How do you keep the retry from creating a second order? > > Candidate: I store the event id and skip any id I have already handled. > > Tracy: Where do you store that id, and what happens if two copies arrive at the same moment? > > Candidate: In the same transaction as the order, with a unique constraint, so the second insert fails. > > Tracy: What does the sender get back when that insert fails? The first answer would earn the same note from someone who has built this and someone who has read about it. The second shows a failure path. A preset list reaches it only if someone wrote it in for every candidate. See [one-way vs two-way video interviews](/blog/one-way-vs-two-way-video-interviews). ![Two columns comparing a preset question list with Tracy's follow-ups and evidence.](/blog/images/tracy-vs-vidcruiter-2.png) *What a follow-up adds to a preset question list. A generic preset list on the left, Tracy on the right. Source: Tracy product description.* A follow-up shows depth that a preset list can show only if someone wrote the second question in advance. ## Do follow-ups show more than a fixed question list? VidCruiter's structured format exists to make answers comparable, and it does that. Tracy keeps the same starting questions for every applicant to a role, then adds what a fixed list leaves out: a follow-up in the person's own field, asked after she waits for the end of the thought. A coding task inside the interview gets follow-ups on approach and trade-offs, as [AI technical interviews](/blog/ai-technical-interviews) explains. A director at an accelerator residency programme says: "Tracy sometimes brings up exactly those questions I might have forgotten to ask, which has been a pleasant surprise." For the same question against another enterprise suite, see [Tracy vs HireVue](/compare/tracy-vs-hirevue). ## Can a reviewer check a VidCruiter or a Tracy result? VidCruiter's AI screening writes "a written rationale for every score, citing specific things the candidate said," and a person can override it. Tracy's assessment covers each question and the follow-up answers behind it, with the recording and a timestamped transcript to open. A person decides in both. For integrity, VidCruiter's fraud page lists controls that lock the test down, such as locking screens and browsers. Tracy works inside the conversation. It captures tab changes, gaze, live face detection, delays before answers, fullscreen exits and external screens, and shows the evidence to a person. These are signals for a human reviewer, not proof of cheating, and nothing is decided in the interview. See [AI interview reports](/blog/ai-interview-reports) and [AI interview cheating detection](/blog/ai-interview-cheating-detection). ## Does Tracy hold up at volume? Tracy has completed 600K+ interviews for 300+ clients, with 99% interview completion. One customer cut its average time to hire from 44 days to 7, and Tracy did the work of 53 full-time recruiters in one month. An account executive at a US IT staffing firm says of candidates: "By the time they finish with this platform, they're even laughing and joking." See [hiring at volume](/use-cases/volume-hiring). ## What do VidCruiter and Tracy cost? VidCruiter does not publish a price. Its pricing page is a form covering nine modules, and it says pricing is "custom pricing based on the modules you select and your organization's hiring volume." Tracy's price is published. A credit costs $1 and covers each started 20 minutes of an interview that produced a result, so a 15-minute interview costs $1 and a 34-minute one $2. The first 3 results are free with no card, there are no seats, and a session with no answers is never charged. 100 interviews of up to 20 minutes cost $100. Enterprise is priced to your volume and adds a dedicated customer manager, a dedicated support channel and a setup customised to your workflow in days. See [Tracy's pricing](/pricing) and [AI interview cost](/blog/ai-interview-cost). ## How do you test both before you commit? Run one technical role with at least 20 expected applicants for two weeks. For the mechanics first, read [how AI interviews work](/blog/how-do-ai-interviews-work). 1. In VidCruiter's demo, get the quote in writing before the pilot starts. 2. In Tracy, describe the role and edit the drafted questions, or send us the role and we build the plan. The first 3 results are free. 3. Have one reviewer read the reports from your next 20 applicants. Count the follow-ups that landed, the results verified from the recording in under a minute, and the cost per interview. 4. Stop if the reviewer cannot name a candidate the reports separated that your usual screen would have merged. ## Which path fits you? - Technical hiring: [technical interviews](/use-cases/technical-interviews) covers role-specific interviews with follow-ups. - High-volume hiring: [volume hiring](/use-cases/volume-hiring) covers pilots for large applicant pools. - Expert vetting: [expert vetting](/use-cases/expert-vetting) covers interviews for specialists and contributors. - Enterprise and integrations: [Connect](/connect) lists the integrations, and [pricing](/pricing) explains Enterprise pricing to your volume. ## Common questions ### Is VidCruiter a one-way video interview tool? It offers both. Pre-recorded interviews are one-way, and live interviews are two-way with a human interviewer. Tracy is a two-way interview run by an AI interviewer. ### Does VidCruiter have an AI interviewer? VidCruiter says its AI tools are assistive, not autonomous, and its home page navigation lists an AI Interviewer as coming soon. ### How much does VidCruiter cost? VidCruiter publishes no price. You request a quote, based on modules and hiring volume. Tracy's price is on its pricing page. ### Is VidCruiter better than Tracy? It depends on the job. VidCruiter is the better pick for live panels, scheduling, native HR connectors and 24/7 support. Tracy is the better pick when the first round must show how technical candidates reason. ### Will candidates drop out of a Tracy interview? Candidates rate Tracy 92 on NPS. They join from a browser link at any hour, with no account, and are told on the join screen that the conversation is recorded and transcribed. --- Send us one role and we will build the interview plan, with a recording and report you can check. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=tracy-vs-vidcruiter&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=tracy-vs-vidcruiter&utm_content=try-free) ## Sources - [VidCruiter, Pricing](https://vidcruiter.com/pricing/), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Connect](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. # VidCruiter Alternatives (Nov 2026) > VidCruiter alternatives for technical hiring: Tracy, HireVue, Spark Hire, Hireflix, Willo and Sapia.ai on format, evidence, stack and price, 9 Nov 2026. - URL: https://tracyinterview.ai/compare/vidcruiter-alternatives - Published: 2026-11-09 - Author: Tracy team > **In short** Tracy AI Interviewer is the pick when VidCruiter's preset questions are not deep enough: a two-way interview with follow-ups in the person's own field, a recording and timestamped transcript behind every result, and a published price of $1 per credit with the first 3 results free. > > HireVue is the closest enterprise suite and publishes no price. Spark Hire (from $249 a month billed annually) and Hireflix ($75 or $150 a month billed yearly) are the low flat-price recorded options. Willo charges £49 per live role, and Sapia.ai runs a text-only interview priced per hire. > > Stay with VidCruiter if you need live panel interviews, scheduling and native Workday or SAP SuccessFactors connectors. Teams look for VidCruiter alternatives when the quote, the scope or the question list stops fitting. When the first round has to show how an engineer reasons, Tracy AI Interviewer is the first tool to try. Below: Tracy first, then five tools for other needs, each with who it suits and where it falls short. ## Why do teams look for VidCruiter alternatives? Three reasons come from VidCruiter's own pages. The price is a form: "Fill out the form to get pricing information on our solutions." The scope is enterprise: it says it is "designed for enterprise HR, government, and healthcare teams requiring compliance and structured rubrics." The questions are preset: the interviewer "asks each candidate the same questions in the same order," with no follow-up on what the candidate just said. Its navigation also labels the AI Interviewer "COMING SOON". Some teams should stay. If you run live panel or multiple mini interview formats, need self-serve scheduling, or need native connectors to Workday, SAP SuccessFactors, Oracle, ADP or UKG, VidCruiter lists all of it and the tools below mostly do not. ## Which VidCruiter alternative suits which team? - Tracy AI Interviewer: technical or expert roles where the first round must show depth, at a published price per result. - HireVue: large enterprise teams that want governance documents, audits and one suite, and accept a quote. - Spark Hire: small and mid-size teams that want recorded and live video at a published monthly price, with humans deciding. - Hireflix: teams under 250 people that want the lowest flat price for one-way video. - Willo: teams that want to pay per live role, or an assessment library with ISO 27001. - Sapia.ai: companies hiring more than 500 people a year that are happy with a text-only interview. The sections below give the detail and the limits of each. ## Tracy AI Interviewer: follow-ups in the person's field Tracy interviews every applicant by two-way video. After each answer she asks about what was said, in the person's field, and waits for them to finish. A director at an accelerator residency programme said Tracy "sometimes brings up exactly those questions I might have forgotten to ask, which has been a pleasant surprise." Behind every result sit a recording, a timestamped transcript, a written assessment per question and integrity signals shown to a human reviewer. Bring your own questions or let Tracy draft them, and add a coding task for technical roles (see [technical interviews](/use-cases/technical-interviews)). She starts on click at any hour, connects through API and MCP, Slack, WhatsApp and Zapier, and is SOC 2 Type II attested. She has completed 600K+ interviews for 300+ clients, with 99% interview completion, and one customer cut its average time to hire from 44 days to 7. Where others are better: a flat recorded plan costs less at high volume (1,000 Tracy interviews cost $1,000; Hireflix's small plan is $75 a month), HireVue publishes more governance material, and VidCruiter lists native HR-system connectors and live panel scheduling. See [Tracy vs HireVue](/compare/tracy-vs-hirevue) for the enterprise-suite view. ![Recreated Tracy result report with a recording, a timestamped transcript and three questions, each with what the candidate said and quoted evidence.](/blog/images/vidcruiter-alternatives-1.png) *The second question follows from the first answer. Illustration with sample data* The report shows what the candidate said to each question, with the quote and the transcript behind it. ## HireVue: the enterprise suite with governance documents HireVue is the closest like-for-like swap for an enterprise buyer. Its AI interviewer has agents that "ask and record candidate responses, including probing follow-up questions, in a natural two-way conversation," it lists more than 200 on-demand coding skills assessments, and it says "every score ties back to evidence from the interview and your competency rubric." It publishes adverse impact testing and an annual SOC 2 Type 2 audit, says it is FedRAMP certified, and lists no plan or price. Pick it when governance documents are a procurement requirement. See [HireVue pricing](/compare/hirevue-pricing). ## Spark Hire: recorded and live video at a published price Spark Hire starts from $249 a month billed annually, with unlimited live and one-way video interviews, jobs and users, 40+ ATS integrations, transcription in 50+ languages and an AI summary of each response. Its stance is human judgment: "Faster Candidate Screening. No Automated Decisions." Every candidate answers the same questions. Pick it for a published price and a lean team that reviews recordings. ## Hireflix: the lowest flat price for one-way video Hireflix describes "the most affordable unlimited one-way video interviewing plans": $75 a month billed yearly under 50 employees, $150 for 50 to 250, custom above. Every plan has unlimited seats, positions and responses, automatic transcriptions and a first month free, and it says it is SOC 2 compliant. Pick it to screen volumes of recorded answers on a small budget. ## Willo: pay per live role Willo's Lite plan is £49 per live role with no contract, with 150 assessed candidates per live role; Enterprise is from £2,999 a year. Both list custom scorecards, anti-cheat, 18+ languages and ISO 27001, with 50+ native integrations and 1,500+ assessments. Its page lists "follow-up questions" without saying how they work. Pick it if you hire for a few roles at a time. ## Sapia.ai: a text-only interview for large volumes Sapia.ai interviews by chat: "No video, no voice, no face. Text only, by design." It says every score is "traceable to a candidate's own words," prices per hire with no public rate, and is designed for companies that hire more than 500 people a year. Pick it for frontline or graduate volumes where a written answer is enough. ## VidCruiter and its alternatives compared on four pillars and price | Tool | Depth | Volume | Stack | Evidence | Price | | --- | --- | --- | --- | --- | --- | | Tracy AI Interviewer | Two-way video, follow-ups in the person's field | Starts on click, any hour, no scheduling | API, MCP, Slack, WhatsApp, Zapier | Recording, transcript, assessment per question, integrity signals | $1 per credit, first 3 results free | | VidCruiter | One-way and live; same questions, same order; AI Interviewer labelled "COMING SOON" | Pre-recorded link 24/7; live needs booking | 75+ integrations incl. Workday, SAP SuccessFactors | Transcript; AI screening rationale | By request | | HireVue | AI interviewer with probing follow-ups; 200+ coding assessments | Mid-market to enterprise first rounds | Scores shown in the ATS | Scores tied to evidence; governance detail | None published | | Spark Hire | One-way and live; same questions for everyone | Unlimited interviews and users | 40+ ATS integrations | AI summary of each response | From $249 a month, billed annually | | Hireflix | One-way recorded answers | Unlimited responses and seats | Zapier, Make and ATS integrations | Automatic transcriptions | $75 or $150 a month, billed yearly | | Willo | Structured questions; lists follow-up questions | 150 assessed candidates per live role on Lite | 50+ native integrations | Custom scorecards | £49 per live role; Enterprise from £2,999 a year | | Sapia.ai | Text-only chat interview | Aimed at companies hiring 500+ a year | Not checked for this page | Scores traceable to the candidate's words | Per hire, no public price | ![Grid of interview format against public price for Tracy, HireVue, VidCruiter, Sapia.ai, Spark Hire and Hireflix.](/blog/images/vidcruiter-alternatives-2.png) *Where the shortlist sits on interview format and public price, from each vendor's own pages. Willo is left off because its pages list follow-up questions without saying how they work. Source: each vendor's pricing and product pages, Nov 2026.* Of the tools whose pages plainly describe a two-way interview, only Tracy also shows a price. ## What do the four pillars leave out? Other columns: vendors' own pages, read 9 Nov 2026. Tracy: approved product records. "Not stated" and "Not verified" mean the page does not say. | Criterion | Tracy AI Interviewer | VidCruiter | Verified edge elsewhere | | --- | --- | --- | --- | | Implementation and service | Enterprise includes a dedicated customer manager, a setup customised to your workflow and a walkthrough for your team. Tracy is customised in days per customer need. | Implementation included with a dedicated specialist; dedicated CSM, custom setup and training. | Not verified. | | Accessibility | Not stated. | States WCAG 2.2 AA. | Sapia.ai: untimed, free text, screen reader and speech-to-text compatible (vendor reply). | | Multilingual | A spoken-language check scores fluency, coherence, grammar, pronunciation and vocabulary from 0 to 100 beside the transcript. Tracy publishes no language count. | Notes and transcripts in English and French; live support in 8+ languages. | Willo lists 18+ languages. | | Security and compliance | SOC 2 Type II attested, with the report on request at hello@tracyinterview.ai. | SOC 2 Type II certified; data residency options. | HireVue lists FedRAMP; Willo ISO 27001. | | Support | Enterprise has a dedicated support channel and a dedicated customer manager. | Live agents respond within 1 minute in 8+ languages, around the clock. The stronger published claim. | Not verified. | ## How do you choose a VidCruiter alternative? Work down four questions. 1. Do you need live human interviews? Stay with VidCruiter or look at Spark Hire. 2. What must the first round show? If recorded answers to the same questions are enough, Hireflix or Spark Hire cost least. If you need to see how people reason in their field, you need a tool that asks the second question: Tracy, or HireVue at a quote. See [one-way vs two-way video interviews](/blog/one-way-vs-two-way-video-interviews). 3. How will you pay? A flat monthly plan favours steady volume. A per-result price like Tracy's $1 per credit favours uneven volume. 4. What must it connect to? Name your ATS and ask each vendor for the connector. The [Connect page](/connect) lists Tracy's. Then run one role: take the next 20 applicants through your top two and have one reviewer read both outputs. For wider lists, read [best AI interview software](/blog/best-ai-interview-software). Which path fits you: - Engineers and other technical roles: [technical interviews](/use-cases/technical-interviews). - Hundreds or thousands of applicants: [volume hiring](/use-cases/volume-hiring). - Subject experts to vet: [expert vetting](/use-cases/expert-vetting). - Programme or fellowship applications: [applications](/use-cases/applications). - A first-round budget: [Tracy's pricing](/pricing). ## Common questions ### What is the best VidCruiter alternative? It depends on the job. Tracy AI Interviewer suits technical and expert roles that need follow-ups and evidence at a published price. HireVue is the nearest enterprise suite, Spark Hire and Hireflix are the low flat-price recorded options, Willo charges per live role, and Sapia.ai runs a text-only interview. ### Is there a VidCruiter alternative with public pricing? Yes. Tracy costs $1 per credit with the first 3 results free. Spark Hire starts from $249 a month billed annually, Hireflix is $75 or $150 a month billed yearly, and Willo is £49 per live role. VidCruiter, HireVue and Sapia.ai publish no price. ### Which alternative works for technical interviews? Tracy can run a coding task inside the interview and ask follow-ups on approach and trade-offs. HireVue lists more than 200 on-demand coding skills assessments. ### Who are VidCruiter's competitors? HireVue competes for enterprise buyers, Spark Hire and Hireflix for recorded video interviews, Willo for per-role screening and Sapia.ai for text-based volume hiring. Tracy competes where the first round needs follow-ups. --- Send us one role and we build the interview plan, customised in days, with a recording and report you can check. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=vidcruiter-alternatives&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=vidcruiter-alternatives&utm_content=try-free) ## Sources - [Tracy AI Interviewer, Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Connect](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. # VidCruiter Pricing (Nov 2026) > VidCruiter pricing is quote-only. Tracy's is public: $1 per credit. See a four-pillar comparison and what to ask on a VidCruiter quote call. - URL: https://tracyinterview.ai/compare/vidcruiter-pricing - Published: 2026-11-09 - Author: Tracy team > **In short** Tracy AI Interviewer publishes her price: $1 per credit, one credit per started 20 minutes. 100 interviews of up to 20 minutes cost $100, 500 cost $500 and 1,000 cost $1,000. The first 3 results are free, with no seats and no upfront cost. > > VidCruiter publishes no price. Its pricing page is a request form for nine modules, and it says pricing is custom, based on the modules you pick and your hiring volume. > > Our reasoning, not a VidCruiter price: each $1,000 of first-year quote equals 1,000 Tracy interviews of up to 20 minutes. VidCruiter pricing is not public. Tracy AI Interviewer's is, on [Tracy's pricing](/pricing): $1 per credit, the first 3 results free, no seats and no upfront cost. This page sets out what VidCruiter's own pages say about cost, what Tracy costs at 100, 500 and 1,000 interviews, how the two compare, and what to ask on a VidCruiter quote call. ## What is public about VidCruiter pricing? Nothing that is a price. The pricing page says "Fill out the form to get pricing information on our solutions" and lists nine modules above a REQUEST PRICING button. The table lists the rest of what VidCruiter's pages say about cost. | What the page covers | What VidCruiter says | Exact words | | --- | --- | --- | | Plans and prices | No plans and no prices. A request form. | "Fill out the form to get pricing information on our solutions" | | What is priced | Nine modules: pre-recorded and live video interviews, scheduling, structured interviews, automated references, virtual hiring events, skills testing, video proctoring and an applicant tracking system. | "Fill out the form to get pricing information on our solutions, including:" | | Pricing basis | Custom, based on the modules you select and your hiring volume. | "custom pricing based on the modules you select and your organization's hiring volume" | ## What does Tracy cost at 100, 500 and 1,000 interviews? Tracy charges one credit for each started 20 minutes of an interview that produced a result, and a credit costs $1: a 15-minute interview is $1, a 34-minute one is $2. Credits never expire and a session with no answers is never charged. The first 3 results are free, so the first three interviews in any column cost nothing; the totals ignore that. | Interviews | Tracy, up to 20 minutes each (1 credit) | Tracy, 21 to 40 minutes each (2 credits) | VidCruiter | | --- | --- | --- | --- | | 100 | $100 | $200 | Not public. Quote on request. | | 500 | $500 | $1,000 | Not public. Quote on request. | | 1,000 | $1,000 | $2,000 | Not public. Quote on request. | ![Three figures: $100 for 100 interviews, $500 for 500 and $1K for 1,000, each up to 20 minutes.](/blog/images/vidcruiter-pricing-2.png) *Tracy cost at 100, 500 and 1,000 interviews of up to 20 minutes each. $1K is $1,000. Source: Tracy pricing page, Nov 2026.* Tracy's cost is a number you can compute today: $1 per interview of up to 20 minutes at any volume. ## How does a quote compare per interview? Our reasoning, not a VidCruiter price. A quote with several modules, implementation and a term comes down to one yearly number. Divide it by the interviews you complete. Each $1,000 of first-year quote equals 1,000 Tracy interviews of up to 20 minutes, so if you will complete fewer interviews than your quote has dollars, pay as you go costs less per interview. At 200 interviews a month, Tracy costs $2,400 a year. For wider context, see [AI interview cost](/blog/ai-interview-cost). ## How do VidCruiter and Tracy compare on the four pillars? Compare the two on the four pillars, then on the criteria they leave out. For the head-to-head, see [Tracy vs VidCruiter](/compare/tracy-vs-vidcruiter). | Pillar | Tracy AI Interviewer | VidCruiter | What to ask | | --- | --- | --- | --- | | Depth | Two-way video interview. She asks follow-ups in the person's own field and waits for the answer. Questions come from your brief or your list, and a coding task can sit inside. | Pre-recorded questions answered once on camera, or live interviews with a human, with the same questions in the same order. Its navigation labels the AI Interviewer "COMING SOON". | Ask for a live run on your own role. | | Volume | Every applicant starts on click, at any hour, with no scheduling. Ready in under 10 seconds. | Pre-recorded interviews use a unique link, 24/7. Live interviews use self-serve booking with calendar sync. | Ask what happens when 200 start in one day. | | Stack | API and MCP connections, Slack, WhatsApp and Zapier, beside the Workable and Breezy HR connectors. | 75+ integrations, including Workday, SAP SuccessFactors, Oracle and ADP, plus an API. | Name your ATS and ask which connector is included. | | Evidence | A recording of both sides, a timestamped transcript, a written assessment per question and integrity signals shown to a human reviewer. | A timestamped transcript, optional AI screening with a rationale, reviewer override and lockdown controls. | Open one result and get from the assessment to the sentence said. | | Price | $1 per credit, per started 20 minutes. First 3 results free, no seats, no upfront cost. | Not public: a request form, custom by modules and hiring volume. | Ask for the unit and the first-year total at your volume and at twice your volume. | ## What do the four pillars leave out? VidCruiter: its own pages, read 9 Nov 2026. Tracy: approved product records. "Not stated" means the page does not say. | Criterion | Tracy AI Interviewer | VidCruiter | Who has the edge on what we verified | | --- | --- | --- | --- | | Implementation and service | Enterprise includes a dedicated customer manager, a setup customised to your workflow and a walkthrough for your team. Tracy is customised in days per customer need. | Says implementation is included with a dedicated specialist; lists a dedicated CSM, custom setup and training. | Level; Tracy customises in days. | | Accessibility | Not stated. | States its platform meets WCAG 2.2 AA. | VidCruiter. | | Multilingual | A spoken-language check scores fluency, coherence, grammar, pronunciation and vocabulary from 0 to 100 beside the transcript. Tracy publishes no language count. | Notes, summaries and transcripts in English and French; live support in 8+ languages. Candidate-language count not verified. | VidCruiter on published languages. | | Security and compliance | SOC 2 Type II attested, with the report on request at hello@tracyinterview.ai. | SOC 2 Type II certified; data residency options. | Level on SOC 2 Type II. | | Support | Enterprise has a dedicated support channel and a dedicated customer manager. | Live agents respond within 1 minute in 8+ languages, around the clock. | VidCruiter. | ![Recreated Tracy interviews list: eleven backend engineer interviews, each Finished with its length in minutes and a View result link.](/blog/images/vidcruiter-pricing-1.png) *Each finished interview is a result, charged one credit per started 20 minutes. Illustration with sample data* Each finished interview is one row with its length in minutes, so the cost and the list sit side by side. ## What should you ask on a VidCruiter quote call? Start at the form on VidCruiter's pricing page, then ask: 1. Which modules are in the quote? Is AI candidate screening one of them? 2. What is the unit, and what is the first-year total at our volume and at twice our volume? 3. Is implementation included, and what does it cover? 4. What are the term, renewal terms and notice period? 5. Are ATS connectors and the API in the price? 6. What does the AI Interviewer cost, and can we run it live on our role? Its navigation labels it "COMING SOON". Bring the same list to every vendor. For other vendors, see [HireVue pricing](/compare/hirevue-pricing). ## What does VidCruiter say about contracts? On its page comparing itself with HireVue, VidCruiter says every product "can be bought standalone", there is "no minimum applicant volume commitment", implementation is included "with a dedicated specialist", and annual terms come "without multi-year lock-in". Get each in writing. Its video interviewing page also says smaller teams looking only for lightweight one-way interviewing "may find simpler tools more cost-effective." ## Is VidCruiter worth the quote, and for whom? For teams that run live panel interviews, need scheduling, references and testing from one vendor, or must connect to Workday or SAP SuccessFactors, VidCruiter is built for the job and its quote may be worth chasing. It says it is designed for "enterprise HR, government, and healthcare teams requiring compliance and structured rubrics." For a technical team that wants the first round to show how each applicant reasons, Tracy is the cheaper thing to test: the first 3 results are free, there is no seat to buy, and the next 100 cost $100. One customer cut its average time to hire from 44 days to 7, and Tracy did the work of 53 full-time recruiters in one month. Enterprise is priced to how many interviews you run; book a demo for a figure. Which path fits you: - Engineers and other technical roles: [technical interviews](/use-cases/technical-interviews). - Hundreds or thousands of applicants: [volume hiring](/use-cases/volume-hiring). - Subject experts to vet: [expert vetting](/use-cases/expert-vetting). - A first-round budget: [Tracy's pricing](/pricing). ## Common questions ### How much does VidCruiter cost? VidCruiter does not publish a price. Its pricing page is a request form for nine modules, and the company says pricing is custom, based on the modules you select and your hiring volume. We do not estimate a figure. ### Does VidCruiter have a free trial? Its pricing page does not mention one. Tracy's first 3 results are free, whatever their length, and no card is needed. ### What does Tracy cost for 1,000 interviews? $1,000 for interviews of up to 20 minutes at $1 per credit, with one credit for each started 20 minutes. There are no seats, no subscription and no upfront cost. ### What does Tracy Enterprise cost? Enterprise is priced to how many interviews you run, with a dedicated customer manager and a setup customised to your workflow. Book a demo for a figure. --- Send us one role and we build the interview plan, customised in days, with a recording and report you can check. The first 3 results are free, then $1 per credit. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=vidcruiter-pricing&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=vidcruiter-pricing&utm_content=try-free) ## Sources - [VidCruiter, Pricing](https://vidcruiter.com/pricing/), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. # CodeSignal Alternatives (Nov 2026) > Five CodeSignal alternatives compared in Nov 2026 for teams that want a conversation first: Tracy, HackerRank, Karat, CoderPad and HireVue, with prices where published. - URL: https://tracyinterview.ai/compare/codesignal-alternatives - Published: 2026-11-06 - Updated: 2026-11-11 - Author: Tracy team Checked on 3 Nov 2026 against each vendor's own pages ([see Sources](#sources)). > **In short** > - Choose **Tracy AI Interviewer** if you want a conversation with every applicant rather than a test. It runs a video interview from the link, asks CV-based questions, follows up on each answer and returns a recording, transcript and integrity signals with evidence. > - Stay with **CodeSignal** if validated coding assessments, a full IDE, proctoring and live technical interviews are the point. Its AI Interviewer sits in the same plans as those tests. > - HackerRank, Karat, CoderPad and HireVue each fit a different gap, listed below with what their own pages state. CodeSignal is a strong product, and the case for an alternative is narrow. Its certified assessments are written by subject matter experts and validated by I/O psychologists, each backed by 2,800 hours of research and testing. Candidates work in a full IDE, and AI use can be fully allowed or limited to syntax and documentation. Its live interviews run in a cloud IDE with video, audio and a library of 4,000+ questions across 70+ coding languages and frameworks. It also has a Suspicion Score, AI proctoring and ID verification, and its AI Interviewer has agents for sales, marketing, finance and other non-engineering roles. If that is what you hire on, none of the tools below replaces it. **On this page** - [Why do teams look beyond CodeSignal?](#why-do-teams-look-beyond-codesignal) - [Which alternative is best for what?](#which-alternative-is-best-for-what) - [Tracy AI Interviewer](#tracy-ai-interviewer) - [HackerRank](#hackerrank) - [Karat](#karat) - [CoderPad](#coderpad) - [HireVue](#hirevue) - [How do the five compare?](#how-do-the-five-compare) - [How to choose](#how-to-choose) - [Questions](#questions) - [Sources](#sources) ## Why do teams look beyond CodeSignal? The reasons below come from what CodeSignal's own pages state, not from review sites. - **The AI interview is one product inside a test platform.** CodeSignal sells an AI Interviewer, and its site also lists on-camera AI Video Avatars. The AI Interviewer page suggests four uses: a first pass for volume roles, a customer-facing screen before role-plays, Support and Ops scenarios, and an engineering behavioral screen before technical rounds. Coding sits in the assessments and live interviews, so a team that wants the interview to be the main product is buying a platform built around tests. - **Follow-ups follow your criteria.** CodeSignal says its AI Interviewer asks follow-up questions based on your criteria. That keeps scoring consistent. It also means the interview does not wander into what the candidate knows best. - **Credits hide the per-interview price.** A credit is used when a candidate starts an assessment, an AI interview or a live interview, and the pages do not say how many credits an AI interview takes. Our [CodeSignal pricing](/compare/codesignal-pricing) page works through the numbers. - **You may not need tests.** A team that wants a conversation with every applicant and reads the transcript may not want a platform built around assessments. ## Which alternative is best for what? | If you want | Look at | |---|---| | A video interview with follow-ups in the person's own field, priced per result | Tracy AI Interviewer | | A coding test plus an AI interviewer that works on a live canvas | HackerRank | | Human Interview Engineers on call for engineering hires | Karat | | Coding pads, with agentic interviews on higher plans | CoderPad | | One enterprise suite with governance and audit trails | HireVue | ## Tracy AI Interviewer **Good at:** an interview for every applicant that adapts. Tracy runs a two-way video interview with no install for the person being interviewed. Each interview starts when the person clicks the link, at any hour, so nobody waits for a slot. With CV check on, Tracy asks 3 to 5 questions tailored to the CV and the role's must-haves. It follows up on what the person says and waits for the end of their thought, can introduce a coding task and probe the trade-offs, and drafts the questions from your brief, which you can edit, or you bring your own. Each follow-up depends on the answer before it, where CodeSignal's pages describe follow-ups based on your criteria. The result is a recording, a transcript timed to the recording, a written assessment per question with its reasoning and integrity signals with their evidence, so you read the words and then decide. Interviews can start from an invite link, an apply page or your own system, and Enterprise plans include API and MCP connections and a setup customised to the workflow ([pricing](https://tracyinterview.ai/pricing)). **Price:** the first 3 results are free, then $1 per credit. One credit covers each started 20 minutes of a result, so a result up to 20 minutes costs $1. There is no seat fee, and a session with no answers is never charged. **Pick Tracy if** the interview itself is the product you need, you want every applicant interviewed in depth, you want the evidence behind each result, or you want to pay per result. **Pick something else if** you need validated tests or an IDE: Tracy can set a coding task and probe it, but it is not a test platform. Our [AI technical interviews guide](/blog/ai-technical-interviews) covers what a coding probe looks like. ## HackerRank **Good at:** coding tests and a technical AI interviewer in one place. Its Chakra AI Interviewer opens a canvas with a real-world problem and asks follow-ups on the work, and HackerRank's Chakra page contrasts this with AI interviewers it calls voice-only. HackerRank Interview adds pair-programming interviews. Chakra is included on Starter, Pro and Enterprise. **Price:** These HackerRank figures were read in a browser on 3 Nov 2026 and not re-checked on 4 Nov, when its pricing page blocked automated access. Starter is $79 a month billed annually with 60 candidate attempts a year. Pro is $419 a month billed annually with 360 attempts a year. Enterprise is custom. Extra attempts cost $20 each on Starter and Pro, or $15 each when pre-purchased on annual Pro. HackerRank lists a free trial. **Pick HackerRank if** you want the closest like-for-like swap for CodeSignal on engineering hires. **Pick something else if** you hire widely outside tech: the Chakra demo roles include a product manager and a customer experience engineer, but the product is positioned for technical hiring. The page does not state that Chakra uses candidate video. See also [HackerRank AI Interviewer alternatives](/compare/hackerrank-ai-interviewer-alternatives). ## Karat **Good at:** human technical interviews on demand. Karat's Interview Engineers run the interview, with consistent scoring rubrics, around the clock including evenings and weekends. Reports carry an overall recommendation, a quantitative score and written feedback per question, with video and code playback on the platform. A flagged interview is rewatched in full by a second trained Interview Engineer. **Price:** karat.com publishes no price and its calls to action are Request a Consult and Request demo. **Pick Karat if** you want expert humans on engineering interviews and do not want to build the process. **Pick something else if** you need an automated interview for every applicant: the pages we read describe interviews led by Interview Engineers and describe engineering evaluation only. More in [Karat alternatives](/compare/karat-alternatives) and [Tracy vs Karat](/compare/tracy-vs-karat). ## CoderPad **Good at:** pads for working code. CoderPad lists real-work, AI-assisted pads in 40+ languages and frameworks, with AI Assist on every plan. Frontier Lite and Enterprise unlock agentic interviews, and Frontier Lite adds AI summaries and interview transcripts. **Price:** Starter is $80 a month on annual billing (60 interviews a year) or $120 billed monthly (5 a month). Team is $400 a month annually (360 a year) or $500 monthly (30 a month). Frontier Lite is $1,000 a month, annual billing only, with 360 interviews a year. Enterprise is custom. Extra tests or interviews cost $30 each. **Pick CoderPad if** you want working-code pads and an agentic interview on a higher plan. **Pick something else if** you want the interview to be the main product. ## HireVue **Good at:** enterprise governance and scale. HireVue offers one-way on-demand video and live interviews, an AI Interviewer that evaluates each response against role-specific criteria and explains why, more than 200 coding skills assessments, and an "AI Informs. Humans decide." position. It lists annual independent audits and adverse impact testing before models go live, and it focuses on enterprise and mid-market. **Price:** none published. Its pricing page lists no plan names or prices and points to a plan conversation. See [HireVue alternatives](/compare/hirevue-alternatives) and [Tracy vs HireVue](/compare/tracy-vs-hirevue). **Pick HireVue if** your talent acquisition team needs governance, audit trails and one suite from engagement to interview. **Pick something else if** you want a published price. ## How do the five compare? | | Tracy | HackerRank | Karat | CoderPad | HireVue | |---|---|---|---|---|---| | Who interviews | AI, video, started on click at any hour | AI (Chakra) and humans | Human engineers | Pads; agentic interviews on Frontier Lite and Enterprise | AI, one-way video, live | | Coding environment | Can introduce a coding task | Agentic IDE, canvas | Code playback | Pads, 40+ languages | 200+ coding assessments | | Published price | $1 per credit, 3 free | $79 and $419 a month, annual | None | $80, $400 and $1,000 a month, annual; $30 per extra | None | | Plan fee | None | Monthly or annual plan | Not stated | Monthly or annual plan | Not stated | | Focus | Roles you describe in a brief | Tech hiring | Engineering | Coding pads | Enterprise and mid-market | HackerRank's prices and the $20 extra attempt were read in a browser on 3 Nov 2026 and not re-checked on 4 Nov. "Not stated" means the pages we read do not say, not that the vendor lacks it. Each cell comes from the vendor's own page on 3 Nov 2026; CodeSignal's own pricing is in [CodeSignal pricing](/compare/codesignal-pricing), and the side-by-side with Tracy is in [Tracy vs CodeSignal](/compare/tracy-vs-codesignal). ## How to choose Decide first whether you are replacing the test or adding to it. If CodeSignal's assessments work, keep them and add an interview only for roles where a test does not fit. If you are replacing it, match the tool to the part you valued: HackerRank for the closest swap, Karat for human judgment, CoderPad for pads, HireVue for enterprise governance, Tracy for the conversation and a per-result price. Run the same role through two tools before you sign a plan. Our [AI interview cost guide](/blog/ai-interview-cost) helps you compare price models. ## Common questions ### What is the best alternative to CodeSignal? It depends on what you want to replace. For an interview of every applicant with CV-tailored questions, follow-ups in the person's own field and evidence behind each result, Tracy. For a coding test and an AI interviewer from the same family, HackerRank. For human interviewers on call, Karat. CoderPad suits teams that want coding pads, and HireVue suits enterprise teams that need governance. ### Which CodeSignal alternatives publish a price? Tracy ($1 per credit, first 3 results free), HackerRank (Starter $79 a month billed annually, Pro $419 a month billed annually; read in a browser on 3 Nov and not re-checked on 4 Nov) and CoderPad (Starter $80, Team $400 and Frontier Lite $1,000 a month on annual billing) publish prices. Karat and HireVue publish none. ### Is there a CodeSignal alternative without a yearly plan? Tracy sells results by credit with no plan or seat fee. HackerRank and CoderPad sell plans with extra attempts at $20 (HackerRank, read 3 Nov, not re-checked) and $30 (CoderPad) each, and Karat and HireVue list no pay-per-result price in the pages we read. ### Can I use an alternative for roles outside engineering? Yes for Tracy, which interviews from a brief you write for the role, and for HireVue. CodeSignal also covers sales, marketing and other roles, and HackerRank is positioned for tech hiring. ### Should I switch from CodeSignal at all? Not if your hiring runs on its validated assessments and IDE. Add a second tool only for the part you need and CodeSignal does not cover, such as an interview priced per result. --- **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=codesignal-alternatives&utm_content=demo)** to run one of your roles through Tracy, or [try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=codesignal-alternatives&utm_content=try-free) with your first three results. ## Sources All sources retrieved 3 Nov 2026. - CodeSignal, [Pricing](https://codesignal.com/pricing/): credits, plans, AI interviewers by plan. - CodeSignal, [AI Interviewer](https://codesignal.com/ai-interviewer/): follow-ups based on your criteria, role agents, positioning before technical rounds. - CodeSignal, [Technical Assessments](https://codesignal.com/technical-assessments/): validated assessments, IDE, AI assistance settings. - CodeSignal, [Live Tech Interviews](https://codesignal.com/live-tech-interviews/): cloud IDE with video, audio and chat; 4,000+ questions; 70+ languages and frameworks. - CodeSignal, [Cheating and fraud](https://codesignal.com/cheating-and-fraud/): Suspicion Score, proctoring and ID verification. - Tracy AI Interviewer, [home](https://tracyinterview.ai/): video interviews, follow-ups that depend on the answer before, coding task with a probe on the trade-offs, your own questions, integrity signals, 'you read the words, then you decide'. - Tracy AI Interviewer, [product](https://tracyinterview.ai/product): two-way video interview, no install, CV check, three ways in, transcript timed to the recording. - Tracy AI Interviewer, [pricing](https://tracyinterview.ai/pricing): $1 per credit, 3 free results, billing by started 20 minutes, no-answer sessions not charged, Enterprise with API and MCP connections and a customised setup. - HackerRank, [Pricing](https://www.hackerrank.com/pricing): Starter, Pro and Enterprise prices, attempts, overage and Chakra inclusion (read in a browser on 3 Nov 2026; not re-checked on 4 Nov, the page served no prices to scripts). - HackerRank, [Chakra](https://www.chakra.sh/): canvas-based AI interviewer, comparison with voice-only tools and free trial. - Karat, [Karat Interview](https://karat.com/karat-interview/): human Interview Engineers, on-demand scheduling, no pricing page. - Karat, [home](https://karat.com/): engineering talent focus. - Karat, [Customer FAQ](https://karat.com/customer-faq): report contents and second-engineer review of flagged interviews. - CoderPad, [Pricing](https://coderpad.io/pricing/): Starter, Team and Frontier Lite prices, interviews per plan, $30 extra, pads, AI Assist and agentic interview plans. - HireVue, [AI Interviewer](https://www.hirevue.com/platform/ai-interviewer): explained evaluation, human decision statement, audits. - HireVue, [Pricing](https://www.hirevue.com/pricing): no published plans or prices; enterprise and mid-market focus. - HireVue, [Coding assessments](https://www.hirevue.com/products/coding-assessments): 200+ coding skills assessments. - HireVue, [Video interviewing](https://www.hirevue.com/platform/video-interviewing): one-way on-demand video and live interviews. # CodeSignal Pricing (Nov 2026) > What CodeSignal Hire costs in Nov 2026: Build, Grow and Pro, how credits work, the $20 extra credit, and a cost model at 100, 500 and 1,000 interviews next to Tracy. - URL: https://tracyinterview.ai/compare/codesignal-pricing - Published: 2026-11-06 - Updated: 2026-11-11 - Author: Tracy team Checked on 3 Nov 2026 against CodeSignal's own pricing and AI Interviewer pages ([see Sources](#sources)). > **In short** > - Tracy AI Interviewer has no plan or seat fee. Each applicant starts when they click the link and gets CV-tailored questions and answer-led follow-ups. Every result brings a recording, a transcript timed to it and integrity signals with evidence. > - CodeSignal Hire brings validated assessments, a full IDE, proctoring and live technical interviews, from **$79 a month billed annually**. Its credits cover assessments, AI interviews and live interviews alike, and extra credits cost **$20** each. Its pages do not say how many credits one AI interview uses, so the model below shows the floor. > - On price, Tracy has no plan or seat fee: the first 3 results are free, then **$1 per credit**, with one credit per started 20 minutes. CodeSignal sells validated coding assessments and an AI Interviewer inside the same plans. So one set of credits buys a test, an AI interview or a live session with a human. Tracy AI Interviewer is a different kind of purchase: an interview for each applicant, with CV-tailored questions, follow-ups and evidence behind every result. This page walks through CodeSignal's plans, shows how far the credits go at 100, 500 and 1,000 interviews next to Tracy, and says where the published numbers stop. **On this page** - [What does CodeSignal cost?](#what-does-codesignal-cost) - [How do CodeSignal credits work?](#how-do-codesignal-credits-work) - [What does the pricing page not say?](#what-does-the-pricing-page-not-say) - [What does it cost at 100, 500 and 1,000 interviews?](#what-does-it-cost-at-100-500-and-1000-interviews) - [How does Tracy pricing compare?](#how-does-tracy-pricing-compare) - [Is CodeSignal worth the price?](#is-codesignal-worth-the-price) - [Questions](#questions) - [Sources](#sources) ## What does CodeSignal cost? CodeSignal Hire has three plans. Build and Grow show a price; Pro is quoted. | Plan | Annual billing | Monthly billing | Credits | AI interviewers | |---|---|---|---|---| | Build | $79 a month | $99 a month | 60 a year, or 5 a month | Product, Design, Engineering | | Grow | $479 a month | $599 a month | 420 a year, or 35 a month | Adds Sales, CS, Marketing | | Pro | Custom | Custom | Custom | All | Build also lists access to 2,000 questions, AI proctoring and ID verification. Grow adds basic ATS integrations, and the page's Core ATS list names Gem and Ashby. Pro adds advanced fraud prevention, role-based access control and enterprise ATS integrations such as Workday, iCIMS and Oracle. Every plan states SOC 2 Type II, GDPR and CCPA compliance and encryption. Two things sit behind those cards. The annual price is a monthly figure, so Build is $948 a year and Grow is $5,748 a year. Dividing by the credits gives $15.80 per credit on Build and $13.69 on Grow with annual billing, and $19.80 on Build and $17.11 on Grow billed monthly ($99 for 5 credits, $599 for 35). These are our calculations from the published figures, not prices CodeSignal lists. ## How do CodeSignal credits work? A credit is used when a candidate starts an assessment, an AI interview or a live interview, not when you send the invitation. Credits are interchangeable across the three formats, so a team can spend them on whichever mix it needs, and an unopened invitation costs nothing. When the included credits run out, hiring does not stop. CodeSignal charges on-demand credits at $20 each, billed on the next cycle. That structure suits a team that wants one vendor for tests, AI screens and live technical interviews. The same flexibility makes the cost of any single format hard to isolate. ## What does the pricing page not say? - **Credits per AI interview.** The pricing page says what starts a credit. It does not give a credit count for an AI interview, and neither does the AI Interviewer page. - **Credits per assessment or live interview.** Not listed either, so the mix a plan covers cannot be worked out from the page. - **A free trial.** The plan cards offer Start with Build or Grow and carry no trial wording. The AI Interviewer page does let you try a live sample interview by phone or browser. - **Pro.** The price is custom, so nothing past Grow can be modeled. Because of the first gap, every CodeSignal cost per interview below depends on an assumption of ours: one credit for each started AI interview. If CodeSignal uses more, the cost rises by the formula in the next section. ## What does it cost at 100, 500 and 1,000 interviews? The model counts 12 months, annual billing, and one credit per AI interview that a candidate starts. Each table cell is a 12-month total. The cost is the annual fee plus $20 for every credit beyond the plan's allowance. Build includes 60 credits and Grow includes 420. | Interviews | Build | Grow | Lower | |---|---:|---:|---| | 100 | **$1,748** | $5,748 | Build, $17.48 each | | 500 | $9,748 | **$7,348** | Grow, $14.70 each | | 1,000 | $19,748 | **$17,348** | Grow, $17.35 each | The working: Build at 100 interviews is $948 plus 40 extra credits at $20. At 500 it is $948 plus 440 extra credits, and at 1,000 it is $948 plus 940. Grow at 100 interviews is the $5,748 fee with credits to spare. At 500 it is $5,748 plus 80 extra credits, and at 1,000 it is $5,748 plus 580. Pro could be cheaper or dearer at 1,000 and is left out because its price is not published. Of the three volumes, the lowest cost per interview falls at 500, close to where Grow's 420 credits run out. Past that point every extra interview costs $20, so the average climbs again. If an AI interview uses two credits, apply the same formula to double the credit count. At 500 interviews that is 1,000 credits, which on Grow is the same $17,348 the table shows for 1,000 interviews at one credit. ## How does Tracy pricing compare? Tracy AI Interviewer sells results. There is no seat fee and no plan fee, and each result is a full interview: the person starts when they click the link, at any hour, so a team can interview every applicant without booking anyone's calendar. With CV check on, Tracy asks 3 to 5 questions tailored to the CV and the role's must-haves, then follows up on what the person says and waits for the answer. The recruiter gets the recording, a transcript timed to the recording, a written assessment per question and integrity signals with their evidence, for a person to review. Teams can start an interview from an invitation link, an apply page or their own system, and Enterprise plans include API and MCP connections and a setup customised to the team's workflow ([pricing](https://tracyinterview.ai/pricing)). | Result length | Credits | Price | |---|---:|---:| | Up to 20 minutes | 1 | $1 | | 21 to 40 minutes | 2 | $2 | | 41 to 60 minutes | 3 | $3 | The first 3 results are free, credits never expire, and a session with no answers is never charged. For results of up to 20 minutes, 100 interviews cost 97 x $1 = $97 after the free three, 500 cost $497, and 1,000 cost $997. Longer results cost more: a result between 21 and 40 minutes takes two credits. CodeSignal says its AI phone screen typically runs 15 minutes and can be configured, so a 20-minute ceiling is a fair way to line the two up. Both comparisons are list prices. They leave out volume agreements and anything either vendor adds in a contract. The two products also price different things. CodeSignal's credits cover tests and live interviews with human interviewers as well as the AI interview. Tracy prices one thing, a video interview with a result you can check. Our [AI interview cost guide](/blog/ai-interview-cost) sets out how seat, plan and per-result models differ, and the [HireVue pricing](/compare/hirevue-pricing) and [Ribbon pricing](/compare/ribbon-pricing) pages show two vendors that publish less. ## Is CodeSignal worth the price? For a team whose hiring is mostly engineering and who wants validated assessments, a full IDE, proctoring and live technical interviews under one contract, the Build and Grow plans are the place to start. CodeSignal states that its certified assessments are validated by I/O psychologists and backed by 2,800 hours of research each, and it lists a 4,000+ question library for live interviews. A team that needs that depth should price CodeSignal first and buy Tracy only if it also wants interviews for roles where a test does not fit. Tracy fits when the job is the conversation itself: interviewing every applicant, in the person's own field, and reading what each one said with a recording and transcript to check it against. On the modeled numbers it also costs less, without a yearly commitment. For how the two products differ beyond price, read [Tracy vs CodeSignal](/compare/tracy-vs-codesignal) or the [CodeSignal alternatives](/compare/codesignal-alternatives) shortlist. If your first rounds include coding, [AI technical interviews](/blog/ai-technical-interviews) explains how Tracy handles a coding task. ## Common questions ### How much does CodeSignal cost? CodeSignal Hire lists Build at $79 a month billed annually ($99 billed monthly) and Grow at $479 a month billed annually ($599 billed monthly). Pro is custom. An extra credit costs $20. ### What is a CodeSignal credit? A credit is used when a candidate starts an assessment, an AI interview or a live interview. Credits work across all three formats, and an invitation nobody opens costs nothing. ### How many credits does one AI interview use? The pricing and AI Interviewer pages do not say. The cost model on this page assumes one credit as a floor; ask CodeSignal for the number in writing. ### Does CodeSignal have a free trial? The pricing page offers Start with Build or Grow and shows no trial wording. The AI Interviewer page lets you try a live sample interview. ### How does Tracy pricing compare? Tracy has no plan or seat fee. The first three results are free, then $1 per credit, and one credit covers each started 20 minutes of a result. A session with no answers is not charged. --- **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=codesignal-pricing&utm_content=demo)** to price your own volume, or [try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=codesignal-pricing&utm_content=try-free) with your first three results. ## Sources All sources retrieved 3 Nov 2026. - CodeSignal, [Pricing](https://codesignal.com/pricing/): Build, Grow and Pro prices, credit counts, what uses a credit, $20 on-demand credit, plan inclusions, ATS integrations and security wording. - CodeSignal, [AI Interviewer](https://codesignal.com/ai-interviewer/): typical length of the AI interview, live sample interview and what the report contains. - CodeSignal, [Technical Assessments](https://codesignal.com/technical-assessments/): validated assessments and 2,800 hours of research per assessment. - CodeSignal, [Live Tech Interviews](https://codesignal.com/live-tech-interviews/): 4,000+ question library across 70+ coding languages and frameworks. - Tracy AI Interviewer, [home](https://tracyinterview.ai/): follow-ups that depend on the answer before, integrity signals list, 'you read the words, then you decide', starts at any hour. - Tracy AI Interviewer, [product](https://tracyinterview.ai/product): CV check with 3 to 5 questions, three ways in (invite link, apply page, your own system), transcript timed to the recording, video interview. - Tracy AI Interviewer, [volume hiring](https://tracyinterview.ai/use-cases/volume-hiring): each applicant starts on click at any hour, no recruiter and no calendar. - Tracy AI Interviewer, [pricing](https://tracyinterview.ai/pricing): $1 per credit, 1 credit per started 20 minutes, first 3 results free, no-answer sessions not charged, no seats, Enterprise with API and MCP connections and a customised setup. # HackerRank AI Interviewer Alternatives (Nov 2026) > HackerRank's Chakra asks follow-ups on candidates' code. Here is when to pick it and when Tracy AI Interviewer, CodeSignal, CoderPad or HireVue fits better. - URL: https://tracyinterview.ai/compare/hackerrank-ai-interviewer-alternatives - Published: 2026-11-06 - Updated: 2026-11-11 - Author: Tracy team > **In short** > - Pick **Tracy AI Interviewer** when you are leaving HackerRank's Chakra for non-engineering roles, or for interviews with every applicant. It can tailor questions from the CV, follow up on each answer and leave a recording with a transcript timed to it. > - Pick **HackerRank's Chakra** when the hire is an engineer and you want to watch them work in a real codebase. It asks follow-ups on that work and gives a reason for each score. > - CodeSignal, CoderPad and HireVue each cover a different gap, compared below. HackerRank's Chakra is a strong product, and some teams should stay with it. If you are searching for alternatives, ask what Chakra assumes about your hiring, and whether your roles match. Facts about every tool come from its own pages, read on 3 Nov 2026. Tracy AI Interviewer's case is reach and depth: it interviews everyone who applies on video, reads the CV to tailor its questions, follows up on what each person says and leaves a recording, a transcript timed to the recording and a written assessment per question with its reasoning. ## What does HackerRank's Chakra do well? HackerRank describes Chakra as "an agentic interviewer that spins up a canvas with a real-world problem, observes how candidates work and evaluates judgment, critical thinking & AI fluency." Its own comparison with other AI interviewers is direct: most are voice-only, and Chakra "opens a live canvas like a dev environment and watches candidates actually do the work," then asks follow-up questions on that work. That is a real strength for engineering hiring. From HackerRank's own pages: - Follow-ups on the work. Real-time follow-ups are listed on every plan, and Chakra's follow-ups are on the candidate's work. - A reason for every score. Chakra gives "clear scores for each competency, a summary of the candidate's strengths and gaps, and a clear 'why' for every score," backed by transcripts and work. - Integrity. It flags unauthorized apps such as Cluely and help from an external device or another person. - Your questions and your ATS. Its FAQ says you can use your own questions or the Chakra library, and that it connects with popular ATS providers directly, though the pricing page lists ATS and calendar integrations from Pro up, not on Starter. The pricing page lists 55+ languages and, on Enterprise, a 7,500+ question library. - A free trial. The Chakra page has a "Start free trial" button next to "Book a demo". - Scale, as HackerRank states it. 500k+ candidates interviewed, rated 4.8 out of 5 on average. Both are vendor figures, and no independent survey backs them. HackerRank also says it runs frequent bias analyses with third-party auditors "in compliance with the NYC Bias audit law (144)." That is HackerRank's statement about its own system, and we do not repeat it as a claim about any other tool. ## Why look beyond Chakra? No independent reviews of Chakra turned up and no complaints came with them, so what follows are questions of fit. - Are all your hires engineers? Chakra's canvas is a dev environment. HackerRank's Chakra page also says you can turn any job description into a tailored Chakra with custom sections, questions and rubrics, so non-engineering roles are possible. We read that as partial coverage, because the product is built around hands-on work. - Do you want a plan, or a per-credit price? HackerRank sells plans of candidate attempts, and a completed attempt counts against your allowance. - Do you want the interview itself to be a video conversation? Chakra's pages describe a canvas and a conversation without naming a format, so confirm with HackerRank which formats it runs. - Is the AI interview a first conversation, or a work sample? Chakra puts a task in front of the candidate. Some first rounds only need a conversation. ## Which alternative fits which team? | Tool | Best for | | --- | --- | | HackerRank Chakra | Engineers working in a real codebase, with follow-ups on their code and a reason for each score | | Tracy AI Interviewer | Interviewing every applicant on video when they click, with CV-aware questions, follow-ups and a recording, transcript and written assessment per question, across hiring, expert vetting and accelerator applications | | CodeSignal | Teams that want AI interviewers, assessments and live coding on one credit pool | | CoderPad | Engineering teams that run live pads in 40+ languages and want agentic interviews on a higher plan | | HireVue | Enterprise hiring that wants an AI interviewer, one-way and live video and coding assessments in one suite | ### Tracy AI Interviewer Tracy runs a two-way video interview. Its product page says each follow-up depends on the answer before it and a vague answer gets a sharper question, and the person being interviewed needs no install. Tracy can introduce a coding task and probe the trade-offs. Its site lists use cases beyond engineering: hiring at volume, expert vetting that checks domain depth before an expert joins a project, applications to accelerators and fellowships, and customer research. Where Tracy differs from Chakra, on named dimensions: - Volume. Tracy interviews everyone who applies: each person starts when they click the link, at any hour, with no scheduling and no recruiter present, so a role with hundreds of applicants is interviewed in full instead of cut to a shortlist first. HackerRank sells plans with a set number of candidate attempts a year (60 on Starter, 360 on Pro) and $20 for each extra attempt. - CV and follow-ups. Tracy can read the CV and ask 3 to 5 questions tailored to it and to the role's must-haves, then follow up on what the person says and wait for the end of an answer. Chakra's follow-ups are on the candidate's work in the canvas, which is its strength for engineers. We did not find a CV check described on Chakra's pages. - Evidence. Each interview leaves the recording of both sides, a transcript timed to the recording and a written assessment per question with its reasoning. On video interviews Tracy also marks integrity signals with their evidence on the recording: tab changes and fullscreen, face and gaze (live face detection and where the person is looking), a long delay before an answer and an external screen, listed on [Tracy's home page](https://tracyinterview.ai/). They are signals, not a verdict, and we give no detection rate. Chakra also flags unauthorized apps and outside help and gives a reason for each score, so neither tool has an edge on integrity. Teams that want numeric competency scores out of the box should look at Chakra. - Stack. Tracy offers API and MCP connections, and an interview can start from an invitation link, an apply page or your own system. Enterprise teams can ask for a customised setup. Chakra connects too: its FAQ says it works with popular ATS providers and offers a Chakra MCP, and HackerRank's pricing page lists API and MCP access on Enterprise, with ATS and calendar integrations from Pro up. Compare the plan that includes them. - Format. Tracy runs a two-way video interview, with no install for the person being interviewed. Chakra's pages name no interview format, so confirm it with HackerRank. - Range of use. Tracy's pages list hiring, expert vetting, applications and customer research. Chakra's pages centre on engineering work. Pick Chakra instead if your core need is watching an engineer work through a real task in an IDE with files and a terminal. Tracy's pages describe coding as a task Tracy can introduce and probe, and they do not describe an IDE like that. Tracy's integrity signals watch the video interview, not a coding canvas, so if monitoring inside a coding environment is a requirement, ask each tool how it handles it before you decide. For a longer view of Tracy on technical screens, read [AI technical interviews](/blog/ai-technical-interviews) and [AI interview cheating detection](/blog/ai-interview-cheating-detection). ### CodeSignal CodeSignal's Build plan is $79 a month billed annually with 60 annual credits, and a credit is used when a candidate starts an assessment, an AI interview or a live interview. Build includes Product, Design and Engineering AI Interviewers; Grow adds Sales, CS and Marketing ones. Its AI Interviewer is typically 15 minutes, probes "if answers are thin," and produces a report with an overall score; recruiters decide. Its assessments run in a full IDE, and its live interviews use an AI-native cloud IDE with a human interviewer. We did not find a coding task stated inside the AI Interviewer itself. Pick CodeSignal if you want assessments, AI interviews and live interviews on one credit pool. See [CodeSignal alternatives](/compare/codesignal-alternatives) and [Tracy vs CodeSignal](/compare/tracy-vs-codesignal). ### CoderPad CoderPad's pricing page lists "real-work, AI-assisted pads in 40+ languages and frameworks," with AI Assist on every plan and agentic interviews unlocked on its Frontier Lite and Enterprise plans. AI summaries and interview transcripts come with Frontier Lite. Its pricing page lists Starter at $80 a month billed annually (60 interviews a year), Team at $400 (360 a year) and Frontier Lite at $1,000 (360 a year, annual billing only), each with a free trial, and $30 per additional test or interview. Monthly billing costs more: $120 for Starter and $500 for Team. Pick CoderPad if your engineers run the interview live and want a pad with AI help built in. ### HireVue HireVue's AI Interviewer has AI agents ask and record candidate responses, "including probing follow-up questions, in a natural two-way conversation," and it explains why each response meets or misses your criteria. For code, HireVue lists more than 200 OnDemand coding skills assessments in languages such as Python, Java and JavaScript, and says it combines coding challenges and interviews in one experience. It publishes no price; its pricing page points to a plan conversation. Pick HireVue if you are an enterprise team that wants one suite for AI interviews, video and coding assessments. Our [Tracy vs HireVue](/compare/tracy-vs-hirevue) page covers that comparison. ## How do the prices compare? | | Published price | What you get for it | | --- | --- | --- | | HackerRank | Starter $79 a month billed annually ($948) for 60 attempts a year; Pro $419 a month billed annually ($5,028) for 360 attempts a year; extra attempts $20 each (Pro on an annual plan can pre-purchase them at $15) | Chakra, Screen and Interview are included on every plan | | Tracy AI Interviewer | $1 per credit; one credit covers up to 20 minutes, $3 for a 41 to 60 minute session | First 3 results free; no-answer sessions never charged; credits never expire | | CodeSignal | Build $79 a month billed annually (60 credits); Grow $479 a month billed annually (420 credits); $20 per extra credit | A credit covers an assessment, AI interview or live interview | | CoderPad | Starter $80 a month billed annually (60 interviews a year); Team $400 (360 a year); Frontier Lite $1,000 (360 a year); $30 per additional test or interview | Agentic interviews on Frontier Lite and Enterprise | | HireVue | None published | Plan conversation | If you use every HackerRank attempt, Starter works out to $15.80 each ($948 divided by 60) and Pro to $13.97 ($5,028 divided by 360). That is our arithmetic, not a HackerRank figure. The units also differ: HackerRank charges per completed attempt, and Tracy charges per started 20 minutes. Use it as a rough guide and test one role in each tool. ## How should you choose? 1. What does the hire have to show? A built solution in a codebase points to Chakra or CodeSignal's IDE. A reasoned conversation about a field points to Tracy. 2. Which roles? Only engineers: start with Chakra. A mix of engineers, experts, applicants to accelerators and non-technical roles: Tracy or HireVue. 3. How many people should get an interview? A set number of engineers a year fits a plan of attempts. If everyone who applies should be interviewed, Tracy starts each interview when the person clicks, and its per-credit pricing, with credits that never expire, follows uneven volume. 4. What can you try without a call? HackerRank has a free trial on the Chakra page, CoderPad has one on Starter, Team and Frontier Lite, and Tracy gives 3 free results. Run the same role through two tools and compare what your reviewers would decide. For the review side, see [AI interview reports](/blog/ai-interview-reports), and for engineering questions to pair with either tool, [software engineer interview questions](/blog/software-engineer-interview-questions). ## Common questions ### What does HackerRank's Chakra AI Interviewer do? HackerRank describes Chakra as an agentic interviewer that opens a live canvas with a real-world problem, watches how a candidate works, asks follow-up questions on that work and gives scores per competency with a stated reason for each score. It flags unauthorized apps and help from another person or device, and it takes your own questions. ### How much does HackerRank's Chakra cost? HackerRank's pricing page lists Chakra as included on Starter ($79 a month, billed annually at $948, 60 candidate attempts a year), Pro ($419 a month, billed annually at $5,028, 360 attempts a year) and Enterprise (custom). Extra attempts are $20 each on Starter and Pro. ### What is a good alternative to HackerRank's AI interviewer for non-engineering roles? Tracy AI Interviewer interviews everyone who applies, on video, whenever they click the link, for hiring first rounds, expert interviews, accelerator and fellowship applications and customer research. It can read the CV, tailor questions to it and follow up on answers, and it leaves a recording, transcript and written assessment per question. It costs $1 per credit (one credit per started 20 minutes) with the first 3 results free. HireVue and CodeSignal also list AI interviewers for roles beyond engineering. ### Does Tracy replace HackerRank for coding tests? No. Tracy can introduce a coding task and probe the trade-offs behind an answer, and it can interview every applicant who clicks, but its pages do not describe an IDE with a terminal and files. If you need candidates to build in a real codebase, HackerRank's Chakra is built for that. ### Which HackerRank alternative has a free trial? HackerRank offers a free trial on the Chakra page, and CoderPad lists one on Starter, Team and Frontier Lite. Tracy gives your first 3 results free, and needs no card to start. We did not find a free trial stated on CodeSignal's AI Interviewer page, which offers a self-demo. --- **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=hackerrank-ai-interviewer-alternatives&utm_content=demo)** to see Tracy interview a candidate for one of your roles, or [try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=hackerrank-ai-interviewer-alternatives&utm_content=try-free) with your first 3 results on us. ## Sources All sources retrieved 3 Nov 2026. - HackerRank, [Pricing](https://www.hackerrank.com/pricing) (blocks scripted fetches; figures read in a real browser on 3 Nov 2026 and kept in the pipeline snapshot): plan prices, attempts, extra-attempt price, Chakra included on every plan, question libraries, 55+ languages, real-time follow-ups, SOC 2 listing, ATS integrations from Pro, the $15 pre-purchase price, and the FAQ definition of an attempt. - HackerRank, [Chakra](https://www.chakra.sh/): what Chakra does, how it differs from voice-only tools, scores with a reason, integrity flags, own questions, ATS, free trial, 500k+ candidates and 4.8 rating (vendor figures), bias-audit wording (HackerRank's own statement). - HackerRank, [Interview](https://www.hackerrank.com/products/interview/): live pair programming in a real codebase. - CodeSignal, [Pricing](https://codesignal.com/pricing/): Build and Grow prices and credits, credit rule, extra-credit price, AI Interviewers per plan. - CodeSignal, [AI Interviewer](https://codesignal.com/ai-interviewer/): length, probing, report with overall score, recruiters decide. - CodeSignal, [Technical assessments](https://codesignal.com/technical-assessments/): full IDE. - CodeSignal, [Live tech interviews](https://codesignal.com/live-tech-interviews/): AI-native cloud IDE with a human interviewer. - CoderPad, [Pricing](https://coderpad.io/pricing/): plan prices and interview allowances, extra-interview price, free trials, pads in 40+ languages, agentic interviews on higher plans, summaries and transcripts. - HireVue, [AI Interviewer](https://www.hirevue.com/platform/ai-interviewer): two-way conversation with probing follow-ups, explanation of each response. - HireVue, [Coding assessments](https://www.hirevue.com/products/coding-assessments): 200+ OnDemand coding assessments, coding and interviews combined. - HireVue, [Video interviewing](https://www.hirevue.com/platform/video-interviewing): one-way on-demand video and live interviews. - HireVue, [Pricing](https://www.hirevue.com/pricing): no published plans or prices. - Tracy AI Interviewer, [home page](https://tracyinterview.ai/): integrity signals list, use cases, 'you read the words, then you decide'. - Tracy AI Interviewer, [Product page](https://tracyinterview.ai/product): two-way video interview, no install for the person being interviewed, CV check, follow-ups that depend on the previous answer, recording with a timed transcript, three ways in, API and MCP. - Tracy AI Interviewer, [Hiring at volume](https://tracyinterview.ai/use-cases/volume-hiring): everyone who applies, on click, at any hour, no recruiter and no calendar, the same depth for applicant one and applicant five hundred. - Tracy AI Interviewer, [Technical interviews](https://tracyinterview.ai/use-cases/technical-interviews): a coding task, then a probe on the trade-offs. - Tracy AI Interviewer, [Expert vetting](https://tracyinterview.ai/use-cases/expert-vetting): check domain depth before an expert joins a project. - Tracy AI Interviewer, [Connect page](https://tracyinterview.ai/connect): MCP server and API connection. - Tracy AI Interviewer, [Pricing](https://tracyinterview.ai/pricing): credit price, billing unit, free results, no card needed, credits never expire, no-answer sessions never charged, Enterprise API and MCP connections and customised setup. # Karat Alternatives (Nov 2026) > Karat sells interviews run by expert human engineers. Compare Tracy AI Interviewer, HackerRank, CodeSignal, CoderPad and HireVue by who runs the interview, price and fit. - URL: https://tracyinterview.ai/compare/karat-alternatives - Published: 2026-11-06 - Updated: 2026-11-11 - Author: Tracy team > **In short** > - Pick **Tracy AI Interviewer** when you want to replace booked expert slots with an interview each applicant starts on click, at any hour. It can read the CV to tailor questions, ask follow-ups that depend on the previous answer and write an assessment per question with its reasoning. > - Pick **Karat** when expert human interviewers, one standard rubric and benchmark data from Karat's interview pool matter more than volume. > - HackerRank, CodeSignal, CoderPad and HireVue sit between the two; each is compared below. Karat has a clear model: it supplies the interviewer. Its site describes "live, human-led interviews conducted by experienced, certified interview engineers." Anyone searching for alternatives is usually asking one of two things: how to interview more people than human interviews allow, or how to get the same consistency for less. Facts come from each vendor's own pages, read on 3 Nov 2026, except where we mark a number as third-party. ## What does Karat do well? - Expert human interviewers. Karat's interviews are run by its own Interview Engineers. Its home page says every engineer, "regardless of team, geography, or vendor, is held to a rigorous, consistently enforced technical bar by expert interviewers." - A standard rubric. Karat describes structured interview formats and consistent scoring rubrics. Its interview report carries an overall recommendation, a quantitative score and written feedback on the candidate's response to each question. - Availability. Its interviewing is 24/7 and on demand, same-day, evenings and weekends. - Scale and benchmarks. Karat's home page says 600,000+ technical interviews and invites you to see how your talent measures up against them. A customer quote on the same page credits Karat with interview performance data and competitive benchmarking across roles and geographies. Its product page quotes a lower figure, 350k+ interviews in 70+ countries with 500+ interview engineers. These are Karat's own numbers and we did not verify either. - Integrity with a human check. Its Interview Engineers watch for behaviours such as typing outside the browser window and looking between monitors, and a flagged interview is rewatched in full by a second trained engineer. - AI in the loop, not in charge. Karat says AI never makes final hiring decisions. Its NextGen page describes human-led, AI-enabled interviews, and its home page lists live AI-native simulations led by its Interview Engineers. - An earlier step. Karat Qualify is an on-demand, adaptive multiple-choice screen of 15 questions, 15 to 30 minutes for most candidates, that you can send to all applicants before the Karat Interview. It is a multiple-choice screen: no code test and no AI interviewer. - Security paperwork. Karat says it provides a SOC 2 Type II attestation report on request. ## Why do teams look for Karat alternatives? Karat has no independent reviews we could cite and we saw no complaints, so what follows are questions of fit. - There is no public price. Karat's site has no pricing page; its calls to action are Request a Consult and Request demo. Vendr, a third-party procurement platform, reports a median buyer spend of $175,695 a year, a range of $57,000 to $1,069,900, and per-interview rates of $200 to $450 for a standard technical screen. These are Vendr's figures, not Karat's, and your terms will differ. - A human interview per candidate is a limit on volume. At those third-party rates, a team may reserve the Karat Interview for candidates who already passed an earlier step, such as Karat's own Qualify screen. - The scope is engineering. Karat describes evaluating engineering talent. If you also interview experts, founders or non-technical hires, you need something else for those. - You may want the interviewer to be software. Karat's pages do not describe an AI interviewer that replaces its engineers. ## Which Karat alternative fits which team? | Tool | Who runs the interview | Best for | | --- | --- | --- | | Karat | Karat's Interview Engineers | Engineering teams that want expert human interviewers and one standard rubric | | Tracy AI Interviewer | AI, on video | A real first interview for every applicant on click, CV-aware and with follow-ups, in engineering and other fields | | HackerRank | Chakra (AI) or your own team in pair programming | Watching engineers work in a real codebase | | CodeSignal | AI Interviewer, or your team in a live IDE | AI interviews, assessments and live coding on one credit pool | | CoderPad | Your own team, with agentic interviews on higher plans | Teams that run live technical interviews in-house | | HireVue | AI Interviewer, or live video | Enterprise teams that want AI interviews and coding assessments in one suite | We did not research other outsourced human-interview services, so this list covers software that replaces or reduces the human step. ### Tracy AI Interviewer Tracy runs a two-way video interview. There is no booking step: each person starts from a link when they are ready, at any hour, and the person being interviewed needs no install. A vague answer gets a sharper question, and each follow-up depends on the answer before it. Tracy can introduce a coding task and probe the trade-offs, and its site lists hiring at volume, expert vetting, applications to accelerators and fellowships, and customer research. What Tracy adds to a first round, against Karat's model of one booked engineer per interview: - Volume. Every applicant is interviewed when they click the link, at any hour, with no scheduling and no recruiter present, so hundreds of applicants can each get a full interview, not a shortlist. - Depth. Tracy can read the CV and ask 3 to 5 questions tailored to it and to the role's must-haves, then follows up on what the person says and waits for the end of an answer. - Evidence. Each interview leaves a recording, a transcript timed to the recording and a written assessment per question with its reasoning. You read the words, then you decide. - Stack. API and MCP connections, three ways to start an interview (an invitation link, an apply page or your own system) and customised setups for enterprise teams. Price is published too: $1 per credit, credits never expire, one credit per started 20 minutes, first 3 results free, and a session with no answers is never charged. Pick Karat instead if you need an experienced engineer to judge a candidate's code and reasoning, or you want Karat's benchmark data. Tracy's pages describe a coding task it can introduce and probe, not an expert's review of the code, and they do not offer benchmarks across other companies' candidates. Karat's engineers also watch for cheating; Tracy's video interviews carry integrity signals with their evidence for a reviewer instead: tab changes and fullscreen, face and gaze (live face detection and where the person is looking), a long delay before an answer and an external screen, listed on [Tracy's home page](https://tracyinterview.ai/). They are signals, not a verdict, and we give no detection rate. For the AI-versus-human question in general, read [AI interview vs human interview](/blog/ai-interview-vs-human-interview), and for Tracy on technical screens, [AI technical interviews](/blog/ai-technical-interviews). The head-to-head is in [Tracy vs Karat](/compare/tracy-vs-karat). ### HackerRank HackerRank's Chakra is "an agentic interviewer that spins up a canvas with a real-world problem," asks follow-ups on the work, and gives a reason for each score. Its Interview product is live pair programming in a real codebase, run by your own engineers. Starter is $79 a month billed annually for 60 candidate attempts a year; Pro is $419 a month billed annually for 360, with Chakra included on every plan. Pick HackerRank if you want to keep your own engineers in the room or want an AI to run a hands-on task. Our [HackerRank AI Interviewer alternatives](/compare/hackerrank-ai-interviewer-alternatives) page goes deeper. ### CodeSignal CodeSignal's Build plan is $79 a month billed annually with 60 annual credits, usable on an assessment, an AI interview or a live interview. Live interviews run in an AI-native cloud IDE with a human interviewer, which is your own. The AI Interviewer is typically 15 minutes and produces a report with an overall score, and recruiters decide. Pick CodeSignal if you want assessments, AI interviews and live coding on one credit pool. See [CodeSignal alternatives](/compare/codesignal-alternatives). ### CoderPad CoderPad's pricing page lists AI-assisted pads in 40+ languages and frameworks, with agentic interviews on its Frontier Lite and Enterprise plans and AI summaries and transcripts on Frontier Lite. Its pricing page lists Starter at $80 a month billed annually (60 interviews a year), Team at $400 (360 a year) and Frontier Lite at $1,000 (360 a year), with $30 per additional test or interview. Pick CoderPad if your engineers will keep running the interview and you want a pad that supports it. ### HireVue HireVue's AI Interviewer has AI agents ask and record answers, "including probing follow-up questions, in a natural two-way conversation," and explains why each response meets or misses your criteria. It lists more than 200 OnDemand coding skills assessments and says it combines coding challenges and interviews in one experience. It publishes no price. See [Tracy vs HireVue](/compare/tracy-vs-hirevue) and [HireVue alternatives](/compare/hirevue-alternatives). Pick HireVue if you are an enterprise team that wants an AI interview, live video and coding assessments from one vendor. ## How should you choose? 1. Who should run the interview? If the answer is an expert engineer, Karat is built for that. If it is software, so every applicant can be interviewed, look at Tracy, Chakra, CodeSignal or HireVue. 2. What do you need to see? Code worked through in an IDE points to HackerRank or CodeSignal. A reasoned conversation about a field points to Tracy or HireVue. 3. How many people will you interview? Third-party Karat rates of $200 to $450 per interview suit finalists. Tracy starts an interview whenever an applicant clicks, with no slot to book, and $1 per credit, one credit per started 20 minutes, suits a first round for everyone, which is the case in [how to screen candidates](/blog/how-to-screen-candidates). 4. Do you need to share a price with finance before a call? Tracy, HackerRank, CodeSignal and CoderPad publish prices on their sites; Karat and HireVue do not. Many teams will use both kinds: an AI first round for everyone, then expert humans for finalists. [AI interview cost](/blog/ai-interview-cost) puts the published prices side by side. ## Common questions ### What does Karat sell? Karat sells technical interviews run by its own Interview Engineers, who are human, with consistent scoring rubrics. It offers on-demand interviewing, same-day and on evenings and weekends, and an interview report with an overall recommendation, a score and written feedback per question. ### How much does Karat cost? Karat does not publish a price. Its site offers Request a Consult and Request demo. Vendr, a third-party procurement platform, reports a median buyer spend of $175,695 a year and per-interview rates of $200 to $450 for a standard technical screen, depending on volume, term and tier. Those are Vendr's figures, not Karat's. ### Is there an AI alternative to Karat? Yes. Tracy AI Interviewer, HackerRank's Chakra, CodeSignal's AI Interviewer and HireVue's AI Interviewer all run the interview without a human interviewer. Each has different strengths; Chakra and CodeSignal are built around code, Tracy and HireVue around conversation. ### Can an AI interviewer replace Karat's expert interviewers? It depends on the round. Karat's strength is expert human judgement and a standard rubric on engineering roles. An AI interviewer such as Tracy fits a first conversation with every applicant, started on click and built from the role and the CV. Many teams will want both: an AI first round for volume and expert humans for finalists. ### Does Karat work for roles outside engineering? Karat's site describes evaluating engineering talent. Tracy AI Interviewer lists hiring first rounds, expert interviews, accelerator and fellowship applications and customer research. --- **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=karat-alternatives&utm_content=demo)** to see Tracy interview a candidate for one of your roles, or [try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=karat-alternatives&utm_content=try-free) with your first 3 results on us. ## Sources All sources retrieved 3 Nov 2026. - Karat, [home page](https://karat.com/): 600,000+ interviews (vendor figure), the compare-globally wording, a customer quote on benchmarking, standards wording, live AI-native simulations, engineering scope. - Karat, [Karat Interview](https://karat.com/karat-interview/): human-led interviews, 24/7 on-demand, rubrics, 350k+ interviews and 500+ engineers (vendor figures), no pricing page. - Karat, [Customer FAQ](https://karat.com/customer-faq): report contents, cheating checks and second review, AI never decides, SOC 2 Type II attestation report on request. - Karat, [NextGen](https://karat.com/evaluate-ai-ready-engineers/): human-led, AI-enabled interviews. - Karat, [Qualify](https://karat.com/karat-qualify/): on-demand adaptive multiple-choice screen, 15 questions, 15 to 30 minutes, not a code test. - Vendr, [Karat](https://www.vendr.com/marketplace/karat): third-party median spend, range and per-interview rates. Not Karat's figures. - HackerRank, [Pricing](https://www.hackerrank.com/pricing) (blocks scripted fetches; figures read in a real browser on 3 Nov 2026 and kept in the pipeline snapshot): plan prices and attempts, Chakra included. - HackerRank, [Chakra](https://www.chakra.sh/): what Chakra does. - HackerRank, [Interview](https://www.hackerrank.com/products/interview/): live pair programming in a real codebase. - CodeSignal, [Pricing](https://codesignal.com/pricing/): Build price and credits, credit rule. - CodeSignal, [AI Interviewer](https://codesignal.com/ai-interviewer/): length, report, recruiters decide. - CodeSignal, [Live tech interviews](https://codesignal.com/live-tech-interviews/): AI-native cloud IDE with a human interviewer. - CoderPad, [Pricing](https://coderpad.io/pricing/): plan prices and allowances, extra-interview price, pads, agentic interviews on higher plans, summaries. - HireVue, [AI Interviewer](https://www.hirevue.com/platform/ai-interviewer): two-way conversation, probing follow-ups, explanation of responses. - HireVue, [Coding assessments](https://www.hirevue.com/products/coding-assessments): 200+ OnDemand coding assessments. - HireVue, [Video interviewing](https://www.hirevue.com/platform/video-interviewing): one-way on-demand video and live interviews. - Tracy AI Interviewer, [home page](https://tracyinterview.ai/): integrity signals list, use cases, 'you read the words, then you decide'. - Tracy AI Interviewer, [Product page](https://tracyinterview.ai/product): two-way video interview, no install for the person being interviewed, CV check, follow-ups that depend on the previous answer, recording with a timed transcript, three ways in, API and MCP. - Tracy AI Interviewer, [Hiring at volume](https://tracyinterview.ai/use-cases/volume-hiring): everyone who applies, on click, at any hour, no recruiter and no calendar, the same depth for applicant one and applicant five hundred. - Tracy AI Interviewer, [Technical interviews](https://tracyinterview.ai/use-cases/technical-interviews): a coding task, then a probe on the trade-offs. - Tracy AI Interviewer, [Expert vetting](https://tracyinterview.ai/use-cases/expert-vetting): check domain depth before an expert joins a project. - Tracy AI Interviewer, [Connect page](https://tracyinterview.ai/connect): MCP server and API connection. - Tracy AI Interviewer, [Pricing](https://tracyinterview.ai/pricing): credit price, billing unit, free results, credits never expire, no-answer sessions never charged, Enterprise API and MCP connections and customised setup. # Tracy vs CodeSignal: A Test or a Technical Interview? (Nov 2026) > Tracy vs CodeSignal in Nov 2026: validated assessments and an IDE against a video interview with follow-ups. Where each wins, what each costs, and who should pick which. - URL: https://tracyinterview.ai/compare/tracy-vs-codesignal - Published: 2026-11-06 - Updated: 2026-11-11 - Author: Tracy team Checked on 3 Nov 2026 against both vendors' own pages ([see Sources](#sources)). > **In short** > - Tracy gives every applicant a two-way video interview, with CV-based questions, follow-ups that depend on each answer and a written assessment per question with its reasoning, plus the recording and transcript behind it. > - Pick **CodeSignal** when you want validated assessments, a full IDE, proctoring and live technical interviews from one vendor. > - The two overlap on one thing, an AI screen. CodeSignal sells an AI Interviewer and on-camera AI Video Avatars beside its tests; Tracy's interview is the whole product. ## Verdict: a test platform or an interview CodeSignal and Tracy answer different first questions. CodeSignal asks whether a candidate can pass a validated test, and it has built a lot around that: certified assessments, a full IDE, a Suspicion Score, live interviews and, in the same plans, an AI Interviewer. Tracy asks what a person knows when you talk to them and follow up. It runs a video interview for each applicant when they click the link, with no scheduling, and returns a recording, a transcript timed to the recording and a written assessment per question with its reasoning. If your engineers want to read passing code, pick CodeSignal. If you want a conversation with hundreds of applicants where the second question matters, and you do not want a plan, pick Tracy. Many teams will use both. The rest of this page says where each is better and on which dimension. **On this page** - [Who is each for?](#who-is-each-for) - [How do they compare side by side?](#how-do-they-compare-side-by-side) - [Where does CodeSignal win?](#where-does-codesignal-win) - [Where does Tracy win?](#where-does-tracy-win) - [How does pricing compare?](#how-does-pricing-compare) - [Using both, or switching](#using-both-or-switching) - [Questions](#questions) - [Sources](#sources) ## Who is each for? **CodeSignal** is for teams that hire engineers and other roles and want measured, validated screening: assessments built by subject matter experts, a coding environment where candidates build, run and test, and live interviews in a cloud IDE. Its AI Interviewer adds role agents for sales, marketing, finance and more, as a first pass for volume roles or a behavioral screen before technical rounds. **Tracy** is for teams that want a first conversation with everyone who applies, for roles where a test does not tell you enough and you want to hear the reasoning. You describe who you want to interview, share a link, and the person joins at any hour with no install. With CV check on, Tracy tailors 3 to 5 questions to the CV and the role's must-haves. If the format is new to you, start with [what an AI interviewer is](/blog/what-is-an-ai-interviewer). ## How do they compare side by side? | Dimension | CodeSignal | Tracy AI Interviewer | |---|---|---| | Starting point | Validated assessments, IDE, live interviews; AI Interviewer in the same plans | A video interview from a brief | | AI interview format | AI Interviewer, whose demo runs by phone or browser; AI Video Avatars run on-camera video screens | Two-way video interview, no install for the person | | Volume and scheduling | AI Phone Screens: candidates complete the screen without waiting for a live call; AI Interviewer: more applicants are considered with the same team even when volume spikes | Each applicant starts when they click the link, at any hour, with no scheduling | | CV or resume | Not stated for the AI Interviewer | CV check: 3 to 5 questions tailored to the CV and the role's must-haves | | Follow-up questions | Yes, based on your criteria | Yes: each follow-up depends on the answer before it | | Coding in the interview | Not stated for the AI Interviewer; coding sits in assessments and live interviews with a full IDE | Can introduce a coding task and probe the trade-offs | | Your own questions | Yes: custom competencies, tone and follow-ups | Yes: Tracy drafts the questions from your brief, which you can edit, or you bring your own | | Report | Overall score, skill evaluation, key insights, full transcript with rationale | Recording, transcript timed to the recording and a written assessment per question with its reasoning | | Connections | ATS integrations by plan (Gem and Ashby, Workday, iCIMS, Oracle) | API and MCP connections; start from an invitation link, an apply page or your own system; customised setups for enterprise | | Integrity checks | Suspicion Score, LeakSweep, AI proctoring, ID verification | For video interviews, signals listed on Tracy's home page: tab changes and fullscreen, face and gaze, a long delay before an answer, external screens | | Non-engineering roles | Yes: sales, marketing, finance and more | Roles you describe in a brief | | Published price | Build $79 and Grow $479 a month billed annually; Pro custom | First 3 results free, then $1 per credit | | Human decision | Humans review and decide | You read the words, then you decide | "Not stated" means the pages we read do not say. It does not mean the vendor lacks the feature. ## Where does CodeSignal win? CodeSignal does these things, and Tracy does not claim them. - **Validated assessments.** CodeSignal's certified assessments are written by subject matter experts and validated by I/O psychologists, each backed by 2,800 hours of research and testing. Tracy has no assessment library. - **A full IDE.** Candidates build, run and test applications. CodeSignal also lets you allow full AI assistance or limit it to syntax and documentation help. Tracy can set a coding task and probe it in conversation, but it is not an IDE. - **Live technical interviews.** A human-run session in a cloud IDE with video, audio, chat and whiteboarding, and 4,000+ questions and templates across 70+ coding languages and frameworks. - **Integrity tooling.** Suspicion Score looks at solution similarity, telemetry and copy-paste activity. LeakSweep watches for leaked questions, and the platform adds AI proctoring and ID verification. For assessments where cheating is the main risk, this is deep. Our guide to [AI interview cheating detection](/blog/ai-interview-cheating-detection) covers how such checks work in general. - **Calibration to your reviewers.** CodeSignal says its I/O psychologists compare the AI's ratings with your reviewers' and tune until they are on par, and a pilot can include an adverse impact study. - **Range of roles and ATS.** Role agents for sales, customer success, marketing and more; Gem and Ashby in its Core ATS list, and Workday, iCIMS, Oracle and more in its Enterprise ATS list. - **Stated security.** Every plan states SOC 2 Type II, GDPR and CCPA compliance and data encryption. We have not compared security here, so ask both vendors for current documentation. A team hiring engineers against a rubric, with proctoring and a measured pass mark, should start with CodeSignal. Our [AI technical interviews](/blog/ai-technical-interviews) guide covers what a conversation adds on top of a test. ## Where does Tracy win? Each point is measured against CodeSignal's own wording. - **Coding task with trade-off probing.** CodeSignal's AI Interviewer page positions the interview as a first pass for volume roles, a customer-facing screen, Support and Ops scenarios, or an engineering behavioral screen before technical rounds. It does not describe a coding task inside the AI interview; coding sits in the assessments and live interviews. Tracy can introduce a coding task in the interview and probe the trade-offs behind the answer. This is Tracy's approach, not a finding that CodeSignal's AI interview cannot do it. - **Follow-ups that start from the answer.** CodeSignal says its AI Interviewer asks follow-up questions based on your criteria. Tracy's follow-ups depend on the answer before them, and it waits for the end of the thought, with longer pauses allowed on hard questions. CodeSignal's follow-ups come from the criteria you set. - **A first round for every applicant.** On volume, the two are close. CodeSignal's AI Phone Screens page says candidates complete the screen without waiting for a live call, and its AI Interviewer page says more applicants are considered with the same team even when volume spikes, with no further phone screen to schedule. Tracy also removes the scheduled call: each applicant starts when they click the link, at any hour, with no recruiter and no calendar. The differences are in what the round contains. Tracy's CV check adds 3 to 5 questions from each CV into the conversation, each follow-up depends on the answer before it, and the result is a recording, a transcript timed to the recording and a written assessment per question with its reasoning. Interviews can also start three ways: an invite link, an apply page, or your own system handing the person the link. Dropped interviews can be rejoined within 10 minutes and get a reminder after a day. - **The CV in the conversation.** With CV check on, Tracy asks 3 to 5 questions from each CV and the role's must-haves, on top of the role's questions. The pages we checked describe no CV step in CodeSignal's AI Interviewer. - **Tools you already use.** Tracy's Enterprise plan includes API and MCP connections and a setup customised to the workflow ([pricing](https://tracyinterview.ai/pricing)), and an interview can start from an invite link, an apply page or your own system. CodeSignal's ATS integrations, listed above, are broader on its Enterprise plan than anything we can state for Tracy. - **Interview signals, stated as signals.** CodeSignal has the deeper integrity tooling for assessments: Suspicion Score, LeakSweep, AI proctoring and ID verification, covered above. For video interviews, Tracy's home page lists tab changes and fullscreen, face and gaze, a long delay before an answer and external screens as integrity signals. We state no rates. - **Evidence behind each question.** Tracy's result is a written assessment per question with its reasoning and the timestamps behind it, over the recording and a transcript timed to it, and you read the words, then you decide. CodeSignal also gives a transcript with rationale, so this is a difference of shape: CodeSignal's report leads with an overall score and skill evaluation. Whether each CodeSignal score links to the specific answer is not stated. Read [AI interview reports](/blog/ai-interview-reports) for what a reviewer should look for. ## How does pricing compare? Tracy charges $1 per credit with the first 3 results free and no seat fee or plan fee, and a session with no answers is never charged. CodeSignal's credits are public for Build and Grow, but a credit count for an AI interview is not stated. CodeSignal's published plans and Tracy's per-result price do not measure the same thing, so the table models both for a year under stated assumptions. CodeSignal uses annual billing, one credit per AI interview (its pages do not state the number) and $20 for each credit past the plan's allowance. Tracy uses results of up to 20 minutes and the first 3 free. The CodeSignal column shows the cheaper of Build and Grow in each row. | Interviews a year | CodeSignal | Tracy | |---|---:|---:| | 100 | $1,748 (Build) | $97 | | 500 | $7,348 (Grow) | $497 | | 1,000 | $17,348 (Grow) | $997 | Longer Tracy results cost more: 21 to 40 minutes is two credits. CodeSignal's credits also cover assessments and live interviews, which Tracy does not sell, so a team buying all three formats is not comparing like with like. The full working is in [CodeSignal pricing](/compare/codesignal-pricing). ## Using both, or switching If CodeSignal's assessments work for your engineering roles, keep them. Add Tracy for roles where a test misses the point: experts, hard-to-test skills, applicants you want to hear before you test them. If you are replacing CodeSignal's AI Interviewer only, run the same role through both with your own questions and compare the transcripts and reports. For a wider shortlist, see [CodeSignal alternatives](/compare/codesignal-alternatives) and [software engineer interview questions](/blog/software-engineer-interview-questions). ## Common questions ### What is the main difference between Tracy and CodeSignal? CodeSignal sells validated assessments, a coding IDE and live technical interviews, with an AI Interviewer and on-camera AI Video Avatars. Tracy AI Interviewer is a video interview, started when each applicant clicks the link, that asks questions from the CV, follows up on each answer and returns a written assessment per question with a recording and a transcript behind it. ### Does CodeSignal run AI interviews on video? Its AI Phone Screens page lists AI Video Avatars, on-camera video screens with an AI avatar, and that link leads to the AI Interviewer page, whose demo runs by phone or browser. CodeSignal's separate Live Tech Interviews, run by a human, include built-in video. Tracy runs a two-way video interview. ### Can Tracy run a coding interview? Tracy can introduce a coding task and probe the trade-offs behind an answer. It is not a test platform with an IDE, validated assessments or proctoring, which CodeSignal offers. ### Which is cheaper, Tracy or CodeSignal? Tracy charges $1 per credit with the first 3 results free and no plan fee. CodeSignal Hire starts at $79 a month billed annually, with $20 extra credits, and does not state credits per AI interview. On a one-credit-per-interview assumption, 500 interviews in a year cost about $7,348 on CodeSignal Grow and $497 on Tracy. ### Can I use both? Yes. A team may want CodeSignal's validated tests for engineering roles and an interview for roles where a test does not fit. Each is priced and run on its own. --- **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=tracy-vs-codesignal&utm_content=demo)** to run one of your roles through Tracy, or [try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=tracy-vs-codesignal&utm_content=try-free) with your first three results. ## Sources All sources retrieved 3 Nov 2026. - CodeSignal, [Pricing](https://codesignal.com/pricing/): plans, credits, $20 extra credit, ATS integrations, AI interviewers by plan, security wording. - CodeSignal, [AI Interviewer](https://codesignal.com/ai-interviewer/): demo by phone or browser, follow-ups based on your criteria, report contents, calibration, role agents, custom interviewers, human decision. - CodeSignal, [AI Phone Screens](https://codesignal.com/ai-phone-screens/): AI Video Avatars, on-camera video screens with an AI avatar. - CodeSignal, [Technical Assessments](https://codesignal.com/technical-assessments/): validated assessments, IDE, AI assistance settings. - CodeSignal, [Live Tech Interviews](https://codesignal.com/live-tech-interviews/): cloud IDE with video, audio and chat; 4,000+ questions; 70+ languages and frameworks. - CodeSignal, [Cheating and fraud](https://codesignal.com/cheating-and-fraud/): Suspicion Score, LeakSweep, proctoring and ID verification. - Tracy AI Interviewer, [home](https://tracyinterview.ai/): video interviews, follow-ups that depend on the answer before, coding task with a probe on the trade-offs, your own questions, integrity signals list, 'you read the words, then you decide'. - Tracy AI Interviewer, [product](https://tracyinterview.ai/product): two-way video interview, no install, CV check, three ways in, rejoin and reminder, transcript timed to the recording. - Tracy AI Interviewer, [volume hiring](https://tracyinterview.ai/use-cases/volume-hiring): starts on click at any hour with no recruiter and no calendar, rejoin within 10 minutes, reminder after a day. - Tracy AI Interviewer, [pricing](https://tracyinterview.ai/pricing): $1 per credit, 3 free results, billing by started 20 minutes, no-answer sessions not charged, Enterprise with API and MCP connections and a customised setup. # Tracy vs Karat: AI or Human Interviewers for Technical Rounds (Nov 2026) > Karat books expert human interviewers for engineering roles. Tracy AI Interviewer interviews every applicant on video. Who each is for, and what each costs. - URL: https://tracyinterview.ai/compare/tracy-vs-karat - Published: 2026-11-06 - Updated: 2026-11-11 - Author: Tracy team > **In short** > - Pick **Tracy AI Interviewer** over Karat for a first round when no applicant should wait for a booking: every one gets a video interview with CV-tailored questions, and your team reads the recording and integrity signals with evidence before deciding. > - Pick **Karat** when expert human interviewers, a standard rubric and benchmark data on engineering roles matter more than reach. > - Many teams will use both: Tracy for the first round, Karat for finalists. These are two different answers to one problem. Karat supplies the interviewer: a trained engineer who runs the technical interview for you. Tracy AI Interviewer is the interviewer: it runs the conversation on video and writes up what each person said. Facts come from both vendors' own pages, read on 3 Nov 2026, except where we mark a number as third-party. The Karat sections are written to be useful if you pick Karat. ## Who is each for? **Karat** is for engineering teams that want expert interviewers to take the first or second technical interview off their own engineers, scored on one rubric across teams and geographies. **Tracy** is for teams that want to interview everyone who applies, including candidates outside engineering, with questions built from the role and the CV, and review a recording, transcript and written assessment instead of booking a person for each one. ## How do they compare? | Dimension | Tracy AI Interviewer | Karat | | --- | --- | --- | | Who interviews | AI, in a two-way video interview | Karat Interview Engineers, who are human | | Format | Two-way video interview; no install for the person being interviewed | Live video with an engineer; camera strongly encouraged, not required | | Scheduling and volume | None: each applicant joins from a link when they click, at any hour, so everyone who applies can be interviewed | 24/7 on demand, same-day, evenings and weekends; each interview is an engineer's booked session | | CV or resume | Can read the CV and ask 3 to 5 questions tailored to it and to the role's must-haves | Its FAQ says customers may share a candidate's resume data; how interviewers use it is not stated | | Coding | Can introduce a coding task and probe the trade-offs behind an answer | Code playback; engineer-led coding interviews | | What you get back | A recording, a transcript timed to the recording and a written assessment per question with its reasoning; you read the words, then you decide | Overall recommendation, a quantitative score and written feedback per question | | Standard across interviews | The same depth for applicant one and applicant five hundred, per Tracy's hiring-at-volume page | Consistent scoring rubrics | | Benchmark data | Not stated on Tracy's pages | Compare against 600,000+ interviews (Karat's claim) | | Integrity | On video interviews, integrity signals with their evidence for a reviewer, not a verdict (Tracy's home page) | Interview Engineers watch for cheating; a second engineer rewatches flagged interviews | | Roles covered | Hiring, expert interviews, accelerator and fellowship applications, customer research | Engineering talent | | Connections | API and MCP connections; start from an invitation link, an apply page or your own system; customised setups for enterprise | Not stated on the pages we checked | | Price | $1 per credit; first 3 results free | No published price; Request a Consult or Request demo | | Human decision | Stays with you | AI never makes final hiring decisions, per Karat | ## Where does Karat win? - Expert human judgement. Karat's interviews are "live, human-led interviews conducted by experienced, certified interview engineers." An engineer can read code and reasoning the way your own engineers would. Tracy is software that asks follow-ups; its pages do not describe an expert's review of a candidate's code. - One standard across teams. Karat describes structured formats and consistent scoring rubrics, and says every engineer is held to a consistently enforced technical bar. If your problem is that different interviewers grade differently, a single outside bench of interviewers addresses that directly. - Benchmark data. Karat's home page invites you to see how your talent measures up against its 600,000+ interviews, and a customer quote on the same page credits it with competitive benchmarking across roles and geographies. Tracy's pages do not describe benchmarks across other companies' candidates. - Scale of its own practice. Karat's home page says 600,000+ technical interviews; its product page says 350k+ in 70+ countries with 500+ interview engineers. These are Karat's numbers and we did not verify either. - An earlier step. Karat Qualify is an on-demand, adaptive multiple-choice screen you can send to all applicants before the Karat Interview. It takes 15 to 30 minutes for most candidates and is not a code test, so Karat has an answer to volume too, though it is not an AI interview. - A human check on cheating. Karat's engineers watch for behaviours such as typing outside the browser window and looking between monitors, and a flagged interview is rewatched in full by a second trained engineer. - Security paperwork. Karat says it provides a SOC 2 Type II attestation report on request. - AI in the loop. Its NextGen page describes human-led, AI-enabled interviews, and its home page lists live AI-native simulations led by its Interview Engineers, so a human can run a modern, AI-allowed interview. ## Where does Tracy differ? Each point names its dimension and the Karat fact it is set against. - Volume. Tracy has no booking step: every applicant is interviewed when they click the link, at any hour, with no recruiter present, so hundreds of applicants can each get a full interview. Karat's on-demand scheduling is fast, but each interview is one engineer's session. The difference is in the model: with an AI interviewer, you do not need a person for each applicant. - Depth from the CV. Tracy can read the CV and ask 3 to 5 questions tailored to it and to the role's must-haves, on top of the role's questions. Its product page says each follow-up depends on the answer before it and a vague answer gets a sharper question, and Tracy can introduce a coding task and probe the trade-offs. It waits for the end of an answer before it follows up. Karat's pages do not state how its interviewers use a CV or follow up, and we make no claim about it. - Evidence. Each Tracy interview leaves the recording of both sides, a transcript timed to the recording and a written assessment per question with its reasoning. You read the words, then you decide. Karat's report also carries written feedback per question and a quantitative score, so the evidence has a different shape on each side. - Integrity signals. Tracy's video interviews include integrity signals with their evidence for a reviewer: tab changes and fullscreen, face and gaze (live face detection and where the person is looking), a long delay before an answer and an external screen, listed on [Tracy's home page](https://tracyinterview.ai/). They are signals, not a verdict, and we give no detection rate. Karat's checks are made by people, so the two take different routes to the same concern. - Stack. Tracy connects through an API and an MCP server, and an interview can start from an invitation link, an apply page or your own system. Enterprise teams can ask for a customised setup. - Fields beyond engineering. Tracy's site lists hiring at volume, expert vetting that checks domain depth before an expert joins a project, applications to accelerators and fellowships, and customer research. Karat's site describes engineering talent. ## What does each cost? Karat does not publish a price. The only figures we found are third-party, from Vendr's procurement data, and they are not Karat's: a median buyer spend of $175,695 a year, a range of $57,000 to $1,069,900 a year, and per-interview rates of $200 to $450 for a standard technical screen depending on volume, term and tier. Tracy's price is its own and is on its pricing page: $1 per credit, one credit for each started 20 minutes, so up to 20 minutes is $1, 21 to 40 minutes is $2 and 41 to 60 minutes is $3. Your first 3 results are free. These are not like-for-like. A Karat interview is an engineer's time with a rubric, and a Tracy interview is an AI conversation with a written result. Compare them by round: what a first round for everyone costs, and what a second round for finalists costs. See [AI interview cost](/blog/ai-interview-cost) for more published prices. ## How would you get started? 1. Run a first round on one open role. Invite applicants to a Tracy video interview; your first 3 results are free. 2. Read the results next to your usual first screen. Check whether the people Tracy surfaced are the ones you would have chosen. [AI interview reports](/blog/ai-interview-reports) shows what a reviewer should look for. 3. Keep expert humans where they add the most. If Karat or your own engineers interview finalists, the AI first round can narrow who reaches them. [AI screening interviews](/blog/ai-screening-interviews) explains how a first round runs, [AI interview vs human interview](/blog/ai-interview-vs-human-interview) covers what each does well, [AI technical interviews](/blog/ai-technical-interviews) shows a coding screen, and [Karat alternatives](/compare/karat-alternatives) lists other options. For engineering questions to pair with either, see [software engineer interview questions](/blog/software-engineer-interview-questions). ## Common questions ### Is Tracy a replacement for Karat? For a first conversation with every applicant, it can be. For expert human judgement on engineering candidates, with Karat's rubric and benchmark data, Karat is built for that and Tracy is not. Tracy's case is volume, depth and evidence: every applicant interviewed on click, questions built from the role and the CV, and a recording, transcript and written assessment per question. Many teams will want an AI first round and expert humans for finalists. ### How do Tracy and Karat pricing compare? Tracy publishes $1 per credit, with one credit per started 20 minutes, so a 60-minute session is $3, and the first 3 results are free. Karat publishes no price. Vendr, a third party, reports per-interview rates of $200 to $450 for a standard technical screen and a $175,695 median annual spend. The two are not like for like: Karat's interviewer is a human engineer. ### Does Tracy handle coding interviews? Tracy can introduce a coding task and probe the trade-offs behind an answer. Karat's interviews include code playback and are run by expert engineers. Tracy's pages do not describe an expert review of the code. ### Can Tracy interview roles other than engineers? Yes. Tracy's site lists hiring first rounds, expert interviews, accelerator and fellowship applications and customer research. Karat's site describes evaluating engineering talent. ### Does Tracy or Karat make the hiring decision? Neither. Karat says AI never makes final hiring decisions, and Tracy says the decision stays with you. --- **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=tracy-vs-karat&utm_content=demo)** to see Tracy interview a candidate for one of your roles, or [try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=tracy-vs-karat&utm_content=try-free) with your first 3 results on us. ## Sources All sources retrieved 3 Nov 2026. - Karat, [home page](https://karat.com/): 600,000+ interviews (vendor figure), the compare-globally wording, a customer quote on benchmarking, standards wording, live AI-native simulations, engineering scope. - Karat, [Karat Interview](https://karat.com/karat-interview/): human-led interviews, 24/7 on-demand, rubrics, 350k+ interviews and 500+ engineers (vendor figures), no pricing page. - Karat, [Customer FAQ](https://karat.com/customer-faq): report contents, live video with the camera encouraged but optional, code playback, resume data may be shared by customers, cheating checks and second review, AI never decides, SOC 2 Type II attestation report on request. - Karat, [NextGen](https://karat.com/evaluate-ai-ready-engineers/): human-led, AI-enabled interviews. - Karat, [Qualify](https://karat.com/karat-qualify/): on-demand adaptive multiple-choice screen, 15 to 30 minutes, not a code test. - Vendr, [Karat](https://www.vendr.com/marketplace/karat): third-party median spend, range and per-interview rates. Not Karat's figures. - Tracy AI Interviewer, [home page](https://tracyinterview.ai/): integrity signals list, use cases, 'you read the words, then you decide'. - Tracy AI Interviewer, [Product page](https://tracyinterview.ai/product): two-way video interview, no install for the person being interviewed, CV check, follow-ups that depend on the previous answer, recording with a timed transcript, three ways in, API and MCP. - Tracy AI Interviewer, [Hiring at volume](https://tracyinterview.ai/use-cases/volume-hiring): everyone who applies, on click, at any hour, no recruiter and no calendar, the same depth for applicant one and applicant five hundred. - Tracy AI Interviewer, [Technical interviews](https://tracyinterview.ai/use-cases/technical-interviews): a coding task, then a probe on the trade-offs. - Tracy AI Interviewer, [Expert vetting](https://tracyinterview.ai/use-cases/expert-vetting): check domain depth before an expert joins a project. - Tracy AI Interviewer, [Connect page](https://tracyinterview.ai/connect): MCP server and API connection. - Tracy AI Interviewer, [Pricing](https://tracyinterview.ai/pricing): credit price, billing unit, free results, no-answer sessions never charged, Enterprise API and MCP connections and customised setup. # AI Interview Tools for High-Volume Skilled Hiring (Nov 2026) > Six AI interview tools for teams hiring skilled roles at volume, compared on depth, capacity, evidence and stack fit, with where each one wins. Facts checked 3 Nov 2026. - URL: https://tracyinterview.ai/blog/ai-interview-tools-for-high-volume-hiring - Published: 2026-11-05 - Author: Tracy team > **In short** > - For skilled roles at volume, test **Tracy AI Interviewer** first: everyone who applies gets a video interview at any hour, with a CV check, a coding task it can probe on the trade-offs, and a written assessment per question with its reasoning. > - Tracy sets one interview language at setup and connects through API and MCP. It reaches Workable and Breezy HR through those vendors' own connectors, and other systems through Zapier; results are read in Tracy. HireVue, HeyMilo, Ribbon, Greenhouse Voice AI and Alex are stronger on ATS integration, language coverage or scheduling. > - The ranking rates what each vendor's own pages state across four pillars. It is not a test, and we make one of the six tools. Checked on 3 and 4 Nov 2026 against each vendor's own pages. A mark of "not stated" means we read the pages and found no mention, not that the product lacks the feature. ## What does a team hiring skilled roles at volume need? Hourly hiring mostly needs fast eligibility checks and scheduling. Skilled hiring at volume adds a harder job: telling a strong candidate from a polished one, for every applicant, without a recruiter on every call. In a Robert Half survey fielded in November 2025, 67% of US HR leaders said reviewing AI-generated applications had slowed their hiring. That is HR leaders' perception, not a measured delay. Some teams respond with a first conversation that does not depend on the written application. See [how to screen candidates when there are too many applicants](/blog/how-to-screen-candidates) for the wider problem. We judge the tools on four pillars: - **Depth:** asks follow-ups on the person's own answer, and can handle a technical task. - **Capacity:** states how many interviews it has run or is built for volume, and says what volume costs. - **Evidence:** the result points back to what the candidate said. - **Stack:** connects to the ATS you already run. ## How do the six tools rank on the four pillars? For each pillar a mark means "stated on the vendor's own pages", "partly" or "not stated". Capacity is "stated" only when a vendor gives both a volume figure and something about what volume costs; a vendor with a volume figure but no published price is "partly". We scored a stated mark 2, a partly mark 1 and a not stated mark 0, and ordered the tools by total. Ties go first to Depth and then to Evidence, the two pillars that separate skilled hiring from hourly. Tools that match on every mark share a rank and are listed alphabetically, which implies no order between them. | Tool (rank) | Depth | Capacity | Evidence | Stack | Published price | |---|---|---|---|---|---| | Tracy AI Interviewer (1) | Stated | Stated | Stated | Partly | $7 per credit (1 credit per started 20 minutes), first 3 results free | | Ribbon (2) | Partly | Stated | Stated | Stated | $499 a month for 100 interviews (blog, annual) | | Greenhouse Voice AI (3=) | Partly | Partly | Stated | Stated | Quote on request | | HeyMilo (3=) | Partly | Partly | Stated | Stated | None for the interview platform | | HireVue (3=) | Partly | Partly | Stated | Stated | None | | Alex (6) | Partly | Stated | Partly | Stated | From $36K a year for staffing firms (blog) | Totals: Tracy 7, Ribbon 7 (Tracy first on Depth), Greenhouse, HeyMilo, HireVue and Alex 6 (Alex last on Evidence). Greenhouse, HeyMilo and HireVue match on every mark. "Partly" on Depth means follow-up questions are stated but a technical task is not stated, or sits in a separate product. "Partly" on Capacity means a volume figure or a pricing basis is given but no price. Capacity figures are vendor claims. Tracy's marks rest on its own pages: Depth (follow-ups, a coding task it can introduce, CV-based questions), Capacity (100K+ interviews on the home page, the same depth for applicant one and applicant five hundred on the hiring-at-volume page, and a published price), Evidence (recording, transcript, an assessment per question with its reasoning) and Stack (API and MCP connections on the pricing page, but no described ATS write-back). ## Tracy AI Interviewer Tracy interviews everyone who applies, not a shortlist. Each applicant gets a two-way video interview in the browser when they click the link, at any hour, with no calendar and no recruiter on the call, and there is no install for the person interviewed. Tracy's home page reports 100K+ interviews completed, and hundreds of applicants can be interviewed over a hiring round. Depth comes from three things. With CV check on, Tracy can read the CV and ask 3 to 5 questions tailored to it and to the role's must-haves. It follows up on what the person actually said, waiting for the end of the answer, and it can introduce a coding task and probe the trade-offs behind an answer. Evidence comes with every result: a recording of both sides, a transcript timed to the recording, a written assessment per question with its reasoning, and integrity signals with their evidence for a reviewer, so you read the words and then decide. Enterprise teams can also ask for a customised setup. On cost, the first 3 results are free, then $7 per credit, with 1 credit per started 20 minutes, no seats, and no charge for a session with no answers. At that rate, 500 interviews of up to 20 minutes cost $3,479 after the 3 free results. Where others are stronger: Tracy sets one interview language per interview and has no described ATS write-back, which is why Stack is marked partly. It connects through API and MCP, and its MCP server works with Claude, ChatGPT, Mistral, Gemini and Copilot Studio. For Workable and Breezy HR it runs beside each vendor's own connector in Claude or ChatGPT; the ATS shares each candidate's name and email, and Tracy's guides say results stay in Tracy. Zapier connects it to other systems: it can invite people and react to interview events or results. Tracy's home page lists the integrity signals for video interviews: tab changes and fullscreen, where the person is looking, face in frame, a long delay before an answer and an external screen. They are signals, not a verdict, and we state no rates. Pick Tracy when depth in the candidate's field, an interview for every applicant and evidence a reviewer can check matter more than ATS write-back. Pick another tool if your team needs scores to land on the candidate record without an extra step. See [Tracy vs HireVue](/compare/tracy-vs-hirevue). ## Ribbon Ribbon runs structured voice and video interviews that it says adapt in real time, in 10+ languages, and it evaluates candidates from the interview transcripts. It lists 60+ integrations, offers a 7-day free trial and adds sourcing and outreach by SMS, WhatsApp and email. Dayforce reports more than one million completed interviews. Its own blog, dated 23 August 2026, advertises packages from $499 a month for 100 interviews on annual billing, though its pricing page returned a 404 when we checked. Dayforce has signed an agreement to acquire Ribbon, and the deal had not closed in its 23 October 2026 release. Pick Ribbon when you want sourcing, outreach and a first interview in one tool, quickly. Technical tasks are not stated. See [Ribbon alternatives](/compare/ribbon-alternatives). ## Greenhouse Voice AI Greenhouse describes a structured two-way voice interview scored against the team's own rubric, with dynamic follow-ups, a recording, a transcript and per-question scores with explanations. It works with or without the Greenhouse ATS and integrates with 60+ ATS providers, with depth that varies by ATS, in 11 languages. Voice only: the pages we read describe no video, and no coding task. It is sold as an annual subscription scaled to headcount with a fair-use interview allowance, quoted on request, and the pages state no interview count or amount, which is why Capacity is marked partly. Pick Greenhouse Voice AI when you already run Greenhouse or want a voice screen tied to your rubric. See [Greenhouse's AI interviewer vs Tracy](/compare/greenhouse-ai-interviewer-vs-tracy) and [the ATS interviewer vs a dedicated one](/blog/ats-ai-interviewer-vs-dedicated). ## HeyMilo HeyMilo describes itself as an AI recruiter for high-volume hiring that screens and interviews every applicant over voice, video, SMS and forms in 20+ languages. It asks follow-ups in real time; the recruiter sets how many, and they are tied to the scoring rubric. Reports carry per-question scores with a rationale grounded in the candidate's words. It syncs with 20+ ATS partners and writes scores, reports, transcripts and recordings back to the candidate record, and it books calendar slots after the screen. It also reports more than one million candidates screened, a company statement. Its pages list a separate skill assessment with scenario-based evaluation, which is why Depth is partly. It publishes no price for the interview platform, so Capacity is marked partly; the prices it does publish, a free tier and $59 a month, belong to HeyMilo Red, its separate interview proctoring product. Pick HeyMilo for high-volume screening where SMS, forms, language range and ATS write-back matter. See [Tracy vs HeyMilo](/compare/tracy-vs-heymilo). ## HireVue HireVue's AI interviewer asks follow-up questions in a two-way conversation, and says every score ties back to evidence from the interview and your competency rubric. Recruiters view scores in the ATS. HireVue sells more than 200 coding assessments as a separate product and does not say whether the interviewer probes code. Its video interviewing page states 390K interviews completed in 4 months, a vendor claim, and its AI interviewer page lists adverse impact testing before models go live and the line "AI Informs. Humans decide." HireVue publishes no price: its pricing page points to a plan conversation, so Capacity is marked partly. Pick HireVue when governance, an audit trail and one suite at enterprise volume matter most. Compare it in [HireVue alternatives](/compare/hirevue-alternatives). ## Alex Alex interviews over video, phone, SMS and WhatsApp and asks follow-ups based on what the candidate says. It reports 1,000,000+ candidates interviewed, a company claim, and integrates with 33+ ATS platforms. A coordinator books qualified candidates on team calendars. Its pages describe searchable transcripts, role-fit scores and fraud flags; per-score quote tracing is not stated, so Evidence is marked partly. Its current product pages describe behavioral and role-specific interviews rather than coding tasks. Its own blog says pricing starts at $36K a year (stated for staffing firms in two posts), with custom enterprise pricing based on volume; that is the only figure we found. Pick Alex when you want one workflow from outreach to scheduling. See [Tracy vs Alex](/compare/tracy-vs-alex). ## How should you choose? - Skilled roles where every applicant should get a deep first conversation, with CV-based questions and evidence to review: start with Tracy. - Scores must land in your ATS record: HireVue, HeyMilo, Greenhouse Voice AI or Alex, then check how deep the integration goes for your ATS. - Many languages or SMS and WhatsApp channels: HeyMilo or Alex. - Sourcing and a first interview together: Ribbon. - Hourly or frontline volume where scheduling is the bottleneck: this list is not built around it. HeyMilo and Ribbon describe high-volume and frontline use; see [AI interview tools for staffing agencies](/blog/ai-interview-tools-for-staffing-agencies) and [best AI interview software](/blog/best-ai-interview-software). Run two real roles through two tools, include a vague answer, and compare the next question each one asks. For what each model costs per candidate, read [AI interview cost](/blog/ai-interview-cost). ## Common questions ### What is the best AI interview tool for high-volume hiring of skilled roles? It depends on what you need most. Tracy AI Interviewer is the pick when every applicant should get a deep first interview, with CV-based questions, follow-ups and evidence a reviewer can check. HireVue, HeyMilo, Ribbon, Greenhouse Voice AI and Alex are stronger on ATS integration, languages or scheduling. ### How is high-volume skilled hiring different from hourly hiring? Hourly hiring mostly needs fast eligibility checks and scheduling. Skilled hiring needs the first conversation to show how deep a person's knowledge goes, so follow-up questions and evidence in the result matter more than speed alone. ### How much does an AI interview cost at volume? Tracy publishes $7 per credit, 1 credit per started 20 minutes, with the first 3 results free. Ribbon's own blog advertises packages from $499 a month for 100 interviews on annual billing. Alex's blog says from $36K a year, stated for staffing firms in two posts. HireVue, Greenhouse and HeyMilo publish no list price for the interview product. ### Do these tools replace the hiring decision? No. HireVue, Greenhouse, HeyMilo and Alex each state that people make the hiring decision, and Tracy leaves the decision with the customer. ### Can I run a coding interview at volume? Tracy can introduce a coding task and probe the trade-offs behind an answer. HireVue sells separate coding assessments. The other pages we read do not describe a coding interview, which does not mean the product lacks one; ask each vendor. --- **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=ai-interview-tools-for-high-volume-hiring&utm_content=demo)** or [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=ai-interview-tools-for-high-volume-hiring&utm_content=try-free) ## Sources All sources retrieved 3 and 4 Nov 2026. - Tracy AI Interviewer, [Tracy home page](https://tracyinterview.ai/), video interviews, follow-ups, coding task, assessment per question, integrity signals, 100K+ interviews - Tracy AI Interviewer, [Tracy product page](https://tracyinterview.ai/product), CV check, recording and transcript timed to the recording, one interview language chosen at setup, no install for the person interviewed - Tracy AI Interviewer, [Hiring at volume](https://tracyinterview.ai/use-cases/volume-hiring), start on click at any hour, the same depth for applicant one and applicant five hundred - Tracy AI Interviewer, [Tracy pricing](https://tracyinterview.ai/pricing), price, free results, billing unit, no seats, API and MCP connections and a setup customised to your workflow for Enterprise - Tracy AI Interviewer, [Connect Tracy to your AI app](https://tracyinterview.ai/connect), MCP server for Claude, ChatGPT, Mistral, Gemini and Copilot Studio - Tracy AI Interviewer, [Workable guide](https://tracyinterview.ai/help/workable), what moves between Workable and Tracy - HireVue, [HireVue AI interviewer](https://www.hirevue.com/platform/ai-interviewer), follow-ups, evidence, ATS scores, governance, human decision - HireVue, [HireVue video interviewing](https://www.hirevue.com/platform/video-interviewing), interviews completed in 4 months - HireVue, [HireVue coding assessments](https://www.hirevue.com/products/coding-assessments), coding assessments as a separate product - HireVue, [HireVue pricing](https://www.hirevue.com/pricing), no published price - Greenhouse, [Greenhouse Voice AI interviewing](https://www.greenhouse.com/voice-ai-interviewing), voice interview, rubric, evidence, ATS count, languages, pricing status - HeyMilo, [HeyMilo home page](https://www.heymilo.ai/), positioning, languages, write-back to the candidate record - HeyMilo, [HeyMilo interviewing candidates](https://www.heymilo.ai/interviewing-candidates), real-time follow-ups and transcript-backed evidence - HeyMilo, [HeyMilo integrations](https://www.heymilo.ai/integrations), 20+ ATS partners - Ribbon, [Ribbon home page](https://www.ribbon.ai/), real-time conversations, 60+ integrations - Ribbon, [Ribbon AI interviewer](https://www.ribbon.ai/ai-interviewer), voice and video interviews, 10+ languages - Ribbon, [Ribbon vs BrightHire](https://www.ribbon.ai/blog/ribbon-vs-brighthire), $499 a month for 100 interviews, annual billing, 23 Aug 2026 - Dayforce, [Dayforce to acquire Ribbon AI](https://www.dayforce.com/who-we-are/newsroom/dayforce-to-acquire-ribbon-ai-inc-to-advance-agentic-recruiting), signed agreement not closed, interview count, transcript basis - Alex, [Alex AI interviewer](https://www.alex.com/product/ai-interviewer), video, phone, SMS and WhatsApp interviews, follow-ups, transcripts and scores - Alex, [Alex platform](https://www.alex.com/platform), 33+ ATS integrations - HeyMilo, [HeyMilo raises $6 million](https://blog.heymilo.ai/heymilo-raises-6-million), more than one million candidates screened, company statement of 16 Jun 2026 - HeyMilo, [HeyMilo interview scheduling](https://www.heymilo.ai/interview-scheduling), calendar booking after the screen - HeyMilo, [HeyMilo responsible AI](https://www.heymilo.ai/responsible-ai), advisory AI, candidate quotes with timestamps - Alex, [Alex coordinator](https://www.alex.com/product/coordinator), booking after the AI screen - Alex, [Alex llms.txt](https://www.alex.com/llms.txt), 1,000,000+ candidates interviewed, company claim - Tracy AI Interviewer, [Use Tracy with Breezy HR](https://tracyinterview.ai/help/breezy), connector in Claude or ChatGPT, results stay in Tracy - Tracy AI Interviewer, [Connect Tracy to Zapier](https://tracyinterview.ai/help/zapier), triggers and invite action - HeyMilo, [HeyMilo voice and video interviews docs](https://docs.heymilo.ai/deep-dive/interviews/voice-interviews), per-question maximum follow-ups set by the recruiter - HeyMilo, [HeyMilo vs Alex](https://www.heymilo.ai/compare/alex-ai), follow-ups tied to the rubric - HeyMilo, [HeyMilo Red pricing](https://red.heymilo.ai/pricing), prices of the separate proctoring product - Greenhouse, [Voice AI works with 60+ ATS](https://www.greenhouse.com/product-features/voice-ai-now-works-with-60-ats-platforms), integration depth varies by ATS - Alex, [Alex vs JobDiva](https://www.alex.com/blog/jobdiva-vs-alex), pricing from $36K a year for staffing firms - Robert Half, [Robert Half survey: 67% of HR leaders report AI-generated applications are slowing hiring](https://press.roberthalf.com/2026-03-10-Robert-Half-survey-67-of-HR-leaders-report-AI-generated-applications-are-slowing-hiring), 10 Mar 2026 # The AI Interviewer in Your ATS vs a Dedicated One: What You Give Up > What Greenhouse, Workday and Dayforce actually offer for AI interviews today, what a built-in interviewer leaves out, and when a dedicated one earns its place. - URL: https://tracyinterview.ai/blog/ats-ai-interviewer-vs-dedicated - Published: 2026-11-05 - Author: Tracy team > **In short** > - A dedicated interviewer such as **Tracy AI Interviewer** adds a video interview at any hour with no scheduling, a vague answer met with a sharper question, and a recording and transcript the reviewer can check. API and MCP connections let it sit beside any ATS. > - Stay with the interviewer inside your ATS when you want one vendor and results on the candidate record, and its interview covers your roles. Greenhouse's is a two-way voice interview; Workday's Paradox products are described as screening, scheduling and text-based conversations. > - "Workday, Greenhouse and Dayforce all bundle AI interviews" overstates Workday and Dayforce: Dayforce's deal with Ribbon is a signed agreement, not a closed acquisition. Fact-checked on 3 and 4 Nov 2026 against the vendors' own pages ([see Sources](#sources)). You may have heard that Workday, Greenhouse and Dayforce now bundle AI interviews, so a separate tool is redundant. Each vendor's own pages say something narrower. Greenhouse is the one that sells a two-way AI interview, and the other two differ. Workday's Paradox products screen, schedule and run text-based conversations. Dayforce has a signed deal for a voice and video interviewer, Ribbon, that it has not yet closed. A dedicated interviewer adds what a bundled one is not described as doing: Tracy AI Interviewer can read each candidate's CV and tailor 3 to 5 questions to it, follows up on what the person said, interviews every applicant when they click the link at any hour, and returns a recording, a transcript timed to the recording, a written assessment per question and integrity signals for a reviewer. It also runs beside whichever ATS you have. ## What each ATS vendor offers today | ATS vendor | Product | What its own pages describe | Status | |---|---|---|---| | Greenhouse | Voice AI | A two-way voice interview with follow-up questions, a score against your rubric, recording and transcript | Available. Came from the Ezra AI Labs acquisition, completed 27 May 2026 | | Workday | Paradox Candidate Experience Agent and Paradox Conversational ATS | The agent engages candidates, screens them for basic qualifications and schedules interviews. The Conversational ATS lets candidates search for jobs, apply, interview and onboard through short text-based conversations | Available through Workday. Workday completed the Paradox acquisition on 1 Oct 2025, and announced the Conversational ATS on 8 Jan 2026 | | Dayforce | Ribbon, built into Dayforce after the deal closes | Conversational voice and video interviews, with questions and a scoring rubric generated from the job description and relevant past interviews | Ribbon is sold on its own today (see ribbon.ai), with hundreds of customers according to Dayforce's release. Dayforce announced on 23 Oct 2026 that it has signed an agreement to acquire it; not closed, no date for the Dayforce version | Greenhouse's Voice AI is a structured interview, and Greenhouse says it is "not a one-way recording, not a chatbot." It asks dynamic follow-up questions, keeps the score, recording, transcript and explanations, and flags inconsistencies to the moment they happened. For a head-to-head, see [Greenhouse's AI interviewer vs Tracy](/compare/greenhouse-ai-interviewer-vs-tracy). Workday's datasheet says the agent "engages, screens, and schedules candidates automatically", and the screening is for basic qualifications. Workday's 8 Jan 2026 release adds the Conversational ATS, where candidates apply and interview "through short text-based conversations". Neither page describes an AI-led interview with follow-ups and a scored result, so this is a different job from Greenhouse's. Workday says it is built particularly for frontline industries. [Tracy vs Paradox](/compare/tracy-vs-paradox) sets out the difference. Dayforce has not closed its purchase of Ribbon. Its release says it "signed an agreement to acquire Ribbon AI Inc." and that after the close Ribbon will be integrated natively into a new Dayforce Agentic Recruiting experience. Ribbon itself is a voice and video interviewer you can buy from Ribbon today, and [Tracy vs Ribbon](/compare/tracy-vs-ribbon) shows how it compares. The release gives no date for it to ship inside Dayforce. ## What a built-in interviewer does well Greenhouse says scores, transcripts and rationale land on the candidate record in your ATS, with integration depth varying by ATS, so recruiters do not open a second tool. Workday's agent schedules interviews automatically once a candidate qualifies, which keeps scheduling in the same system. And a built-in interviewer is not always confined to one ATS: Greenhouse says Voice AI works with or without the Greenhouse ATS and integrates with 60+ ATS providers. If your roles are high volume and your recruiters live in the ATS, these points count for a lot. ## What you give up The list below is limited to what the vendors' pages say or leave unstated. A page that does not mention something does not prove the product lacks it, so treat each item as a question for the vendor. On Greenhouse Voice AI: - Video. The pages describe a voice interview. If you want to see the person, or run one format for every applicant across roles, ask. - A coding or technical task. None is described on the Voice AI pages. - A published price. Voice AI is an annual subscription that scales with your organisation, with a fair-use interview allowance and no number on the page. The page lists no free trial and describes pilot periods as an opt-out window on the annual agreement. On Workday's Paradox products: - The interview itself. The pages describe screening for basic qualifications, scheduling and text-based conversations, not a follow-up interview with evidence. - A published price. Workday's datasheet ends with sales contacts and no price. On Dayforce: - The Dayforce version. Until the deal closes and the integration ships, there is no Ribbon inside Dayforce. You can buy Ribbon directly, and the Dayforce release gives no pricing. Check the format, the technical depth and the price in each case. ## What a dedicated interviewer adds, and what it costs Tracy's capabilities, set against the bundled options above: - CV-based questions. With CV check on, Tracy asks 3 to 5 questions tailored to the CV and the job's must-have requirements, on top of the role's questions. None of the bundled products above is described on its pages as building questions from each candidate's CV. - Follow-ups that wait. Tracy asks follow-ups on what the person actually said, and turn-taking waits for the end of a thought, so a long technical answer is not cut off. It can introduce a coding task and probe the trade-offs. - Video. Two-way video interviews. Greenhouse describes voice, and Workday's products are described as text-based conversations. - Every applicant, any hour. Each interview starts when the candidate clicks the link, with no recruiter and no calendar, so hundreds of applicants can be interviewed over a hiring round. - Evidence. A recording of both sides, a transcript timed to the recording, a written assessment per question with its reasoning, and integrity signals (face and gaze, response timing, tab changes, leaving fullscreen, an external screen and other unusual patterns) with their evidence, described in Tracy's post [how AI interview cheating detection works](/blog/ai-interview-cheating-detection) and listed on [Tracy's home page](https://tracyinterview.ai/). They are signals for a reviewer, not a verdict. - Stack. API and MCP connections, three ways to start an interview (an invitation link, a self-serve apply page, or your own system), and customised versions that enterprise teams can ask for. The [Connect page](https://tracyinterview.ai/connect) describes the MCP server for Claude, ChatGPT, Mistral, Gemini and Copilot Studio. Pricing is published, which is a separate point from capability: see [what an AI interview costs per candidate](/blog/ai-interview-cost) for the arithmetic. What it costs you: - A second tool. Tracy's Workable page says results stay in Tracy, and you ask the chat to read them there. Greenhouse's sync to the candidate record is more direct. - No native connector for your ATS is described. Tracy's help lists a Workable guide and a Breezy HR guide, where you invite candidates from the same Claude or ChatGPT chat as the ATS's own connector; Workable shares name and email. It also has a Zapier app, whose Result Ready trigger carries the result link and your own reference, such as a CRM record ID. Those can carry a result to the next step in your process. The pattern is in [AI interviews in Workable](/blog/tracy-workable-integration), and the ChatGPT setup is in [AI interviews in ChatGPT](/blog/tracy-chatgpt-integration). - Less around the interview. Tracy's pages describe neither scheduling of human interviews nor a place on the candidate record in your ATS. Greenhouse does ask follow-ups, accepts your own questions, keeps evidence and flags integrity signals, so Tracy's follow-ups and evidence are not exclusive. The differences that hold on the vendors' pages are video, questions built from the CV, the coding task, the integrity evidence, the API and MCP connections and running beside any ATS. ## How to decide Stay with the built-in interviewer when: - recruiters will not leave the ATS and need results on the record; - the interview is voice and you do not need a coding task; - you want one contract and are comfortable with a quote-based annual agreement; - your ATS vendor's interviewer is actually available for the roles you hire. Add a dedicated interviewer when: - every applicant should be interviewed with questions tailored to their CV; - the first round should be on video or include a coding task, with a recording, transcript and integrity signals behind each result; - you want API and MCP connections and a tool that works beside whichever ATS you run; - you want to see a price and run three real interviews before talking to sales; - your ATS vendor's offering describes screening, scheduling and short text-based conversations rather than an AI-led interview with follow-ups; - you are waiting on an announcement, as Dayforce customers are, and need an interview now. To test, run one real role through each candidate. Include one vague answer and watch the next question. Ask where the result lands, what the contract looks like after the pilot, and which formats and tasks are included. The [guide to what an AI interviewer is](/blog/what-is-an-ai-interviewer) lists what to look for, and [one-way vs two-way interviews](/blog/one-way-vs-two-way-video-interviews) explains why the second question matters. ## Common questions ### Do Workday, Greenhouse and Dayforce all include an AI interviewer? Not in the same way. Greenhouse sells Voice AI, a two-way voice interview. Workday sells the Paradox Candidate Experience Agent, which its datasheet describes as engaging, screening and scheduling candidates, and Paradox Conversational ATS, where candidates apply and interview through short text-based conversations. Dayforce has signed an agreement to acquire Ribbon, a voice and video interviewer that is sold on its own today, but the deal had not closed when we checked. ### Does Workday's AI agent run scored interviews? The Workday pages we read describe screening for basic qualifications, automatic interview scheduling and text-based conversations. They do not describe an AI-led scored interview, so ask Workday what the products cover for your roles. ### Has Dayforce bought Ribbon? Dayforce announced on 23 Oct 2026 that it has signed an agreement to acquire Ribbon AI. The release gives no closing date, and says Ribbon will be integrated natively into Dayforce Agentic Recruiting after the close, with no ship date. ### What do you give up with an ATS's built-in AI interviewer? It depends on the ATS. Greenhouse Voice AI describes a voice interview, describes no coding task and publishes no price. Workday's products are described as screening, scheduling and text-based conversations. Dayforce's built-in option is not yet available. Check the format, technical depth and price for your own roles. ### Can I add a dedicated AI interviewer without switching my ATS? Yes. Tracy runs from a link and needs no ATS connection, and it connects through API and MCP. Tracy's help pages describe Workable and Breezy HR guides, where results stay in Tracy, plus a Zapier app that can carry a result link or results to the next step. ### Is a built-in AI interviewer cheaper? We cannot say. Greenhouse, Workday and Dayforce publish no price for their AI interview products on the pages we checked. Tracy publishes $7 per credit with the first three results free. --- **[Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=ats-ai-interviewer-vs-dedicated&utm_content=try-free) with three results on us, or [book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=ats-ai-interviewer-vs-dedicated&utm_content=demo) to see it against your ATS.** ## Sources All sources retrieved 3 and 4 Nov 2026. - Greenhouse, [Voice AI interviewing](https://www.greenhouse.com/voice-ai-interviewing): format, follow-ups, evidence, ATS reach, pricing model and pilot terms. - Greenhouse, [Voice AI now works with 60+ ATS platforms](https://www.greenhouse.com/product-features/voice-ai-now-works-with-60-ats-platforms): results landing on the ATS record and integration depth, dated September 2026. - Greenhouse, [Greenhouse completes acquisition of Ezra AI Labs](https://www.greenhouse.com/newsroom/greenhouse-completes-acquisition-of-ezra-ai-labs-bringing-conversational-ai-to-the-hiring-process): acquisition completed 27 May 2026. - Workday, [Workday completes acquisition of Paradox](https://newsroom.workday.com/2025-10-01-Workday-Completes-Acquisition-of-Paradox): acquisition completed 1 Oct 2025, availability for purchase and frontline focus. - Workday, [Paradox Conversational ATS now available through Workday](https://newsroom.workday.com/2026-01-08-Paradox-Conversational-Applicant-Tracking-System-ATS-Now-Available-Through-Workday,-Helping-Organizations-Hire-Frontline-Workers-Faster): release of 8 Jan 2026, candidates apply, interview and onboard through short text-based conversations. - Workday, [AI agents for Workday Recruiting datasheet](https://www.workday.com/content/dam/web/en-us/documents/datasheets/ai-agents-for-workday-recruiting-datasheet-enus.pdf): the agent engages, screens for basic qualifications and schedules. - Dayforce, [Dayforce to acquire Ribbon AI](https://www.dayforce.com/who-we-are/newsroom/dayforce-to-acquire-ribbon-ai-inc-to-advance-agentic-recruiting): signed agreement of 23 Oct 2026, not closed; Ribbon's voice and video interviews, rubric generation and customer count. - Tracy AI Interviewer, [home page](https://tracyinterview.ai/): video interview, follow-ups, coding task, assessment per question, integrity signals. - Tracy AI Interviewer, [product page](https://tracyinterview.ai/product): questions from a brief or your own, CV check, recording and transcript timed to the recording, three ways in, API and MCP. - Tracy AI Interviewer, [pricing](https://tracyinterview.ai/pricing): credit price, billing by started 20 minutes, free results, no-answer policy, no seats, API and MCP connections and a customised setup for Enterprise. - Tracy AI Interviewer, [AI interview cheating detection](https://tracyinterview.ai/blog/ai-interview-cheating-detection): the primary source for the integrity signals. - Ribbon, [ribbon.ai](https://www.ribbon.ai/): Ribbon sold on its own, with a pricing page and a demo, and a banner saying it has signed an agreement to be acquired by Dayforce. - Tracy AI Interviewer, [Connect page](https://tracyinterview.ai/connect): Tracy runs as an MCP server for Claude, ChatGPT, Mistral, Gemini and Copilot Studio. - Tracy AI Interviewer, [help](https://tracyinterview.ai/help): the Use it with your ATS section (Workable, Breezy HR) and the Zapier entry. - Tracy AI Interviewer, [Workable help page](https://tracyinterview.ai/help/workable): invite flow through chat, shared data and results staying in Tracy. - Tracy AI Interviewer, [Zapier help page](https://tracyinterview.ai/help/zapier): the Result Ready trigger and the reference carried back. - Tracy AI Interviewer, [Breezy HR help page](https://tracyinterview.ai/help/breezy): Breezy HR guide, name and email shared, results stay in Tracy. # Greenhouse's AI Interviewer vs Tracy: Is Your ATS's Interviewer Enough? (Nov 2026) > Greenhouse Voice AI vs Tracy AI Interviewer: format, follow-ups, technical tasks, evidence, ATS fit and price, with the cases where Greenhouse is the better pick. - URL: https://tracyinterview.ai/compare/greenhouse-ai-interviewer-vs-tracy - Published: 2026-11-05 - Author: Tracy team > **In short** > - A team already on Greenhouse might add **Tracy AI Interviewer** for video interviews that start when each applicant clicks, with questions drawn from their CV and a recording, transcript and integrity signals to review. It needs no ATS connection to run. > - Pick **Greenhouse Voice AI** when you already work in Greenhouse, or in one of the 60+ ATS it integrates with, and want a scored voice interview with the result on the candidate record. > - Both ask follow-up questions, accept your own questions, keep evidence and flag integrity signals. They differ on format, technical tasks, price model and where the result lands. Fact-checked on 3 and 4 Nov 2026 against the vendors' own pages ([see Sources](#sources)). Greenhouse added AI interviewing by acquisition. It completed the acquisition of Ezra AI Labs, a voice AI interviewer, on 27 May 2026, and Ezra's interviewer now sits inside Greenhouse as Voice AI. If your applicant tracking system already offers an AI interviewer, is that enough, or should you add a dedicated one such as Tracy AI Interviewer? The wider version of that question is in [the AI interviewer in your ATS vs a dedicated one](/blog/ats-ai-interviewer-vs-dedicated). Tracy is built for the case where a voice screen is not enough. It can read each candidate's CV and ask 3 to 5 questions tailored to it and to the role, and it follows up on what the person actually said, waiting for the end of the answer before it speaks. Every applicant is interviewed when they click the link, at any hour, with no scheduling, so a large pool does not become a shortlist. Each result comes with the recording, a transcript timed to the recording, a written assessment per question with its reasoning and integrity signals for a reviewer. If the category is new to you, [what an AI interviewer is](/blog/what-is-an-ai-interviewer) is the place to begin. ## Greenhouse Voice AI vs Tracy at a glance | Buying question | Tracy AI Interviewer | Greenhouse Voice AI | |---|---|---| | Format | Two-way video interview, recorded on both sides | Two-way voice: "Not a one-way recording, not a chatbot" | | Follow-up questions | Yes: a vague answer gets a sharper question, and each follow-up depends on the answer before it | Yes, dynamic and tied to the team's rubric | | Coding or technical task | Can introduce a coding task and probe the trade-offs | Not described on the Voice AI pages | | CV or resume | Can read the CV and ask 3 to 5 questions tailored to it and to the role's must-haves (a setting) | A resume authenticity layer; questions built from the CV are not described on the pages we read | | Volume and scheduling | Each applicant starts when they click, at any hour, with no recruiter and no calendar | Two-way voice interview; the pages state no interview count | | Your own questions | Yes: questions drafted from the job description or your brief, or bring your own | Yes: recruiters and hiring managers set questions, rubric and role context | | What the reviewer gets | Recording of both sides with a transcript timed to the recording, a written assessment per question with its reasoning, and integrity signals with their evidence | Score, recording, transcript, explanations and flags marked to the moment | | Cheating or fraud checks | Integrity signals in the interview report: tab changes, abnormal eye tracking, live face detection, delay before an answer, external screens and other unusual patterns. They are signals for a reviewer, not a verdict | Cheat detection with timestamps, plus fraud detection and a resume authenticity layer | | Languages | One interview language chosen at setup | 11 | | ATS and connections | No ATS needed; runs from a link or an invitation. API and MCP connections (MCP guides for Claude, ChatGPT and Workable), Breezy HR guide, Zapier; customised versions can be requested by enterprise teams | 60+ providers; works without Greenhouse ATS | | Where results land | In Tracy; a Zapier step can pass the result on | On the candidate record in your ATS, with depth varying by ATS | | Published price | $7 per credit; first 3 results free | No number: annual subscription scaled to your organisation | ## Where Greenhouse Voice AI is the better pick Greenhouse may be enough for your roles. Its pages describe a structured voice interview with dynamic follow-up questions anchored to your rubric. It also answers candidates' questions, drawing only on information your team approved. The evidence list is full: a score, the recording, the transcript, detailed explanations of the score, and inconsistencies in candidate behaviour marked to the moment they happened. Tracy's home page describes a written assessment for every question, with the reasoning and the timestamps behind it, so you read the words and then decide; [what reviewers should see in an AI interview report](/blog/ai-interview-reports) covers that approach. Integrity checks exist on both sides. Greenhouse lists cheat detection that flags signs of outside help, timestamped in the transcript, plus a fraud detection and resume authenticity layer. Tracy's interviews include integrity signals too, described in Tracy's post [how AI interview cheating detection works](/blog/ai-interview-cheating-detection) and listed on [Tracy's home page](https://tracyinterview.ai/). We publish no detection rates for Tracy's signals, and we have none to compare with Greenhouse's. Greenhouse also writes back. It says scores, transcripts and rationale land on the candidate record in your ATS, across 60+ providers, with integration depth varying by ATS. Tracy's help pages show a different route, covered below, and if recruiters live in the ATS and will not open a second tool, that is a real cost of choosing Tracy. Greenhouse also publishes monthly independent bias audits and says SOC 2 Type II audit reports are on its trust portal. Its FAQ says the team makes every hiring decision, and its 60+ ATS page lists candidate actions that let a recruiter move a candidate forward or reject them from inside Voice AI. Tracy's pages say the decision stays with you. ## Where Tracy differs Tracy reads the CV. With CV check on, it asks 3 to 5 questions tailored to the CV and the job's must-have requirements, on top of the role's questions. Greenhouse's pages describe a resume authenticity layer and rubric-tied questions your team sets; they describe no questions built from each candidate's CV. For skilled roles, that is the difference between asking everyone the same thing and asking each person about their own work. Tracy interviews every applicant on their own time. Each interview is its own session that starts when the candidate clicks the link, at any hour, with no recruiter and no scheduling, so hundreds of applicants can be interviewed over a hiring round. Greenhouse's pages state no interview count, so test the volume you expect with each. Tracy's evidence is built for the reviewer who was not there: video of both sides, a transcript timed to the recording, a written assessment per question with its reasoning, and integrity signals with their evidence. Greenhouse offers a recording, transcript and flags too; the differences to test are the video, the per-question reasoning and what each signal shows. Tracy also connects through an API and an MCP server, so interviews can be created and read from the tools a team already works in, and enterprise teams can ask for a customised setup. Greenhouse's 60+ ATS integrations are the more direct route when results must sit on the candidate record. Tracy interviews are two-way video conversations. Greenhouse describes voice. If seeing the person matters for the role, test that difference on a real role. Tracy's home page says it can introduce a coding task and probe the trade-offs behind an answer. No coding or technical task appears on the Greenhouse Voice AI pages we read, which does not mean Greenhouse cannot do it, so ask. If your first round is technical, see how a [technical AI interview works](/blog/ai-technical-interviews). Tracy also asks nothing of the candidate beyond a link. There is no install for the person being interviewed. ## How Tracy fits with your ATS Tracy does not need an ATS connection to run, and it works beside any ATS. It connects through an API and an MCP server, and its help pages describe several ways to use it next to one. - MCP: the [Connect page](https://tracyinterview.ai/connect) describes Tracy as an MCP server you add to Claude, ChatGPT, Mistral, Gemini or Copilot Studio, so you can set up an interview by asking in the chat. - Workable and Breezy HR: the help page lists both under "Use it with your ATS". You connect Tracy and the ATS's own connector to the same Claude or ChatGPT chat, invite candidates from there, and read results back through the chat. On the Workable page, Workable shares each candidate's name and email and the interview results stay in Tracy; Breezy HR has a [matching guide](https://tracyinterview.ai/help/breezy). See [how Tracy works with Workable](/blog/tracy-workable-integration). - Zapier: Tracy is an app with five triggers and an invite action. Result Ready starts a Zap when a result can be read, and carries the link to the result plus your own reference for the person, such as a CRM record ID, so the next step finds the right record. None of those pages describes Tracy writing a score or summary into an ATS record by itself. A Zap or flow can carry the result link or the results to the next step, and that is the route to try if your recruiters stay in the ATS. Greenhouse's sync is the more direct one. ## How pricing compares | Pricing item | Tracy AI Interviewer | Greenhouse Voice AI | |---|---|---| | Published number | $7 per credit | None | | Model | Pay per result: 1 credit per started 20 minutes | Annual subscription scaled to headcount, fair-use interview allowance | | Seats | No seats | Priced as a standalone product, not a per-seat add-on | | Trial | First 3 results free, no card needed to start | No free trial is listed; the page describes optional pilot periods as an opt-out window on the annual agreement, and an interactive demo | | Contract | Credits never expire | Customers typically start with 12 months or more | The cheaper option depends on your volume, and Greenhouse's number is not public, so we cannot give you a break-even point. Get the Greenhouse quote, then run the same arithmetic on Tracy's published rates. Ask Greenhouse what counts as fair use and what happens above the allowance. ## Which should your team choose? Choose Greenhouse Voice AI when: - recruiters work in Greenhouse, Ashby, Lever, Workday or another of its 60+ ATS and want scores on the candidate record; - you want Greenhouse's scored voice interview, cheat detection and fraud layer in one product; - you need interviews in 11 languages; - a bias-audit and governance story matters to your compliance review; - an annual contract is fine and you want one vendor. Choose Tracy AI Interviewer when: - every applicant should get an interview, at any hour, with questions tailored to their CV; - the first round should be on video; - reviewers need the recording, a transcript timed to the recording, a written assessment per question and integrity signals; - a coding task with follow-ups belongs in the first conversation; - you want to pay per result, with no seat fee; - you are not ready to connect an ATS, or you are happy to route results through Zapier; - you want to read the words behind each assessment, next to the recording, and then decide. You can also run both. Greenhouse says Voice AI works with or without its ATS, and Ezra stays available as a standalone product, so neither tool ties you to a single ATS. A reasonable split is Greenhouse for roles where a voice screen is enough and Tracy for technical roles where a coding task and video matter. Run one role through both and compare the second question each one asks. [Tracy vs HireVue](/compare/tracy-vs-hirevue) shows the same test against a different rival. ## Common questions ### Does Greenhouse have an AI interviewer? Yes. Greenhouse Voice AI is a two-way voice interview with follow-up questions, a score against the team's rubric, a transcript and a recording. It comes from Greenhouse's acquisition of Ezra AI Labs, completed on 27 May 2026. ### Do I need the Greenhouse ATS to use Greenhouse Voice AI? No. Greenhouse says it works with or without the Greenhouse ATS and integrates with 60+ ATS providers, with integration depth varying by ATS. ### Does Greenhouse Voice AI do video interviews or coding tasks? The Greenhouse pages we read describe a voice interview. They do not describe video or a coding task, so check both with Greenhouse before you rely on them. ### How much does Greenhouse Voice AI cost? Greenhouse does not publish a number. Voice AI is sold as an annual subscription that scales with your organisation, with a fair-use interview allowance. Tracy publishes $7 per credit, and each started 20 minutes of a result costs one credit. ### Is there a free trial of Greenhouse Voice AI? Greenhouse's page lists no free trial. It describes optional pilot periods as an opt-out window on the annual agreement, and offers an interactive demo. Tracy includes the first three results free. ### Can I use Tracy if my team already works in Greenhouse? Yes. Tracy runs from a link and does not need an ATS connection, so you can invite people from any list, and it connects through API and MCP. Tracy's help pages describe Workable and Breezy HR guides and a Zapier app whose Result Ready trigger carries the result link and your own record reference. They do not describe a native Greenhouse connector. ### Does Tracy have cheating checks like Greenhouse? Tracy interviews include integrity signals in the report: tab changes, abnormal eye tracking, live face detection, average delay before an answer, external screens and other unusual patterns. They are signals for a reviewer to look at, not a verdict. --- **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=greenhouse-ai-interviewer-vs-tracy&utm_content=demo) with one of your own first-round briefs, or [try Tracy free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=greenhouse-ai-interviewer-vs-tracy&utm_content=try-free) with three results on us.** ## Sources All sources retrieved 3 and 4 Nov 2026. - Greenhouse, [Voice AI interviewing](https://www.greenhouse.com/voice-ai-interviewing): format, follow-ups, evidence, cheat detection, questions, languages, ATS integrations, pricing model, pilot terms and governance statements. - Greenhouse, [Voice AI now works with 60+ ATS platforms](https://www.greenhouse.com/product-features/voice-ai-now-works-with-60-ats-platforms): integration depth, results landing on the ATS record and fraud detection, dated September 2026. - Greenhouse, [Greenhouse Voice AI feature page](https://www.greenhouse.com/product-features/greenhouse-voice-ai): scope read for coding or technical tasks, dated August 2026. - Greenhouse, [Greenhouse completes acquisition of Ezra AI Labs](https://www.greenhouse.com/newsroom/greenhouse-completes-acquisition-of-ezra-ai-labs-bringing-conversational-ai-to-the-hiring-process): acquisition completed 27 May 2026 and Ezra remaining a standalone product. - Tracy AI Interviewer, [home page](https://tracyinterview.ai/): video interview, follow-ups, coding task, assessment per question, integrity signals, no card needed for the first 3 results. - Tracy AI Interviewer, [product page](https://tracyinterview.ai/product): questions from a brief or your own, CV check, recording and transcript timed to the recording, one interview language chosen at setup, no install for the person interviewed. - Tracy AI Interviewer, [hiring at volume](https://tracyinterview.ai/use-cases/volume-hiring): each applicant starts on click at any hour, with no recruiter and no calendar. - Tracy AI Interviewer, [pricing](https://tracyinterview.ai/pricing): credit price, billing by started 20 minutes, free results, no-answer policy, no seats and a setup customised to your workflow. - Tracy AI Interviewer, [AI interview cheating detection](https://tracyinterview.ai/blog/ai-interview-cheating-detection): the primary source for the integrity signals (tab changes, abnormal eye tracking, live face detection, average delay before an answer, external screens, other unusual patterns), reported for review while a person decides. - Tracy AI Interviewer, [Connect page](https://tracyinterview.ai/connect): Tracy runs as an MCP server for Claude, ChatGPT, Mistral, Gemini and Copilot Studio. - Tracy AI Interviewer, [help](https://tracyinterview.ai/help): the Use it with your ATS section (Workable, Breezy HR) and the Zapier entry. - Tracy AI Interviewer, [Workable help page](https://tracyinterview.ai/help/workable): invite flow through chat, name and email shared, results staying in Tracy. - Tracy AI Interviewer, [Breezy HR help page](https://tracyinterview.ai/help/breezy): the matching guide, where Breezy HR shares name and email and results stay in Tracy. - Tracy AI Interviewer, [Zapier help page](https://tracyinterview.ai/help/zapier): five triggers, the invite action, the Result Ready trigger and the reference carried back. # Paradox Alternatives for Skilled Roles (Nov 2026) > Paradox suits hourly chat screening. For skilled roles, compare seven interviewers on format, follow-ups, technical tasks and published price. - URL: https://tracyinterview.ai/compare/paradox-alternatives - Published: 2026-11-05 - Updated: 2026-11-11 - Author: Tracy team Checked on 3 Nov 2026 against each vendor's own pages. Where a vendor publishes no price or does not describe a capability, the table says so. The page rests on vendor pages, not on reviews or forum threads. > **In short** > - For skilled roles, **Tracy AI Interviewer** is the first alternative to test: every applicant gets a two-way video interview on click, at any hour, and CV check can add 3 to 5 tailored questions. Paradox still wins hourly chat screening, scheduling and languages. > - HireVue, Greenhouse Voice AI, Ribbon, HeyMilo and Alex also run AI interviews, and Sapia.ai runs a text-chat one. Of the tools here, only Tracy and Ribbon put a price in public, and Ribbon's is an unconfirmed figure on a dated blog post. ## Where does Paradox fit before you leave it? Paradox is built for hourly and frontline hiring. Its home page describes automation for frontline workers and managers, and candidates apply and are screened by SMS, WhatsApp or Facebook Messenger. It claims interviews scheduled within 10 minutes (a vendor claim), handles multi-person, multi-room, multi-location scheduling, and answers candidate questions in more than 100 languages. Its FAQ says it serves both frontline hourly and high-skilled professional roles. For the skilled side, the Candidate Experience Agent page describes Q&A and complex scheduling, and Workday's datasheet says the agent screens applicants "for basic qualifications". Paradox's video page covers screening prompts, viewing responses and hosting video interviews between people. The pages we read do not describe an AI-led interview that asks follow-ups on a candidate's answer. If that is what you need for skilled roles, the tools below are worth testing. If your bottleneck is booking interviews for shift workers, stay with Paradox and read [Paradox pricing](/compare/paradox-pricing) before you negotiate. ## Which alternatives suit skilled roles? | Tool | Format | Follow-up questions | Technical task | Published price | |---|---|---|---|---| | Tracy AI Interviewer | Two-way video, started by the applicant's click, any hour | A vague answer gets a sharper question; waits for the end of a thought; optional CV check adds 3 to 5 tailored questions | Can introduce a coding task | $1 per credit, 3 free | | HireVue | AI interviewer, plus one-way and live video | Yes | Coding assessments alongside the interview | None | | Greenhouse Voice AI | Two-way voice | Dynamic, anchored to the team's rubric | Not described on pages read | Quote on request | | Ribbon | Voice and video | Follows up when an answer needs more detail | Not stated | Blog of 23 Aug: starts at $499/month for 100 interviews, annual; unconfirmed | | HeyMilo | Voice, video, SMS, form | Real time; recruiter sets the maximum | AI-fluency scenario assessment; no coding interview described | None | | Alex | Video, phone, SMS, WhatsApp | Based on what the candidate says | Not on current pages; its YC launch post lists coding | None | | Sapia.ai | Text chat | Not stated | Not stated | Per hire, tailored | | Paradox | SMS, WhatsApp, Messenger chat | Not described on pages read | Not described | None | "Not stated" and "not described" mean we read the vendor's pages and found no mention. They do not mean the product lacks the feature. ## Tracy AI Interviewer Tracy runs a spoken, two-way video interview in the browser, with no install for the person being interviewed. It is built for the case where a skilled role draws hundreds of applicants and a text screen would flatten them. - **Everyone who applies is interviewed.** Each applicant starts when they click the link, at any hour, with no recruiter and no calendar, and the interview is ready in under 10 seconds. A candidate who drops can rejoin within 10 minutes, and a reminder goes out after a day for unfinished interviews. - **It reads the CV.** With CV check on, Tracy asks 3 to 5 questions tailored to the CV and the job's must-have requirements, on top of the role's questions. - **It follows up.** Each follow-up depends on the answer before it, a vague answer gets a sharper question, and hard technical questions get longer pauses. For technical roles it can introduce a coding task and probe the trade-offs. - **It shows its evidence.** Each result has the recording, a transcript timed to it, and a written assessment per question with its reasoning. Integrity signals (tab changes and fullscreen, face and gaze, response timing, external screens) are marked on the recording and go to a reviewer as evidence. - **It fits your stack.** An interview can start from an invite link, an apply page or your own platform through the API or MCP, and Enterprise adds API and MCP connections and a setup customised to your workflow. The first 3 results are free, then $1 per credit for each started 20 minutes, with no seats. Pick Tracy when the role is skilled and the first conversation should test depth for every applicant, not a shortlist. Pick something else if you need scheduling or text-message screening at hourly volume; see [Tracy vs Paradox](/compare/tracy-vs-paradox). ## HireVue HireVue pairs an AI interviewer with one-way and live video, and says it evaluates each response against role-specific criteria and explains why. It also offers coding assessments, which it describes as combining coding challenges and interviews in one experience. It states "AI Informs. Humans decide." It publishes no price: its pricing page points to a plan conversation. Pick HireVue when governance, audit trails and one suite at enterprise volume matter most. The full list is in [HireVue alternatives](/compare/hirevue-alternatives). ## Greenhouse Voice AI Greenhouse describes a structured two-way voice interview scored against the team's own rubric, with dynamic follow-ups, a recording, a transcript and per-question scoring. It works with or without the Greenhouse ATS. Voice only: the pages we read describe no video. Pricing is an annual subscription scaled to headcount, quoted on request. Pick it when your team already runs Greenhouse or wants a voice screen tied to a rubric. Compare it with a dedicated tool in [the ATS interviewer vs a dedicated one](/blog/ats-ai-interviewer-vs-dedicated). ## Ribbon Ribbon runs voice and video interviews that follow up when an answer needs more detail, and also covers sourcing and outreach. Its blog, dated 23 August 2026, says public annual-billing packages start at $499/month for 100 interviews. Its pricing page returned a 404 when we checked, so we could not confirm the figure. Dayforce has signed an agreement to acquire Ribbon, and as of its 23 October 2026 release the deal had not closed. Pick Ribbon when you want sourcing and a first interview in one tool. See [Ribbon alternatives](/compare/ribbon-alternatives). ## HeyMilo HeyMilo screens and interviews over voice, video, SMS and forms, and offers a scenario assessment that scores how a candidate works with an AI assistant. The recruiter sets the questions and a maximum number of follow-ups, and follow-ups tie to the scoring rubric. Its /pricing address redirects to a demo form. Pick HeyMilo for hourly or staffing volume where you want SMS and forms beside the interview, and calendar booking after the screen. See [HeyMilo alternatives](/compare/heymilo-alternatives). ## Alex Alex interviews over video, phone, SMS and WhatsApp and asks follow-ups based on what the candidate says. Its /pricing address redirects to a demo form, and its current product pages describe behavioral and role-specific interviews, with no coding task. Its early YC launch post lists coding among the interview types, so ask. Its coordinator books qualified candidates on team calendars. Pick Alex when you want one recruiting workflow from outreach to scheduling. See [Alex alternatives](/compare/alex-apriora-alternatives). ## Sapia.ai Sapia.ai runs a text-chat interview that works on any phone. Its pricing is per hire at a tailored rate, and it says it fits companies that hire more than 500 people a year. It is the closest in format to Paradox among the tools here, so pick it if you want a scored chat interview rather than a spoken one. ## How should you choose? Start from the job you are hiring for. - Hourly volume where booking is the bottleneck: keep Paradox, or look at HeyMilo or Alex, which book interviews as well. - Skilled roles where the first conversation should test depth: Tracy, HireVue, Greenhouse Voice AI or Ribbon, then confirm what each does with a vague answer. - A coding interview as part of the first round: Tracy, or HireVue with its coding assessments. Alex's early launch post also lists coding, so ask it too. - An ATS-first buy: ask whether the interview result is written back. Tracy offers API and MCP connections and can run beside an ATS connector such as Workable's inside Claude or ChatGPT; the pages we checked describe no direct ATS write-back, so ask. Run two real roles through your shortlist, including a vague answer, and compare the next question each tool asks. The wider buyer list is in [AI interview tools for high-volume skilled hiring](/blog/ai-interview-tools-for-high-volume-hiring) and [best AI interview software](/blog/best-ai-interview-software). ## Common questions ### What is the best Paradox alternative for skilled roles? It depends on the job. Tracy AI Interviewer runs two-way video interviews with follow-ups in the candidate's field and publishes its price. HireVue, Greenhouse Voice AI, Ribbon, HeyMilo and Alex also run AI interviews; compare them on format, follow-ups, technical tasks and price. ### Why would a team leave Paradox? Teams hiring skilled roles may want an interview that goes beyond screening for basic qualifications, as Workday describes the agent's screening, or want a price they can read before a demo. Paradox remains the stronger fit for hourly, frontline volume and scheduling. ### Which Paradox alternatives publish a price? Tracy publishes $1 per credit with 3 results free. Ribbon's own blog, dated 23 Aug 2026, says public annual-billing packages start at $499/month for 100 interviews, but its pricing page returned an error when we checked, so treat that figure as unconfirmed. HireVue, Greenhouse, HeyMilo, Alex and Sapia.ai publish no list price. ### Can an alternative replace Paradox's scheduling? Not all of them. Paradox is strong at self-scheduling, reminders and panel interviews. Some alternatives book a calendar slot after the screen, and some do not describe scheduling. Check this before you switch if scheduling is your bottleneck. ### Which alternatives can run a coding interview? Tracy can introduce a coding task and probe the trade-offs behind an answer. HireVue offers coding assessments alongside its interviews. HeyMilo's scenario assessment targets AI fluency rather than coding, and Alex's current product pages describe behavioral and role-specific interviews, though its earlier YC launch post lists coding among its interview types. We found no coding interview described for Greenhouse, Ribbon or Sapia.ai, which is not the same as saying they lack one; ask each vendor. --- **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=paradox-alternatives&utm_content=demo)** or [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=paradox-alternatives&utm_content=try-free) ## Sources All sources retrieved 3 Nov 2026. - Paradox, [Paradox home page](https://www.paradox.ai/), frontline positioning and demo-only call to action - Paradox, [Paradox FAQs](https://www.paradox.ai/faqs), hourly and skilled-role scope, languages - Paradox, [Candidate Experience Agent](https://www.paradox.ai/products/candidate-experience-agent), channels, scheduling and the skilled-hiring use - Paradox, [Paradox video](https://www.paradox.ai/products/video), video prompts and hosted human video interviews - Workday, [AI agents for Workday Recruiting datasheet](https://www.workday.com/content/dam/web/en-us/documents/datasheets/ai-agents-for-workday-recruiting-datasheet-enus.pdf), screening for basic qualifications - Tracy AI Interviewer, [Tracy home page](https://tracyinterview.ai/), the integrity signals listed for video interviews, you read the words then you decide - Tracy AI Interviewer, [Tracy product page](https://tracyinterview.ai/product), two-way video interview, CV check, follow-ups, coding task, transcript timed to the recording, three ways in - Tracy AI Interviewer, [Tracy hiring at volume](https://tracyinterview.ai/use-cases/volume-hiring), interview on click at any hour, rejoin and reminders - Tracy AI Interviewer, [Tracy pricing](https://tracyinterview.ai/pricing), price, free results, no seats, API and MCP connections and customised setup for Enterprise - HireVue, [HireVue AI interviewer](https://www.hirevue.com/platform/ai-interviewer), follow-up questions, evidence, human decision - HireVue, [HireVue pricing](https://www.hirevue.com/pricing), no published price - HireVue, [HireVue coding assessments](https://www.hirevue.com/products/coding-assessments), coding assessments combined with interviews - HireVue, [HireVue video interviewing](https://www.hirevue.com/platform/video-interviewing), on-demand (one-way) and live video - Greenhouse, [Greenhouse Voice AI interviewing](https://www.greenhouse.com/voice-ai-interviewing), voice interview, follow-ups, rubric, evidence, pricing status, languages - Ribbon, [Ribbon AI interviewer](https://www.ribbon.ai/ai-interviewer), voice and video interviews, follow-ups - Ribbon, [Ribbon vs BrightHire](https://www.ribbon.ai/blog/ribbon-vs-brighthire), advertised $499 a month for 100 interviews, published 23 Aug 2026 - Dayforce, [Dayforce to acquire Ribbon AI](https://www.dayforce.com/who-we-are/newsroom/dayforce-to-acquire-ribbon-ai-inc-to-advance-agentic-recruiting), signed agreement to acquire Ribbon, not closed - HeyMilo, [HeyMilo interviewing candidates](https://www.heymilo.ai/interviewing-candidates), real-time follow-ups set by the recruiter - HeyMilo, [HeyMilo docs: configuring your AI interviewer](https://docs.heymilo.ai/guides/build-your-interviewer/configuring-your-ai-interviewer.md), the recruiter-set maximum number of follow-ups - HeyMilo, [HeyMilo compared with Alex](https://www.heymilo.ai/compare/alex-ai), follow-ups drawn from the scoring rubric - HeyMilo, [HeyMilo AI scenario assessment](https://www.heymilo.ai/product-feature/ai-scenario-assessment), the AI-fluency scenario task - HeyMilo, [HeyMilo pricing](https://www.heymilo.ai/pricing), pricing redirects to a demo form - Alex, [Alex AI interviewer](https://www.alex.com/product/ai-interviewer), video, phone, SMS and WhatsApp interviews with follow-ups - Alex, [Alex coordinator](https://www.alex.com/product/coordinator), qualified candidates book their next round - Alex, [Alex YC launch post](https://www.ycombinator.com/launches/KY6-alex-fka-apriora-your-ai-interviewer-to-identify-the-best-candidates-faster), interview types listed at launch, including coding - Alex, [Alex pricing](https://www.alex.com/pricing), no public price; demo form - Sapia.ai, [Sapia.ai pricing](https://sapia.ai/pricing/), per-hire pricing, best for 500+ hires a year - Sapia.ai, [Sapia.ai interview](https://sapia.ai/platform/interview/), text chat interview # Paradox Pricing (Nov 2026) > Paradox publishes no price. See what is public, what a third party estimates, who owns it now and how Tracy's published pricing compares. - URL: https://tracyinterview.ai/compare/paradox-pricing - Published: 2026-11-05 - Updated: 2026-11-11 - Author: Tracy team Checked on 3 Nov 2026 against Paradox's and Workday's own pages. Paradox publishes no price, so the one outside figure below comes from a third party and is labelled that way. > **In short** > - For skilled roles, **Tracy AI Interviewer** interviews every applicant who clicks the link, at any hour, and can read each CV to ask 3 to 5 tailored questions. Its price is public: the first 3 results are free, then $1 per credit. > - Paradox publishes no price; you get a quote through a demo request, from Paradox or from Workday. It is the stronger fit for hourly and frontline hiring: chat screening, self-scheduling and candidate Q&A in 100+ languages. ## What does Paradox cost? Paradox does not publish a price. Its /pricing address redirects to the home page, and the call to action on the pages we read is "Request demo". Workday completed its acquisition of Paradox on 1 October 2025 and sells the product as the Workday Paradox Candidate Experience Agent, which it says is available "through Workday or Paradox". Workday's recruiting-agent datasheet carries no price either; it ends with Workday's general phone and fax numbers. | What you want to know | What Paradox and Workday publish | |---|---| | Public price | None | | Plan names and tiers | None | | Volume or enterprise tiers | None | | Free trial | None described on the pages we read | | How to buy | Demo request to Paradox, or sales through Workday | ## What does a third party say it costs? Truffle, a third-party site that sells a competing screening product, reports that independent directories list a starting price around $1,000 a month with no self-serve free trial, and that Paradox "has never published a price". Its guide is dated February 2026 and was updated in October 2026. That is not a Paradox statement and we could not confirm it. Use it, if at all, as a starting point for the quote conversation. ## What are you buying? Paradox sells a conversational assistant, Olivia, that handles applying, screening, scheduling and onboarding over text and chat. Its home page describes "mobile-first automation for frontline workers and managers", and its FAQ says it serves "frontline high-volume hourly and high-skilled professional roles". The scheduling work is where Paradox is strongest. It claims interviews scheduled within 10 minutes (a vendor claim), and its Candidate Experience Agent page describes multi-person, multi-room, multi-location scheduling. Candidates apply and are screened by SMS, WhatsApp or Facebook Messenger, and the assistant answers candidate questions in more than 100 languages. It integrates with Workday and SAP SuccessFactors among others. Pick Paradox when: - you hire hourly or frontline staff in volume and the bottleneck is applying, checking eligibility and booking interviews; - you run Workday Recruiting and want one vendor and a direct Workday integration; - SMS or WhatsApp is the channel your candidates already use, or you need candidate Q&A in many languages. A quote is worth requesting in those cases. The price buys scheduling and text or chat messaging at scale. The Tracy pages we checked describe no scheduling step or text-message screening; Tracy's volume case is a different one, set out below. ## What should you ask for in a quote? Because nothing is published, the quote is the whole price. Ask for these in writing: - the unit that is priced: locations, hires, candidates or conversations; - what the base fee includes, and what costs extra, such as text messaging, extra languages or integrations; - implementation and onboarding fees, and the contract length; - whether Workday or Paradox is the seller, and how that changes support; - whether a pilot is possible, since no trial is described. For other quoted tools, compare the same questions in [HireVue pricing](/compare/hirevue-pricing), [HeyMilo pricing](/compare/heymilo-pricing) and [Alex pricing](/compare/alex-apriora-pricing). ## How does Tracy pricing compare? The two do not price the same thing. Paradox is a chat and scheduling layer priced by quote. Tracy AI Interviewer gives every applicant a spoken interview when they click the link, at any hour and with no scheduling, so a large pool for a skilled role is interviewed in full rather than cut to a shortlist first. It prices each result, with no seats: your first 3 results are free, then $1 per credit. One credit covers each started 20 minutes of an interview that produced a result, so a result of up to 20 minutes costs $1. Credits never expire and a session with no answers is never charged. | Results of up to 20 minutes | Tracy, after the 3 free | |---|---| | 100 | $97 | | 500 | $497 | | 1,000 | $997 | Those Tracy totals equal the number of results after the first three, at $1 each. Paradox has no figure to set beside them. The totals count only the interview; hourly volume that needs scheduling and text screening is a different job. The wider cost method is in [AI interview cost per candidate](/blog/ai-interview-cost). ## Is Paradox worth it, and for whom? For hourly volume with a scheduling bottleneck, the price buys what Paradox is built for. For skilled roles the fit is narrower. Paradox's own page lists the corporate and skilled use as Q&A and complex scheduling, and Workday's datasheet says the agent screens applicants "for basic qualifications". The Paradox pages we read do not describe adaptive follow-up questions on an answer, a coding task or a report that ties a result to what the candidate said. Absence from those pages is not proof the product lacks them, so ask in the demo. Tracy covers the interview side of skilled hiring. It runs a two-way video interview and can read the CV and ask 3 to 5 questions tailored to it and to the role's must-haves. A vague answer gets a sharper question, each follow-up depends on the answer before it, and for technical roles it can introduce a coding task and probe the trade-offs. Each result comes with the recording, a transcript timed to it, a written assessment per question with its reasoning, and integrity signals marked on the recording for a reviewer to weigh. Enterprise teams get API and MCP connections and a setup customised to their workflow. Read [Tracy vs Paradox](/compare/tracy-vs-paradox) for the full comparison, or [Paradox alternatives for skilled roles](/compare/paradox-alternatives) for the wider shortlist. ## Common questions ### How much does Paradox cost? Paradox publishes no price. Its pricing address redirects to its home page and the pages we read offer a demo request only. A third-party estimate puts the starting price around $1,000 a month; treat it as unconfirmed and ask for a written quote. ### Does Paradox have a pricing page? No. The /pricing address redirects to the Paradox home page, and the call to action on the pages we read is a demo request. ### Who sells Paradox now? Workday completed its acquisition of Paradox on 1 Oct 2025. Workday says the Workday Paradox Candidate Experience Agent can be bought through Workday or Paradox. ### Does Paradox have a free trial? The Paradox pages we read describe no trial. A third party reports there is no self-serve free trial. Ask about a pilot when you request a quote. ### How does Tracy's pricing compare? Tracy AI Interviewer publishes its price: the first 3 results are free, then $1 per credit, with one credit for each started 20 minutes of a session that produces a result. There are no seats, and a session with no answers is never charged. ### Is Paradox the right tool for skilled roles? Paradox's own pages cover skilled hiring through candidate Q&A and complex scheduling, and Workday describes screening for basic qualifications. If you want every applicant to get a spoken interview that reads the CV and asks follow-ups in the candidate's field, with a recording and integrity evidence, compare an interviewer such as Tracy. --- **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=paradox-pricing&utm_content=demo)** or [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=paradox-pricing&utm_content=try-free) ## Sources All sources retrieved 3 Nov 2026. - Paradox, [Paradox home page](https://www.paradox.ai/), frontline positioning, text and chat screening, and the demo request as the only call to action - Paradox, [Paradox pricing address](https://www.paradox.ai/pricing), the address redirects to the home page; no pricing page exists - Paradox, [Paradox FAQs](https://www.paradox.ai/faqs), hourly and skilled-role scope, 100+ languages, integrations - Paradox, [Candidate Experience Agent](https://www.paradox.ai/products/candidate-experience-agent), channels, scheduling and the skilled-hiring use - Paradox, [Conversational ATS](https://www.paradox.ai/products/conversational-ats), the 10-minute scheduling claim - Workday, [Workday completes acquisition of Paradox](https://newsroom.workday.com/2025-10-01-Workday-Completes-Acquisition-of-Paradox), the completed acquisition of 1 Oct 2025 and how Workday sells the agent - Workday, [AI agents for Workday Recruiting datasheet](https://www.workday.com/content/dam/web/en-us/documents/datasheets/ai-agents-for-workday-recruiting-datasheet-enus.pdf), what the agent does; screening for basic qualifications; no price - Truffle, [Paradox AI pricing](https://hiretruffle.com/blog/paradox-ai-pricing), third-party estimate only, published Feb 2026, updated Oct 2026 and from a vendor of a competing screening product: starting price around $1,000 a month, no self-serve free trial, no published price - Tracy AI Interviewer, [Tracy pricing](https://tracyinterview.ai/pricing), free results, credit price, session-length billing, no-answer policy, no seats, credits never expire, API and MCP connections and customised setup for Enterprise - Tracy AI Interviewer, [Tracy product page](https://tracyinterview.ai/product), two-way video interview, CV check, follow-ups, coding task, recording and transcript, result with assessment per question, integrity signals - Tracy AI Interviewer, [Tracy hiring at volume](https://tracyinterview.ai/use-cases/volume-hiring), interview on click at any hour with no recruiter and no calendar, ready in under 10 seconds # Tracy vs Paradox: Chat Screening or a Real Interview? (Nov 2026) > Paradox screens and schedules hourly candidates by chat. Tracy runs a spoken interview with follow-ups. See which fits your roles, and how each prices. - URL: https://tracyinterview.ai/compare/tracy-vs-paradox - Published: 2026-11-05 - Updated: 2026-11-11 - Author: Tracy team Checked on 3 Nov 2026 against Paradox's, Workday's and Tracy's own pages. Where a Paradox page does not describe a capability, the page says "not described", not "cannot". > **In short** > - For skilled roles, pick **Tracy AI Interviewer**: a two-way video interview of the same depth for applicant one and applicant five hundred, with a coding task and a probe on the trade-offs for technical roles. Paradox wins hourly hiring: chat screening, scheduling and Q&A in 100+ languages. > - Tracy offers API and MCP connections and a customised setup for Enterprise. Paradox publishes no price (quote by demo, from Paradox or Workday); Tracy's first 3 results are free, then $1 per credit. ## Tracy vs Paradox at a glance | Buying question | Tracy AI Interviewer | Paradox | |---|---|---| | Main job | Run a spoken interview and return a result you can check against the recording | Chat and text assistant for applying, screening, scheduling and onboarding | | Format | Two-way video in the browser, no install for the person interviewed | SMS, WhatsApp and Facebook Messenger chat; hosts video prompts and human video interviews | | Volume | Every applicant is interviewed when they click the link, any hour, no recruiter or calendar | Chat screening of applicants at any hour, with scheduling to follow | | CV or resume | Optional CV check: 3 to 5 questions tailored to the CV and the role's must-haves | Not described as an interview step on pages read | | Follow-up questions | A vague answer gets a sharper question; waits for the end of a thought | Not described on pages read | | Technical tasks | Can introduce a coding task and probe the trade-offs | Not described | | Evidence | Recording, a transcript timed to it, a written assessment per question with its reasoning | Screens for basic qualifications, per Workday's datasheet | | Integrity checks | Tab changes and fullscreen, face and gaze, response timing and external screen, marked on the recording for review | Not described on pages read | | Scheduling | Not described on Tracy's pages; people join from a link when it suits them, so there is no scheduling back-and-forth | Self-scheduling, panel and multi-location interviews; 10-minute claim | | Languages | One interview language, chosen at setup | Answers candidate questions in more than 100 | | ATS and stack | API and MCP connections, three ways to start an interview, customised setup for Enterprise; the pages we checked describe no direct ATS write-back | Direct integrations, including Workday | | Price | 3 results free, then $1 per credit | No public price | ## Where does Paradox win? Paradox wins when the work is hourly or frontline volume, and when most of the job is getting people from applying to a booked slot. Its home page describes mobile-first automation for frontline workers and managers. Candidates apply and are screened by SMS, WhatsApp or Facebook Messenger, at any hour. It claims interviews scheduled within 10 minutes, a vendor claim, and its Candidate Experience Agent handles multi-person, multi-room, multi-location scheduling. Tracy does not describe a scheduling feature. It also wins on reach and on stack fit. Paradox says its assistant can answer candidate questions in more than 100 languages, while Tracy sets one interview language at setup, though the two are not the same task: one answers questions, the other conducts the interview. Workday completed its acquisition of Paradox on 1 October 2025 and sells the product as the Workday Paradox Candidate Experience Agent, so a Workday Recruiting shop can buy the conversation layer from its ATS vendor. Pick Paradox if: - you hire shift or hourly staff in volume and the bottleneck is applying, eligibility checks and booking interviews; - you run Workday Recruiting and want one vendor with a direct integration; - SMS or WhatsApp is your candidates' main channel, or you need candidate Q&A across many languages. ## Where does Tracy win? Tracy wins when every applicant for a skilled role should get a real interview, not a chat screen or a shortlist first. Each applicant starts when they click the link, at any hour, with no recruiter and no calendar, and the interview is ready in under 10 seconds, so a pool of hundreds is interviewed over days without anyone booking slots. Paradox's Candidate Experience Agent page describes Q&A and complex scheduling for skilled hiring, and Workday's datasheet says the agent screens "for basic qualifications". That fits an eligibility check better than a test of how someone reasons when you ask the second question. Tracy asks that second question. It waits for the end of the thought first, with longer pauses allowed on hard technical questions, and each follow-up depends on the answer before it. With CV check on, it also reads the CV and asks 3 to 5 questions tailored to it and to the job's must-have requirements, so two applicants for one role do not get the same script. In the invented example on Tracy's home page, a candidate says they chose a B-tree index, and Tracy asks how much faster reads were and what it cost on writes. It can introduce a coding task and probe the trade-offs behind an answer, so it fits engineering and other technical roles. You can edit the drafted questions or replace them with your own. Its video interviews also carry integrity signals that the home page lists (tab changes and fullscreen, face and gaze, a long delay before an answer, an external screen); they are marked on the recording and reach a reviewer as evidence. The person joins from a browser with nothing to install. The report is the other difference. Each person gets the recording, a transcript timed to it and a written assessment per question with its reasoning, so a reviewer reads the words and then decides. The Paradox pages we read describe screening results as qualified or not, and describe no report of this kind. Ask Paradox in a demo before treating that as a gap. How to read such a report is covered in [AI interview reports](/blog/ai-interview-reports), and the idea behind the format in [what an AI interviewer is](/blog/what-is-an-ai-interviewer). ## How does pricing compare? Paradox publishes no price. Its pricing address redirects to the home page and the call to action is a demo request. A third-party site reports a starting price around $1,000 a month; that is not a Paradox figure and we could not confirm it. The detail and the questions to ask are in [Paradox pricing](/compare/paradox-pricing). Tracy publishes its price and charges per result, with no seats. The first 3 results are free, then $1 per credit. One credit covers each started 20 minutes, so a result of up to 20 minutes costs $1. Credits do not expire and a session with no answers is never charged. The two products price different jobs, so compare the cost of the whole workflow you will run. ## Which should you choose? Choose Paradox when: - your roles are hourly or frontline and scheduling is the bottleneck; - you run Workday Recruiting and want a direct integration; - candidates are on SMS or WhatsApp and need answers in many languages. Choose Tracy AI Interviewer when: - the roles are skilled and every applicant, not a shortlist, should get an interview that tests how they handle a follow-up; - you want the CV read and questions tailored to it; - technical roles need a coding task with a probe on the trade-offs; - reviewers need the recording, transcript and integrity evidence behind each assessment; - you want to read a price and start with three free results. A company hiring both hourly and skilled staff does not have to pick one tool for both jobs. If your ATS already bundles an interviewer, see [the ATS interviewer vs a dedicated one](/blog/ats-ai-interviewer-vs-dedicated) and [Greenhouse's AI interviewer vs Tracy](/compare/greenhouse-ai-interviewer-vs-tracy). For the wider shortlist, read [Paradox alternatives for skilled roles](/compare/paradox-alternatives) and [AI screening interviews](/blog/ai-screening-interviews). Tracy does not make the hiring decision; it asks the questions and writes up what people said, and a person decides. Workday likewise says its agent lets recruiters spend more time on human-centered hiring decisions. ## Common questions ### What is the main difference between Tracy and Paradox? Paradox is a chat and text assistant that screens applicants for qualifications, answers their questions and schedules interviews, mostly for hourly and frontline hiring. Tracy AI Interviewer runs a spoken two-way video interview in which a vague answer gets a sharper question, and returns a result with the recording and transcript behind it. ### Is Paradox better for hourly hiring? For hourly and frontline volume, yes on the points that matter most: chat and SMS or WhatsApp screening, self-scheduling and answers to candidate questions in more than 100 languages. Tracy does not describe scheduling or text-message screening on its pages. ### Does Paradox run video interviews? Paradox's video page describes screening prompts, viewing candidate responses and hosting video interviews between people. The pages we read do not describe an AI-led video interview with follow-ups. ### Which one publishes its price? Tracy: the first 3 results are free, then $1 per credit for each started 20 minutes, with no seats. Paradox publishes no price; its pricing address redirects to its home page and you request a demo. ### Does Tracy work with Workday, which now owns Paradox? Paradox integrates directly with Workday. Tracy does not describe a Workday integration; it runs beside ATS tools such as Workable inside Claude or ChatGPT, and results stay in Tracy. ### Who makes the hiring decision? The person. Tracy does not reject anyone and a person decides, and Workday says its agent lets recruiters spend more time on human-centered hiring decisions. --- **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=tracy-vs-paradox&utm_content=demo)** or [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=tracy-vs-paradox&utm_content=try-free) ## Sources All sources retrieved 3 Nov 2026. - Paradox, [Paradox home page](https://www.paradox.ai/), text and chat screening, frontline positioning, demo request as the only call to action - Paradox, [Paradox FAQs](https://www.paradox.ai/faqs), hourly and skilled-role scope, 100+ languages, integrations - Paradox, [Candidate Experience Agent](https://www.paradox.ai/products/candidate-experience-agent), channels, scheduling and the skilled-hiring use - Paradox, [Conversational ATS](https://www.paradox.ai/products/conversational-ats), the 10-minute scheduling claim - Paradox, [Paradox screening](https://www.paradox.ai/products/screening), qualified candidates are scheduled and others dispositioned - Paradox, [Paradox video](https://www.paradox.ai/products/video), video prompts and hosted human video interviews - Paradox, [Paradox and Workday](https://www.paradox.ai/partners/workday), screening against job requirements - Workday, [Workday completes acquisition of Paradox](https://newsroom.workday.com/2025-10-01-Workday-Completes-Acquisition-of-Paradox), ownership, how it is sold, human decisions - Workday, [AI agents for Workday Recruiting datasheet](https://www.workday.com/content/dam/web/en-us/documents/datasheets/ai-agents-for-workday-recruiting-datasheet-enus.pdf), engages, screens and schedules; screening for basic qualifications - Tracy AI Interviewer, [Tracy home page](https://tracyinterview.ai/), the example interview, the integrity signals listed for video interviews, you read the words then you decide - Tracy AI Interviewer, [Tracy product page](https://tracyinterview.ai/product), two-way video interview, CV check, follow-ups, coding task, transcript timed to the recording, one language at setup, a person decides - Tracy AI Interviewer, [Tracy hiring at volume](https://tracyinterview.ai/use-cases/volume-hiring), interview on click at any hour with no recruiter and no calendar, a two-way video interview of the same depth for applicant one and applicant five hundred - Tracy AI Interviewer, [Tracy pricing](https://tracyinterview.ai/pricing), price, free results, no seats, API and MCP connections and customised setup for Enterprise - Tracy AI Interviewer, [Tracy and Workable](https://tracyinterview.ai/help/workable), how Tracy runs beside Workable's connector; no write-back # Alex (formerly Apriora) Alternatives (Nov 2026) > Six alternatives to Alex, formerly Apriora: Tracy AI Interviewer, HeyMilo, HireVue, Sapia.ai, Spark Hire and Ribbon, compared on format, follow-ups, evidence and published price. - URL: https://tracyinterview.ai/compare/alex-apriora-alternatives - Published: 2026-11-04 - Updated: 2026-11-11 - Author: Tracy team > **In short** > - If Alex's funnel is wider than your hiring needs, start with **Tracy AI Interviewer**: it interviews everyone who applies on click, can read the CV to tailor its questions, and hands the reviewer a recording, timed transcript and written assessment per question. > - If you want a bigger suite, look at Ribbon or HeyMilo. For enterprise governance, HireVue. For text on any phone, Sapia.ai. For a small team on a published price, Spark Hire. Fact-checked on 3 Nov 2026 against the vendors' own pages ([see Sources](#sources)). Alex is the new name of Apriora: Apriora Inc. does business as Alex, and apriora.ai redirects to alex.com. This page lists six tools to weigh against it, with what each does well and when to pick it instead. ## Why do teams look beyond Alex? We did not cite buyer reviews of Alex (its platform page claims a G2 ranking, which we could not check), so these reasons come from Alex's own pages and from what they do not say. - Price. The pricing page is a form for a 30-minute demo, and we found no per-interview price, plan tiers or free trial. Alex's own blog says annual contracts start at $36K (stated for staffing firms in its 19 Feb and 2 Mar 2026 posts). See [Alex (formerly Apriora) pricing](/compare/alex-apriora-pricing). - A wider tool than some teams want. Alex describes itself as an AI recruiter for outreach, screening, interviewing, verification and scheduling. A team that only needs the interview is buying a suite. - Evidence wording. Alex's pages describe transcripts, role-fit scores, fraud flags and an audit trail. Its Proctor flags open to timestamps and the transcript, but we found no statement that a role-fit score links to the passage behind it. - Coding. Alex lists coding: its YC launch post names coding among its interview types, and its blog says it assesses coding knowledge through conversational dialogue (16 Nov 2025). Tracy can introduce a coding task and probe the trade-offs. Alex is strong where these do not matter: four channels (video, phone, SMS, WhatsApp), 33+ ATS integrations, an ID check and SOC 2 Type II plus ISO 27001. [Tracy vs Alex (formerly Apriora)](/compare/tracy-vs-alex) covers where it wins. ## The six, with who each is for | Tool | Best for | |---|---| | Tracy AI Interviewer | Every applicant interviewed in depth, with CV-tailored questions, follow-ups and evidence for a reviewer | | HeyMilo | High-volume screening across voice, video, SMS and forms, written back to the ATS | | HireVue | Enterprise teams that need governance and one suite from engagement to interview | | Sapia.ai | Large employers screening by text chat on any phone | | Spark Hire | Small teams that want recorded video interviews on a published price | | Ribbon | Teams that want sourcing, outreach and a first interview in one tool | ## 1. Tracy AI Interviewer Tracy runs a two-way video interview in the browser. The people you interview install nothing, and every applicant is interviewed on their own when they click the link, at any hour, with no scheduling. With CV check on, Tracy adds 3 to 5 questions from each CV and the role's must-haves. Each follow-up depends on the answer before it, and a vague answer gets a sharper question. Its home page says it can introduce a coding task and probe the trade-offs behind an answer, and you can edit the drafted questions or bring your own. Each result carries the recording, a transcript timed to it, a written assessment per question with its reasoning and integrity signals (tab changes, gaze, response timing, an external screen) with their evidence. Interviews can start from an invitation link, an apply page or a customer's own system through the API, the MCP server connects Tracy to chat tools such as Claude and ChatGPT, and enterprise teams can ask for a customised setup. Price: The first 3 results are free. Then $1 per credit, with 1 credit for each started 20 minutes. You pay for results, not seats, and a session with no answers is never charged. Pick Tracy if you want every applicant interviewed in depth, the depth of the second question matters more than the number of channels, or your reviewers need a recording, transcript and written reasoning to check each result. You can also start without a sales call. **Pick something else if** you need phone, SMS or WhatsApp interviews, scheduling, sourcing or results written back to your ATS; the pages we checked describe none of these for Tracy. ## 2. HeyMilo HeyMilo screens and interviews applicants over voice, video, SMS and forms, in 20+ languages. It asks follow-ups in real time, and the recruiter sets the maximum number of follow-ups per question. Reports carry per-question scores with the rationale and candidate quotes with timestamps. It lists 20+ ATS partners. HeyMilo's product pages describe no coding interview. Price: No public price. Its pricing page redirects to a book-a-demo form, and its billing docs say plan options and pricing appear in the product once billing is set up for a workspace. Pick HeyMilo if you hire at volume and want SMS, forms and a phone or video screen in one flow with ATS sync. See [Tracy vs HeyMilo](/compare/tracy-vs-heymilo) and [HeyMilo alternatives](/compare/heymilo-alternatives). ## 3. HireVue HireVue offers on-demand video and live interviews, plus an AI interviewer whose agents ask and record responses, including probing follow-up questions, in a two-way conversation. The AI evaluates each response against role-specific criteria and explains why. HireVue also lists more than 200 coding skills assessments. Its tagline for the AI is "AI Informs. Humans decide." Price: No price on its pricing page, which points to a plan conversation. Pick HireVue if you are an enterprise or mid-market team that wants audits, governance and live, on-demand and AI interviews in one suite. See [Tracy vs HireVue](/compare/tracy-vs-hirevue) and [HireVue alternatives](/compare/hirevue-alternatives). ## 4. Sapia.ai Sapia.ai interviews by text chat only: its page says "No video, no voice, no face." It says every score is traceable to a candidate's own words, and says it is independently audited for bias with the results published. It says it has assessed 10M+ candidates and has built AI interviews since 2018. Those figures are Sapia's. Its pages do not state follow-up questions or technical tasks. Price: None published. Pricing is per hire at a tailored rate, and Sapia says its technology is designed for companies that hire more than 500 people per year. Pick Sapia.ai if you hire frontline or professional volume, want a text-based screen that works on any phone and have the hiring volume its model assumes. ## 5. Spark Hire Spark Hire is built on one-way recorded video and live video interviews. Every candidate answers the same questions, and responses are AI-summarized for reviewers. It lists no AI follow-up questions, and it says it is not designed for teams that want to fully automate screening or remove human judgment. Price: from $249 per month billed annually for unlimited live and one-way video interviews, jobs and users. Pricing varies by company size. Pick Spark Hire if you are a small team that wants to replace phone screens with recorded video, at a price on the page, with a person reviewing every answer. Read [one-way vs two-way video interviews](/blog/one-way-vs-two-way-video-interviews) for what a recording cannot ask. ## 6. Ribbon Ribbon runs structured voice and video interviews, and describes real two-way conversations that adapt in real time. Its scorecard sits beside the recording and transcript, so recruiters can jump to the answer behind a recommendation. It also covers AI sourcing, talent rediscovery in the existing ATS and outreach by SMS, WhatsApp and email. Price: Ribbon's pricing page returned a 404 on 3 Nov 2026. Its blog, dated 23 Aug 2026, says annual-billing packages start at $499 per month for 100 interviews, and its integrations page mentions a 7-day free trial. Confirm both before relying on them. Also: on 23 Oct 2026 Dayforce announced a signed agreement to acquire Ribbon. The release refers to what follows the close of the transaction. Pick Ribbon if you want sourcing, outreach and a first interview in one tool. See [Tracy vs Ribbon](/compare/tracy-vs-ribbon) and [Ribbon alternatives](/compare/ribbon-alternatives). ## Side by side | Tool | Interview format | Follow-ups | Result evidence | Published price | |---|---|---|---|---| | Tracy AI Interviewer | Two-way video, every applicant on click, any hour | A vague answer gets a sharper question; CV check adds questions from each CV | Recording, timestamped transcript, written assessment per question, integrity evidence | First 3 results free, then $1 per credit | | Alex (for reference) | Video, phone, SMS, WhatsApp | Guided by an Interview Guide | Transcripts, role-fit scores, flags with timestamps | Pricing page: none; blog: annual contracts from $36K for staffing firms | | HeyMilo | Voice, video, SMS, forms | Real-time, capped by the recruiter | Scores with rationale and timestamped quotes | None | | HireVue | On-demand and live video, AI interviewer | Probing follow-ups | Criteria-based evaluation with explanations | None | | Sapia.ai | Text chat | Not stated | Says every score traces to the candidate's words | None; per hire | | Spark Hire | One-way recorded and live video | None listed; same questions for everyone | AI summaries per response | From $249 a month, annual | | Ribbon | Voice and video | Adapts in real time | Scorecard linked to the answer | $499 a month for 100 interviews, annual billing, per an Aug 2026 blog | ## How to choose Start from what you are leaving Alex for. If it is price, shortlist the tools with a number on the page: Tracy and Spark Hire publish one, and Ribbon has one in a blog. If it is breadth, Ribbon and HeyMilo are suites. If it is governance, HireVue. If it is depth in a technical or expert interview, run the same brief through Tracy and one other tool and compare the second question each asks. [AI interview cost per candidate](/blog/ai-interview-cost) puts published prices on one scale, and [the best AI interview software for hiring teams](/blog/best-ai-interview-software) ranks the wider market. ## Common questions ### What is Alex, formerly Apriora? Alex is an AI recruiter from Apriora Inc., which does business as Alex. It covers outreach, screening, interviewing, verification and scheduling. apriora.ai redirects to alex.com. ### What is the best alternative to Alex? It depends on what you are leaving. For a two-way interview at a published price, Tracy AI Interviewer. For a wider recruiting suite, Ribbon or HeyMilo. For enterprise governance, HireVue. For text screening, Sapia.ai. For recorded video on a small budget, Spark Hire. ### Which Alex alternatives publish a price? Tracy publishes $1 per credit with the first 3 results free, and Spark Hire publishes plans from $249 per month billed annually. Ribbon's only price is in an August 2026 blog post, and Alex's only price is in its own blog posts from February and March 2026 ($36K a year, stated for staffing firms in two posts). HeyMilo, HireVue and Sapia.ai publish none. ### Do any of these alternatives run coding interviews? Tracy's home page says it can introduce a coding task and probe trade-offs, and HireVue lists coding skills assessments. Alex lists coding in its launch post and its blog says it assesses coding knowledge through conversational dialogue. The other pages we read do not describe coding tasks. ### Is Ribbon being acquired? Dayforce announced on 23 Oct 2026 that it had signed an agreement to acquire Ribbon. The release refers to what follows the close of the transaction, so it was not final when the release was published. --- **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=alex-apriora-alternatives&utm_content=demo)** to run one of your first-round briefs through Tracy, or [try Tracy AI Interviewer free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=alex-apriora-alternatives&utm_content=try-free) with your first 3 results on us. ## Sources All sources retrieved 3 Nov 2026. - Alex, [platform](https://www.alex.com/platform): positioning, 33+ ATS integrations, SOC 2 Type II plus ISO 27001, channels. - Alex, [pricing](https://www.alex.com/pricing): redirects to a 30-minute demo form. - Alex, [blog: JobDiva vs Alex](https://www.alex.com/blog/jobdiva-vs-alex): 19 Feb 2026, pricing starts at $36K annually for staffing firms; the same figure appears in Alex posts dated 17 Feb and 2 Mar 2026. - Alex, [blog: HireVue vs Alex](https://www.alex.com/blog/hirevue-vs-alex): 16 Nov 2025, coding knowledge assessed through conversational dialogue. - Alex, [AI Interviewer](https://www.alex.com/product/ai-interviewer): channels, transcripts, scores and audit trail. - Alex, [Identity](https://www.alex.com/product/identity): government ID plus live selfie. - Y Combinator, [Alex (fka Apriora) launch post](https://www.ycombinator.com/launches/KY6-alex-fka-apriora-your-ai-interviewer-to-identify-the-best-candidates-faster): former name and interview types including coding. The launch post is dated March 2024. - [apriora.ai](https://apriora.ai): redirects to alex.com. - Tracy AI Interviewer, [home page](https://tracyinterview.ai/): follow-ups that depend on the answer before, coding task with a probe on the trade-offs, drafted questions you can edit or replace, integrity signals, API connection. - Tracy AI Interviewer, [product page](https://tracyinterview.ai/product): two-way video interview, no install, CV check, three ways in including the API, recording with a timed transcript, written assessment per question. - Tracy AI Interviewer, [pricing](https://tracyinterview.ai/pricing): free results, credit price, billing, no seats, no-answer policy, enterprise customised setup and API and MCP connections. - HeyMilo, [home](https://www.heymilo.ai): channels and languages. - HeyMilo, [interviewing candidates](https://www.heymilo.ai/interviewing-candidates): real-time follow-ups. - HeyMilo, [responsible AI](https://www.heymilo.ai/responsible-ai): quotes with timestamps. - HeyMilo, [integrations](https://www.heymilo.ai/integrations): 20+ ATS partners. - HeyMilo, [pricing](https://www.heymilo.ai/pricing): redirects to the book-a-demo form. - HeyMilo, [interview types](https://docs.heymilo.ai/guides/build-your-interviewer/interview-types.md): interview types listed in its docs. - HeyMilo, [billing docs](https://docs.heymilo.ai/deep-dive/settings-and-admin/billing.md): plan options and pricing appear once billing is set up for a workspace. - HeyMilo, [interviewer settings](https://docs.heymilo.ai/guides/build-your-interviewer/configuring-your-ai-interviewer.md): maximum follow-ups. - HireVue, [AI Interviewer](https://www.hirevue.com/platform/ai-interviewer): follow-ups, criteria-based evaluation, "AI Informs. Humans decide." - HireVue, [video interviewing](https://www.hirevue.com/platform/video-interviewing): on-demand and live video. - HireVue, [coding assessments](https://www.hirevue.com/products/coding-assessments): 200+ coding skills assessments. - HireVue, [pricing](https://www.hirevue.com/pricing): no price shown. - Sapia.ai, [chat interview](https://sapia.ai/platform/interview/): text only, score traceability, bias audit. - Sapia.ai, [home](https://sapia.ai/): 10M+ candidates assessed, since 2018. - Sapia.ai, [pricing](https://sapia.ai/pricing/): per hire, 500+ hires a year. - Spark Hire, [one-way video interview](https://www.sparkhire.com/one-way-video-interview/): same questions, AI summaries, not for full automation. - Spark Hire, [pricing](https://www.sparkhire.com/pricing/?solution=video-interviews): from $249 a month billed annually. - Ribbon, [home](https://www.ribbon.ai/): two-way conversations that adapt in real time. - Ribbon, [AI Interviewer](https://www.ribbon.ai/ai-interviewer): voice and video, scorecard beside recording and transcript. - Ribbon, [Ribbon vs BrightHire](https://www.ribbon.ai/blog/ribbon-vs-brighthire): $499 a month for 100 interviews, 23 Aug 2026. - Ribbon, [integrations](https://www.ribbon.ai/integrations): 7-day free trial. - Ribbon pricing page (www.ribbon.ai/pricing, not linked because it does not open): returned HTTP 404 on 3 Nov 2026. - Dayforce, [announcement](https://www.dayforce.com/who-we-are/newsroom/dayforce-to-acquire-ribbon-ai-inc-to-advance-agentic-recruiting): signed agreement to acquire Ribbon, 23 Oct 2026; integration follows the close of the transaction. - Tracy AI Interviewer, [Connect page](https://tracyinterview.ai/connect): MCP and API connections. # Alex (formerly Apriora) Pricing (Nov 2026) > Alex, formerly Apriora: its pricing page is a demo form, but its blog says annual contracts start at $36K. What that means, what to ask in a quote, and how Tracy's per-result price compares. - URL: https://tracyinterview.ai/compare/alex-apriora-pricing - Published: 2026-11-04 - Updated: 2026-11-11 - Author: Tracy team > **In short** > - Pick **Tracy AI Interviewer** if you want a video interview for every applicant without booking a call: any hour, no scheduling, questions that can be built from the CV, follow-ups on each answer, integrity signals for a reviewer, and API and MCP connections. A written assessment per question comes with its reasoning, and the price is set out in the Tracy section below. > - Alex's pricing page is a demo form, but its own blog says annual contracts start at $36K. Below: what that figure does and does not tell you, the questions that turn a quote into a cost per interview, and what Tracy costs at 100, 500 and 1,000 results. Fact-checked on 3 Nov 2026 against the vendors' own pages ([see Sources](#sources)). Apriora Inc. does business as Alex, and apriora.ai now redirects to alex.com. If you searched for Apriora pricing, Alex is the product, and this page is about what it costs. ## What does Alex cost? Alex's pricing page does not say. It goes to a booking form that reads "Schedule a 30-minute demo with an AI expert." We found no per-interview price, no plan tiers and no free trial on that page on 3 Nov 2026. Alex's blog does give a starting figure. A post dated 19 Feb 2026 says "Alex pricing starts at $36K annually for staffing firms, with custom enterprise pricing based on volume, integrations, and deployment complexity." A post dated 17 Feb 2026 says "Annual agreements starting at $36K based on interview volume and integration requirements," and a post dated 2 Mar 2026 says "Annual contracts from $36K (staffing); custom enterprise." All three are Alex's own blog, not its pricing page, and the figure is for staffing firms. The 17 Feb post names no buyer type, and we found no other figure. The ROI calculator page adds a framing: Alex is "priced against headcount, not software." The calculator takes your annual hires, your recruiters and your cost per hire. Alex describes itself as handling outreach, screening, interviewing, verification and scheduling, which suggests a quote framed against recruiter time rather than an interview count. What is public: | Pricing item | What Alex states | |---|---| | Price on the pricing page | None; demo form only | | Starting price in its blog (Feb to Mar 2026) | Annual contracts from $36K for staffing firms | | Enterprise price | Custom, based on volume, integrations and deployment complexity (blog) | | Plans or tiers | None published | | Free trial | Not stated; a 30-minute demo is offered | | How the price is framed | Against headcount, not software | ## What decides the quote? Alex sells a bundle, so the quote depends on which parts you buy. Its pages describe interviews over video, phone, SMS and WhatsApp, a Coordinator for scheduling, Talent Match for past applicants in your ATS, and three fraud layers called Verify, Proctor and Identity. It also lists 33+ ATS integrations. Alex's blog says custom pricing depends on "volume, integrations, and deployment complexity." We could not find whether any of the parts above is priced separately. That is the reason to ask the same questions of every vendor. Use this list in your demo call and get the answers in writing: | Ask | Why it changes the cost | |---|---| | What is the unit: completed interview, seat, recruiter, requisition, hire or a flat fee? | Units do not convert into each other without your volumes. | | Does an interview count when the candidate starts and leaves, or never starts? | Tracy never charges a session with no answers. Ask whether Alex does. | | Are phone, SMS and WhatsApp in the base price or charged per message or minute? | Channels can change what each interview costs. | | Are Verify, Proctor, Identity, Coordinator and Talent Match included or extra? | The fraud and scheduling layers may be separate line items. | | What are the contract length, minimum volume and overage terms? | Annual commitments and overage change the real price per interview. | | Is there a trial or a paid pilot? | You can test the report before you commit. | Convert whatever comes back into one number: the cost of a completed interview that a reviewer can use. That is the figure to compare across vendors. ## What does Tracy cost? A credit pays for one video interview slot of up to 20 minutes that ends in a result. Applicants start when they click the link, at any hour, so more applications do not mean more calendar work. With CV check on, Tracy adds 3 to 5 questions from each CV and the role's must-haves, and each follow-up depends on the answer before it. The result holds the recording, a transcript timed to it, a written assessment for each question with its reasoning, and integrity signals with their evidence. Interviews start from an invitation link, an apply page or a customer's own system, and Tracy offers API and MCP connections. Tracy publishes its price. The first 3 results are free. After that, a result costs 1 credit for every started 20 minutes of the session, at $1 per credit. Credits never expire and there are no seats. A session with no answers is never charged. The table assumes every session produces a result, all sessions run the same length, and the 3 free results are used once. | Results | Up to 20 min (1 credit) | 21 to 40 min (2 credits) | |---|---|---| | 100 | $97 | $194 | | 500 | $497 | $994 | | 1,000 | $997 | $1,994 | Each figure is the number of results after the first 3, times the credits per result, times $1. For 100 results of up to 20 minutes that is 97 × $1 = $97. ## Is Alex worth the quote? That depends on how much of the recruiting workflow you want one vendor to run. Alex is the stronger purchase when you need WhatsApp, SMS or phone interviews, government ID checks, scheduling and ATS rediscovery together, because Tracy's pages describe none of them for the interview itself. A quote for the whole bundle can be fair against the recruiter time it replaces, and only a quote can show you. Tracy is the better fit when the interview is the job: every applicant interviewed in depth, questions tailored to the CV, follow-ups, and evidence a reviewer can check, at a price you can check before you talk to anyone. Its pages describe no phone, SMS or WhatsApp channel, ID check or calendar step. If you need the surrounding workflow, price it separately or choose Alex. The two prices do not share a unit, so we do not chart them. Alex's $36K is an annual contract that covers more than the interview. Tracy's $1 is one credit, and one credit buys up to 20 minutes of a session that produces a result. As arithmetic only, $36,000 divided by $1 is 36,000 credits, which means 36,000 paid sessions of up to 20 minutes, not counting the 3 free results. That says nothing about what Alex's contract includes. When you have a quote, divide it by the completed interviews it covers and set it against $1 per credit for the session length you expect. See [Tracy vs Alex (formerly Apriora)](/compare/tracy-vs-alex) for the full feature comparison, [AI interview cost per candidate](/blog/ai-interview-cost) for published prices across the category, and [Alex (formerly Apriora) alternatives](/compare/alex-apriora-alternatives) if you want more options with a stated price. Other vendors that keep prices off the page are covered in [HireVue pricing](/compare/hirevue-pricing) and [HeyMilo pricing](/compare/heymilo-pricing). ## Common questions ### Does Alex (formerly Apriora) publish its pricing? Not on its pricing page, which leads to a form for a 30-minute demo. Its blog says annual contracts start at $36K, for staffing firms in the 19 Feb and 2 Mar 2026 posts, with custom enterprise pricing; a 17 Feb post names no buyer type. We found no plan tiers or free trial on 3 Nov 2026. ### Is Apriora the same company as Alex? Yes. Apriora Inc. does business as Alex, and apriora.ai redirects to alex.com. ### How does Alex frame its price? Its ROI calculator page says Alex is "priced against headcount, not software." The calculator asks for your annual hires, your recruiters and your cost per hire. ### What does Tracy cost compared with Alex? Tracy's first 3 results are free, then $1 per credit, with 1 credit for each started 20 minutes of a session that produces a result. Alex's blog gives an annual contract from $36K, a different unit, so a direct comparison needs a quote. ### What should I ask Alex in a quote? Ask for the unit, what the $36K blog figure covers and whether it applies to you, what counts as a billable interview, whether phone, SMS and WhatsApp and the fraud and scheduling layers are included, the contract length and overage terms, and whether there is a trial or paid pilot. --- **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=alex-apriora-pricing&utm_content=demo)** to price one of your hiring volumes with Tracy, or [try Tracy AI Interviewer free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=alex-apriora-pricing&utm_content=try-free) and run your first 3 results at no cost. ## Sources All sources retrieved 3 Nov 2026. - Alex, [pricing](https://www.alex.com/pricing): redirects to a 30-minute demo form; no price shown. - Alex, [blog: JobDiva vs Alex](https://www.alex.com/blog/jobdiva-vs-alex): 19 Feb 2026, pricing starts at $36K annually for staffing firms. - Alex, [blog: best enterprise recruitment software](https://www.alex.com/blog/best-enterprise-recruitment-software): 17 Feb 2026, annual agreements starting at $36K. - Alex, [blog: Mya chatbot alternatives](https://www.alex.com/blog/mya-chatbot-alternatives-a124d): 2 Mar 2026, annual contracts from $36K for staffing. - Alex, [ROI calculator](https://www.alex.com/roi-calculator): "Priced against headcount, not software." - Alex, [platform](https://www.alex.com/platform): scope, 33+ ATS integrations and channels. - Alex, [AI Interviewer](https://www.alex.com/product/ai-interviewer): video, phone, SMS and WhatsApp interviews. - Alex, [Coordinator](https://www.alex.com/product/coordinator): scheduling. - Alex, [Talent Match](https://www.alex.com/product/talent-match): rediscovery of past applicants in the ATS. - Alex, [Proctor](https://www.alex.com/product/proctor): assistance checks during the interview. - Alex, [Identity](https://www.alex.com/product/identity): government ID plus live selfie. - Alex, [Verify](https://www.alex.com/product/verify): application fraud-risk band. - Alex, [demo page](https://www.alex.com/demo): "Schedule a 30-minute demo with an AI expert." - [apriora.ai](https://apriora.ai): redirects to alex.com. - Tracy AI Interviewer, [pricing](https://tracyinterview.ai/pricing): free results, credit price, billing by started 20 minutes, no seats, no-answer policy, enterprise API and MCP connections. - Tracy AI Interviewer, [product page](https://tracyinterview.ai/product): video interview, CV check, three ways in including the API, recording with a timed transcript, written assessment per question. - Tracy AI Interviewer, [Connect page](https://tracyinterview.ai/connect): MCP and API connections. # HeyMilo Alternatives (Nov 2026) > Six HeyMilo alternatives compared on interview format, follow-ups, public price and who each suits: Tracy, Alex, HireVue, Sapia.ai, Spark Hire and Ribbon. - URL: https://tracyinterview.ai/compare/heymilo-alternatives - Published: 2026-11-04 - Updated: 2026-11-11 - Author: Tracy team Facts checked on 3 Nov 2026 against each vendor's own pages ([see Sources](#sources)). We did not use review sites for this list, so the reasons for leaving HeyMilo come from its own pages, not from buyer reviews. > **In short** > - **Tracy AI Interviewer** is the pick if HeyMilo's multi-step funnel is more than you need: it interviews everyone who applies on video, answers a vague reply with a sharper question and hands a reviewer the recording and integrity signals to check. > - Stay with **HeyMilo** when you want a wide funnel with SMS, forms, scheduling and ATS write-back, all of which it lists. Look at **Alex** for multi-channel interviews (video, phone, SMS, WhatsApp) and calendar coordination, or **Ribbon** for sourcing and outreach around the interview. > - **HireVue** suits enterprise governance, **Sapia.ai** text-first volume hiring and **Spark Hire** a low-cost recorded video screen. HeyMilo is a capable screening funnel, and many teams are right to stay on it. This page lists six alternatives, each judged on the same questions: how the interview runs, whether it follows up, what it costs in public, and when you should pick it instead. ## Why do teams look beyond HeyMilo? From HeyMilo's own pages, three reasons stand out. ### No public price HeyMilo's pricing page for its screening and interview platform leads to a demo form, and its documentation says plan options and prices appear once billing is set up for your workspace. Its separate proctoring product, HeyMilo Red, does publish prices (Free $0, Growth $59 a month, priced by minutes of recorded video analysed), but that is not the price of running an interview. Teams that need a number to put in a budget have to book a call first. Our [HeyMilo pricing page](/compare/heymilo-pricing) goes through what is and is not public. ### Follow-ups bound by a rubric HeyMilo lets the recruiter set how many follow-ups the AI can ask, and in its comparison with Alex it writes that follow-up questions can only come from the scoring rubric. That keeps screening consistent. It also means a follow-up that is not in the rubric does not get asked, which matters when a technical answer needs a question nobody wrote in advance. ### A different shape of tool HeyMilo is built around a high-volume funnel across voice, video, SMS and forms. Teams that want one deep interview, a text-only screen or a recorded video are shopping for something else. Working inside the ATS is a HeyMilo strength, not a reason to leave: its own pages list an embedded widget in Bullhorn, Greenhouse, Lever and Ashby and 20+ ATS partners. ## Quick list: which alternative for which team? | Alternative | Best for | |---|---| | Tracy AI Interviewer | Teams that want every applicant interviewed in depth, with CV-tailored questions, follow-ups and evidence for a reviewer | | Alex (formerly Apriora) | A full recruiting workflow with video, phone, SMS and WhatsApp interviews and identity checks | | HireVue | Enterprise talent teams that need governance, audit trails and one suite | | Sapia.ai | Large employers hiring 500+ people a year who want a text-based screen | | Spark Hire | Small and mid-size teams that want recorded video interviews at a published price | | Ribbon | Teams that want sourcing, outreach and a first-round interview in one tool | ## 1. Tracy AI Interviewer Tracy runs a two-way video interview from a link, with no install for the person being interviewed. Every applicant is interviewed on their own when they click the link, at any hour and with no scheduling, so hundreds of applicants for one role need no extra recruiter time. With CV check on, Tracy adds 3 to 5 questions from each CV and the role's must-haves. Each follow-up depends on the answer before it, a vague answer gets a sharper question, and for technical roles Tracy can introduce a coding task and probe the trade-offs behind an answer. You can edit the drafted questions or bring your own. Each result carries a recording, a transcript with timestamps, a written assessment for each question with its reasoning, and integrity signals (tab changes, gaze, response timing, an external screen) with their evidence, for a person to review. Interviews can start from an invitation link, an apply page or a customer's own system through the API, the MCP server connects Tracy to chat tools such as Claude and ChatGPT, and enterprise teams can ask for a customised setup. Price: $1 per credit, one credit for each started 20 minutes of a session that produces a result. The first 3 results are free, there are no seats, and a session with no answers is never charged. Pick it if you want every applicant interviewed in depth, the roles are skilled or technical and the follow-up question is where you learn the most, or you want to budget before a sales call. Look elsewhere if you need SMS, phone or form steps, calendar scheduling or results written back into an ATS; the pages we checked describe none of these for Tracy. For the full comparison, read [Tracy vs HeyMilo](/compare/tracy-vs-heymilo). ## 2. Alex (formerly Apriora) Alex covers a wider recruiting workflow. Apriora Inc. does business as Alex, and apriora.ai now redirects to alex.com. Alex runs two-way interviews over video, phone, SMS and WhatsApp, and asks follow-ups based on what the candidate says, guided by an Interview Guide. A coordinator books qualified candidates on team calendars. Its fraud layers cover application risk, assistance during the interview, and identity through a government ID plus a live selfie. Its platform page lists 33+ supported ATS integrations and SOC 2 Type II plus ISO 27001. Price: the pricing page leads to a 30-minute demo form. Alex's own blog says its pricing starts at $36K annually for staffing firms (19 Feb and 2 Mar 2026 posts; a 17 Feb post gives the same $36K start without naming a buyer type), a vendor statement. Pick it if candidates need WhatsApp or SMS interviews, you hire across many languages, or identity proof is a requirement. Our [Tracy vs Alex comparison](/compare/tracy-vs-alex) and [Alex alternatives](/compare/alex-apriora-alternatives) pages go further. ## 3. HireVue HireVue is strongest on governance at enterprise scale. It offers on-demand and live video interviews and an AI interviewer that asks probing follow-up questions in a two-way conversation. It says every score ties back to evidence from the interview and your competency rubric. It also lists more than 200 on-demand coding skills assessments, and states an annual SOC 2 Type 2 audit and adverse impact testing before models go live. Price: no published price. The pricing page points to a conversation about plans. Pick it if you are an enterprise talent team that needs audit trails, validated science and one suite from engagement to interview. See [Tracy vs HireVue](/compare/tracy-vs-hirevue), [HireVue alternatives](/compare/hirevue-alternatives) and [HireVue pricing](/compare/hirevue-pricing) for the detail. ## 4. Sapia.ai Sapia.ai is built around text. It describes a structured, text-based AI interview that every applicant can finish on their phone, in their own time, with no video, no voice and no face. It says the interview is untimed, that every score is traceable to the candidate's own words, and that its scoring is independently audited for bias and published. It states SOC 2 Type 2. Price: none published. Sapia.ai says customers pay only for the people they hire, at a tailored cost-per-hire rate, and that its technology is designed for companies that hire more than 500 people a year. Pick it if you hire frontline or professional roles at volume, want applicants to answer by text, and prefer to pay per hire. Its pages we checked do not state whether it asks follow-up questions or runs technical tasks, so test both. ## 5. Spark Hire Spark Hire sells recorded video at a price you can read. Its video interviewing plan starts from $249 a month billed annually, with unlimited live and one-way video interviews, jobs and users; month-to-month starts at $299 with up to 5 jobs and 5 users; the page says pricing varies by company size. Plans that add an ATS, Recruit Pro and Recruit Growth, start from $335 and $499 a month billed annually, and Recruit plans are annual contracts only. Candidates record answers on their own schedule, responses are transcribed in 50+ languages, and reviewers get an AI summary of each response. Spark Hire lists 40+ ATS integrations and a SOC 2 Type II audit on request. Pick it if you are a lean HR team that wants to replace phone screens with recorded video and keep people deciding. Spark Hire says that every candidate answers the same questions under the same conditions, so in its one-way format it is a fixed set of questions with no AI follow-up. It says it is not for teams that want to fully automate screening. ## 6. Ribbon Ribbon covers the work around the interview. It lists AI sourcing, talent rediscovery in your ATS, outreach by SMS, WhatsApp and email, and structured voice and video interviews in 10+ languages that follow up when an answer needs more detail. It lists 60+ integrations and a 7-day free trial, and says a first interview can be live within minutes. Price: Ribbon's pricing page returned a 404 when we checked. Its own blog, dated 23 Aug 2026, says public annual-billing packages start at $499 a month for 100 interviews, which we could not confirm on a live pricing page. Note: on 23 October 2026 Dayforce announced a signed agreement to acquire Ribbon. The deal had not closed in that release. Pick it if sourcing and outreach belong in the same purchase as the interview. See [Tracy vs Ribbon](/compare/tracy-vs-ribbon) and [Ribbon alternatives](/compare/ribbon-alternatives). ## Comparison table | Tool | Interview format | Follow-up questions | Public price | Free start | |---|---|---|---|---| | Tracy AI Interviewer | Two-way video, every applicant on click, any hour | Each follow-up depends on the answer before it; CV check adds questions from each CV | $1 per credit | First 3 results free | | Alex | Video, phone, SMS, WhatsApp | Based on what the candidate says, guided by an Interview Guide | None; demo form | Not stated | | HireVue | On-demand and live video, AI interviewer | Probing follow-ups in the AI interviewer | None | Not stated | | Sapia.ai | Text chat | Not stated | None; per hire | Not stated | | Spark Hire | One-way and live video | One-way: same questions for every candidate, no AI follow-up | From $249 a month billed annually; $299 month-to-month | Not stated | | Ribbon | Voice and video | Follows up when an answer needs more detail | From $499 a month for 100 interviews, billed annually (blog, not a live pricing page) | 7-day trial | | HeyMilo (for reference) | Video, phone, SMS, form | Real-time, set by the recruiter within a rubric | None for the platform; demo form (HeyMilo Red, a separate proctoring product, is priced from $0) | Not stated for the platform; Red has a free tier and a 14-day trial | ## How do you choose? Start from the job. - **If you want every applicant interviewed in depth and the second question matters,** put Tracy first and test it on a real brief. Check the next question the interviewer asks after a vague answer. - **If you screen high volume across several channels,** compare Alex, Ribbon and HeyMilo on ATS write-back, languages and scheduling, and ask each for a written quote. - **If procurement needs governance depth,** start with HireVue and Alex, and ask every vendor for its security documents. - **If candidates are on phones with weak connections,** Sapia.ai's text format removes the camera. Alex's phone and WhatsApp interviews do too. - **If budget is the constraint and you only need recorded answers,** Spark Hire's published price is the easiest to plan around. Whatever you shortlist, run the same role through two tools. Include a plausible answer that leaves out a constraint, a candidate who needs another format and a result you disagree with. For more tools in the category, see our [best AI interview software](/blog/best-ai-interview-software) list, and for per-candidate costs, [AI interview cost per candidate](/blog/ai-interview-cost). ## Common questions ### What is the best HeyMilo alternative? It depends on the job. Tracy AI Interviewer suits skilled and technical roles that need a deeper video interview at a published price. Alex and Ribbon are closest if you want a wide funnel with several channels. HireVue suits enterprise governance, Sapia.ai text-first volume hiring and Spark Hire a recorded video screen. ### Is there a HeyMilo alternative with public pricing? Yes. Tracy publishes $1 per credit with the first 3 results free, and Spark Hire publishes plans from $249 a month billed annually for video interviews. Ribbon's own blog says public annual-billing packages start at $499 a month for 100 interviews, but its pricing page was unavailable when we checked. HeyMilo's screening and interview platform, HireVue and Sapia.ai publish no price, and Alex's pricing page is a demo form, though its own blog says pricing starts at $36K annually for staffing firms. HeyMilo's separate Red proctoring product does (Free $0, Growth $59 a month), but it scans recordings and does not run interviews. ### Which HeyMilo alternative works for technical interviews? Tracy can introduce a coding task and probe the trade-offs behind an answer. HireVue lists more than 200 on-demand coding skills assessments. HeyMilo's pages do not describe a coding interview, so ask if you need one. ### Which alternatives interview by text or WhatsApp? Sapia.ai runs a text-only interview. Alex runs interviews over SMS and WhatsApp as well as video and phone, and Ribbon lists outreach by SMS, WhatsApp and email. ### Why do teams switch from HeyMilo? From HeyMilo's own pages, the likely reasons are the lack of a public price for the platform, follow-ups bound to the recruiter's scoring rubric, and a funnel built for high volume when a team wants one deeper interview. HeyMilo is still the better pick for hourly and staffing volume with ATS write-back. --- **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=heymilo-alternatives&utm_content=demo) and run your own role through Tracy, or [try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=heymilo-alternatives&utm_content=try-free) with your first 3 results.** ## Sources All sources retrieved 3 Nov 2026. - HeyMilo, [pricing page (leads to a demo form)](https://www.heymilo.ai/pricing): no public price for the screening and interview platform. - HeyMilo, [billing documentation](https://docs.heymilo.ai/deep-dive/settings-and-admin/billing): pricing shown after billing set-up. - HeyMilo, [configuring your AI interviewer](https://docs.heymilo.ai/guides/build-your-interviewer/configuring-your-ai-interviewer.md): recruiter sets the maximum follow-ups. - HeyMilo, [compare with Alex](https://www.heymilo.ai/compare/alex-ai): HeyMilo's statement that follow-ups come from the scoring rubric, and its embedded widget in Bullhorn, Greenhouse, Lever and Ashby. - HeyMilo Red, [pricing](https://red.heymilo.ai/pricing): Free $0, Growth $59 a month, $0.04 per minute after the allowance. - HeyMilo, [HeyMilo Red launch post](https://blog.heymilo.ai/heymilo-red-interview-proctoring-api), published 24 Sep 2026: Red's prices and 14-day free trial. - HeyMilo, [home page](https://www.heymilo.ai): positioning across voice, video, SMS and forms. - HeyMilo, [interview types](https://docs.heymilo.ai/guides/build-your-interviewer/interview-types.md): the interview types listed in its documentation. - HeyMilo, [self-serve onboarding documentation](https://docs.heymilo.ai/guides/get-started/self-serve-onboarding.md): onboarding key from the HeyMilo team, no free trial stated. - Tracy AI Interviewer, [home page](https://tracyinterview.ai/): drafted questions you can edit or replace, follow-ups that depend on the answer before, coding task with a probe on the trade-offs, integrity signals with their evidence, "you read the words, then you decide". - Tracy AI Interviewer, [pricing](https://tracyinterview.ai/pricing): credit price, billing unit, free results, no seats, no-answer policy, enterprise customised setup and API and MCP connections. - Alex, [pricing](https://www.alex.com/pricing): demo form. Alex, [JobDiva vs Alex](https://www.alex.com/blog/jobdiva-vs-alex), 19 Feb 2026: pricing starts at $36K annually for staffing firms. Alex, [blog: best enterprise recruitment software](https://www.alex.com/blog/best-enterprise-recruitment-software), 17 Feb 2026: annual agreements starting at $36K. Alex, [blog: Mya chatbot alternatives](https://www.alex.com/blog/mya-chatbot-alternatives-a124d), 2 Mar 2026: annual contracts from $36K (staffing). - Alex, [AI interviewer](https://www.alex.com/product/ai-interviewer): video, phone and chat interviews, follow-ups guided by the Interview Guide. - Alex, [platform](https://www.alex.com/platform): 33+ ATS integrations, WhatsApp interviews. - Alex, [Proctor](https://www.alex.com/product/proctor): in-interview monitoring. - Alex, [Verify](https://www.alex.com/product/verify): application fraud risk band. - Alex, [Identity](https://www.alex.com/product/identity): government ID plus live selfie. - Alex, [Coordinator](https://www.alex.com/product/coordinator): calendar coordination. - HireVue, [AI interviewer](https://www.hirevue.com/platform/ai-interviewer): probing follow-ups, evidence tied to the rubric, governance and testing statements. - HireVue, [enterprise security and compliance](https://www.hirevue.com/platform/enterprise-security-compliance): annual SOC 2 Type 2 audit. - HireVue, [pricing](https://www.hirevue.com/pricing): no published price, mid-market to enterprise. - HireVue, [coding assessments](https://www.hirevue.com/products/coding-assessments): more than 200 on-demand coding skills assessments. - Sapia.ai, [interview](https://sapia.ai/platform/interview/): text-only interview, untimed, scores traceable to the candidate's words, bias audit. - Sapia.ai, [home page](https://sapia.ai/): SOC 2 Type 2. - Sapia.ai, [pricing](https://sapia.ai/pricing/): per-hire pricing, designed for companies that hire more than 500 people a year. - Spark Hire, [video interviewing pricing](https://www.sparkhire.com/pricing/?solution=video-interviews): from $249 a month billed annually, unlimited jobs and users; month-to-month from $299 with up to 5 jobs and 5 users, 40+ ATS integrations, SOC 2 Type II on request. - Spark Hire, [pricing](https://www.sparkhire.com/pricing/): Recruit plan prices and annual contracts. - Spark Hire, [one-way video interview](https://www.sparkhire.com/one-way-video-interview/): recorded format, same questions, transcription languages, human decision stance. - Ribbon, [home page](https://www.ribbon.ai/): outreach, integrations, 7-day trial and a first interview within minutes. - Ribbon, [AI interviewer](https://www.ribbon.ai/ai-interviewer): voice and video interviews, custom questions and rubrics, support for 10+ languages. - Ribbon, [comparison with BrightHire](https://www.ribbon.ai/blog/ribbon-vs-brighthire), published 23 Aug 2026: advertised $499 a month for 100 interviews. - Dayforce, [announcement of the agreement to acquire Ribbon](https://www.dayforce.com/who-we-are/newsroom/dayforce-to-acquire-ribbon-ai-inc-to-advance-agentic-recruiting), 23 Oct 2026: signed agreement, not closed. - Tracy AI Interviewer, [product page](https://tracyinterview.ai/product): two-way video interview, CV check, three ways in including the API, recording with a transcript timed to it, written assessment per question. - Tracy AI Interviewer, [Connect page](https://tracyinterview.ai/connect): MCP and API connections. # HeyMilo Pricing (Nov 2026) > HeyMilo quotes its interview platform by demo and prices its separate Red proctoring product in public. See what is public, what to ask for in a quote and how Tracy AI Interviewer prices per result. - URL: https://tracyinterview.ai/compare/heymilo-pricing - Published: 2026-11-04 - Updated: 2026-11-11 - Author: Tracy team Facts checked on 3 Nov 2026 against HeyMilo's own pages and documentation ([see Sources](#sources)). We found no price for HeyMilo's screening and interview platform on them, and we have not used third-party estimates. > **In short** > - **Tracy AI Interviewer** interviews everyone who applies, at any hour and with no scheduling. It can read the CV and ask tailored questions, follows up on each answer, gives a reviewer integrity signals with their evidence, and connects through an API and MCP. Its price is published in the pricing section below. > - **HeyMilo** does not publish a price for its screening and interview platform. Its pricing page leads to a demo form, and its documentation says plan options and prices appear only after billing is set up for your workspace. > - What is public is the shape of the bill (a subscription plan plus credits for interviews, resume screens and SMS) and the prices of a separate product, HeyMilo Red: Free $0, Growth $59 a month. If you searched for HeyMilo pricing, the short answer is that its interview platform has no number to find, while its separate proctoring product has one. This page sets out what HeyMilo does say about how it charges, how to get a quote that you can compare, and how Tracy AI Interviewer prices the same job. ## What does HeyMilo cost? HeyMilo does not state a plan price or a per-interview rate for its screening and interview platform on the pages we checked. Its pricing address redirects to a book-a-demo form. The form asks for your team size, your ATS, hires per month and applicants per month, and says the team will reach out within 24 hours. The documentation gives the structure. Billing is a subscription plan plus credits, and the docs say "Credits pay for interviews, resume screens, and SMS." Plan options and pricing appear on the billing page once billing is set up for your workspace. The docs also say a new workspace needs an onboarding key sent by the HeyMilo team, and we found no free trial stated for the platform. A plan name appears only as an example in the billing docs ("Professional"). HeyMilo Red is the exception, and it is a separate product. Red scans recorded interviews for integrity findings, including ones HeyMilo did not run, and charges by the minute analysed. On red.heymilo.ai/pricing the Free plan is $0 (about 300 minutes of video analysis a month, card required to start) and Growth is $59 a month (about 3,000 minutes, then $0.04 per minute). HeyMilo's launch post for Red also lists a 14-day free trial and custom Enterprise rates. Those figures price integrity scans, not an interview, so they cannot stand in for a quote for the platform. A review site may show a price in your own search. We did not use one, because we could not tie any figure to HeyMilo's own pages. ## What is public, and what is not? | Pricing item | What HeyMilo publishes | |---|---| | Platform plan names and prices | None published; pricing page leads to a demo form (the billing docs name "Professional" only as an example) | | Billing model | Subscription plan plus credits | | What credits pay for | Interviews, resume screens and SMS | | Price per interview or per credit | Not published | | Free trial | Not stated for the platform | | Seats | Not stated | | Enterprise pricing | Through your HeyMilo account manager; no public figure | | Enterprise services | A dedicated customer success manager and forward deployed engineer; price not public | | HeyMilo Red (separate proctoring product) | Free $0; Growth $59 a month, then $0.04 per minute; Enterprise custom; 14-day free trial named in its launch post | ## What affects what you pay? From the structure above, a quote will depend on three things you control. The first is volume: how many applicants you interview in a month. The second is the mix of steps, because resume screens and SMS draw on the same credits as interviews. The third is the support you buy, since enterprise services are listed as a separate feature and may be priced separately, so ask. HeyMilo's own wording suggests who the price is aimed at. Its funding announcement names staffing, BPO and high-volume corporate hiring as the places it is in production, and the demo form asks how many applicants you see each month. Teams in that position are buying a funnel. ## How do you get a quote you can compare? A quote is only useful if it answers the questions that decide your real cost. Bring these to the demo: 1. What does one interview cost at our usual length, in credits and in dollars? 2. What is the subscription minimum, and what does it include? 3. What does a resume screen or an SMS cost against the same credits? 4. What happens past the credit allowance: overage, top-up or a plan change? 5. Is there a contract term, and can we start with a pilot? 6. Are enterprise services, ATS integration work and training priced separately? 7. Is a session that ends with no answers charged? Ask for the answers in writing, with the date, because the amounts are not on a public page that could be checked later. Our guide to [AI interview cost per candidate](/blog/ai-interview-cost) shows how to turn a quote into a cost per candidate so two vendors can sit in one table. ## How does Tracy price the same job? Before the rate, here is what a credit buys. Each applicant is interviewed on video when they click the link, at any hour, so a rise in applications needs no extra scheduling. With CV check on, Tracy adds 3 to 5 questions from each CV and the role's must-haves, and each follow-up depends on the answer before it. The result carries the recording, a transcript timed to it, a written assessment for each question with its reasoning, and integrity signals with their evidence. Interviews start from an invitation link, an apply page or a customer's own system through the API, and the MCP server lets teams run Tracy from Claude or ChatGPT. Tracy AI Interviewer prices per result and publishes the rate. A result costs one credit for every started 20 minutes of the session, and a credit is $1. Credits never expire. The first 3 results are free and no card is needed to start. There are no seats. | Session length | Credits | Price | |---|---|---| | Up to 20 minutes | 1 | $1 | | 21 to 40 minutes | 2 | $2 | | 41 to 60 minutes | 3 | $3 | A session with no answers is never charged. Here is the arithmetic for 100 interviews of up to 20 minutes. The first 3 are free, so you pay for 97 credits at $1, which is $97. If every session ran 21 to 40 minutes, each paid result costs two credits: 194 credits, or $194. We cannot run this sum for HeyMilo because it has no published rate, and the two products are not priced on the same unit. HeyMilo's credits cover several kinds of step, while a Tracy credit covers spoken interview time. ## Is HeyMilo worth it, and for whom? HeyMilo is worth a demo if you screen hourly, staffing or other high-volume roles and want resume scoring, SMS, forms, phone and video in one flow, written back to an ATS. It lists 20+ ATS partners, a managed enterprise rollout and an integrity layer that includes a trust score, and a quote that covers all of that may be good value for a team with thousands of applicants. Our [Tracy vs HeyMilo comparison](/compare/tracy-vs-heymilo) sets out where it is stronger. If you also want to see how other vendors answer the pricing question, our [Ribbon pricing page](/compare/ribbon-pricing) covers what another AI interviewer publishes. Tracy suits teams that want every applicant interviewed in depth, whether that is hundreds of applicants for one role or a dozen specialists. Questions follow the CV and the answers, reviewers get a recording, transcript and integrity signals to check, and enterprise teams can ask for a customised setup. It is also easy to test, because the first 3 results cost nothing and you do not need a call. If neither fits, the [HeyMilo alternatives](/compare/heymilo-alternatives) page covers other options, and our [best AI interview software](/blog/best-ai-interview-software) list ranks the field. ## Common questions ### How much does HeyMilo cost? HeyMilo does not publish a price for its screening and interview platform. Its pricing page leads to a demo form, and its documentation says plan options and pricing appear once billing is set up for your workspace. Billing is a subscription plan plus credits. Its separate proctoring product, HeyMilo Red, publishes Free at $0 and Growth at $59 a month, priced by minutes analysed. ### Does HeyMilo have a free trial? We found no free trial stated for HeyMilo's screening and interview platform. Its documentation says a new workspace needs an onboarding key sent by the HeyMilo team. Its separate Red proctoring product has a free tier and, per HeyMilo's launch post, a 14-day free trial of its Growth plan. Tracy AI Interviewer gives the first 3 results free with no card needed. ### What do HeyMilo credits pay for? According to HeyMilo's documentation, credits pay for interviews, resume screens and SMS. The price per credit is not published. ### Does HeyMilo charge per seat? HeyMilo does not state whether it charges per seat. Ask in your quote. Tracy has no seats. ### How does Tracy's price compare with HeyMilo's? They cannot be compared in dollars yet, because HeyMilo has no public rate for its platform. HeyMilo Red's per-minute prices are for scanning recordings, not for running interviews. Tracy costs $1 per credit, one credit for each started 20 minutes, with the first 3 results free. Ask HeyMilo for a written quote at your length and volume, then convert both to a cost per interview. --- **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=heymilo-pricing&utm_content=demo) to see Tracy's price on your own volume, or [try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=heymilo-pricing&utm_content=try-free) with your first 3 results.** ## Sources All sources retrieved 3 Nov 2026. - HeyMilo, [pricing page (leads to a demo form)](https://www.heymilo.ai/pricing): no public price, the fields the demo form asks for, the 24-hour reply. - HeyMilo, [billing documentation](https://docs.heymilo.ai/deep-dive/settings-and-admin/billing): pricing shown after billing set-up, "Professional" as an example plan name, credits for interviews, resume screens and SMS, enterprise pricing through the account manager. - HeyMilo, [self-serve onboarding documentation](https://docs.heymilo.ai/guides/get-started/self-serve-onboarding.md): onboarding key from the HeyMilo team. - HeyMilo, [enterprise services](https://www.heymilo.ai/product-feature/enterprise-services): customer success manager and forward deployed engineer. - HeyMilo, [funding announcement](https://blog.heymilo.ai/heymilo-raises-6-million), published 16 Jun 2026: where HeyMilo says it is in production. - HeyMilo, [integrations](https://www.heymilo.ai/integrations): 20+ ATS partners. - HeyMilo, [cheating detection](https://www.heymilo.ai/product-feature/cheating-detection): trust score. - HeyMilo, [home page](https://www.heymilo.ai): resume scoring, voice, video, SMS and form steps. - HeyMilo, [HeyMilo Red launch post](https://blog.heymilo.ai/heymilo-red-interview-proctoring-api), published 24 Sep 2026: Red's Free and Growth prices, 14-day free trial, what Red scans. - HeyMilo Red, [pricing](https://red.heymilo.ai/pricing): Free $0, Growth $59 a month, $0.04 per minute after the allowance, card required to start. - Tracy AI Interviewer, [pricing](https://tracyinterview.ai/pricing): credit price, billing unit, free results, no-answer policy, no seats, enterprise API and MCP connections and customised setup. - Tracy AI Interviewer, [home page](https://tracyinterview.ai/): no card needed to start, the result with its reasoning and integrity signals. - Tracy AI Interviewer, [product page](https://tracyinterview.ai/product): CV check, three ways in including the API, recording and timed transcript, written assessment per question. - Tracy AI Interviewer, [Connect page](https://tracyinterview.ai/connect): MCP and API connections. # Tracy vs Alex (formerly Apriora) (Nov 2026) > Tracy AI Interviewer vs Alex, the AI recruiter formerly called Apriora: channels, follow-ups, evidence in reports, integrity checks and price, with where each one wins. - URL: https://tracyinterview.ai/compare/tracy-vs-alex - Published: 2026-11-04 - Updated: 2026-11-11 - Author: Tracy team > **In short** > - Pick **Tracy AI Interviewer** when you want one focused job done well: a video interview for each applicant that starts from their CV, with a coding task you can add, and a recording, timed transcript and per-question assessment a reviewer can check. Alex's pages do not say its questions change with the resume. > - Pick **Alex** when you want one vendor for outreach, interviewing, identity checks and scheduling, or when candidates need phone, SMS or WhatsApp. Fact-checked on 3 Nov 2026 against the vendors' own pages ([see Sources](#sources)). You may have found Alex under its old name. Apriora Inc. does business as Alex, and apriora.ai now redirects to alex.com. This page compares Tracy with Alex as it is described on alex.com today. Apriora in a search result and Alex on the site are the same product. ## Tracy vs Alex at a glance | Buying question | Tracy AI Interviewer | Alex (formerly Apriora) | |---|---|---| | What it is | A two-way spoken interview and a written-up result | An "AI recruiter": outreach, screening, interviewing, verification and scheduling | | Interview channels | Video, in the browser | Video, phone, SMS and WhatsApp | | Volume and scheduling | Every applicant is interviewed on their own when they click the link, at any hour, with no calendar | AI interviews run around the clock, and Coordinator books qualified candidates onto team calendars afterwards | | CV or resume | Optional CV check: 3 to 5 questions tailored to the CV and the role's must-haves | Resume Screens listed as a separate product; its platform page calls the resume "a commodity signal" and puts the weight on the interview | | Follow-up questions | Listens for specifics and asks for the missing detail; waits longer on hard technical questions | Follow-ups based on what the candidate says, guided by an Interview Guide | | Coding | Home page says it can introduce a coding task and probe trade-offs | Lists coding: YC launch post names coding among its interview types; blog (Nov 2025) says it assesses coding knowledge through conversation | | Integrity checks | Face, gaze, response timing and browser signals (tab changes, leaving fullscreen, an external screen) flagged with evidence for a reviewer, not a verdict | Verify, Proctor and Identity (government ID plus live selfie) | | What the reviewer gets | The recording, a transcript with timestamps, a written assessment per question with its reasoning and the integrity evidence | Searchable transcripts, role-fit scores, fraud flags and an audit trail | | ATS and stack | API and MCP connections, three ways to start an interview (invitation link, apply page, a customer's own system) and a Zapier trigger that passes a result link on; no direct ATS write-back described. Customised versions for enterprise | 33+ officially supported ATS integrations on its product pages | | Languages | One interview language, chosen at setup (product page) | 26+ in the text of its platform page, 30 in a counter on the same page, 30+ in its llms.txt file | | Price | On its pricing page: first 3 results free, then $1 per credit | Pricing page is a 30-minute demo form; its blog (Feb to Mar 2026) says annual contracts start at $36K for staffing firms | ## Where does Alex win? Alex is the wider product, and for several buyers that settles it. Channels and languages. Alex runs interviews over video, phone, SMS and WhatsApp. Its phone page says "No webcams, apps, or internet required." If your candidates are on shifts, in markets where WhatsApp is the default, or hiring across many languages, Tracy's browser video interview will not reach them the way Alex does. Alex states 26+ to 30+ languages across its own pages (26+ in the text of its platform page, a counter of 30 on the same page, 30+ in its llms.txt); Tracy runs each interview in one language, chosen at setup. Proof that the person is the person. Alex's Identity layer asks for a government-issued ID and a live selfie during the interview. Its Verify layer scores each application for fraud risk and puts it in a low, medium or high band. Tracy's pages describe neither an ID check nor an application fraud score. Its signals start when the interview does. The work around the interview. Alex's Coordinator books qualified candidates onto team calendars right after the AI screen, and Talent Match re-contacts past applicants who are already in your ATS. Tracy's pages describe the interview, the invitations and the result. They do not describe calendar scheduling or sourcing. ATS depth and procurement evidence. Alex lists 33+ official ATS integrations, with Workday, iCIMS, Greenhouse, Lever and SmartRecruiters among them. Its platform page states SOC 2 Type II plus ISO 27001, ISO 42001, and a SOC 2 report available under NDA. With Workable, Tracy runs beside Workable's own connector inside Claude or ChatGPT and keeps results in Tracy, so a team that wants results written back to the candidate record should look at Alex first. Technical interviews by conversation. Alex's blog, in a post dated 16 Nov 2025, says it "conducts technical question interviews with contextual follow-ups that assess coding knowledge through conversational dialogue," covering implementation approaches, edge cases and system design. If you want coding knowledge assessed by talking it through, Alex describes that directly. Track record. Alex raised $20M, including a $17M Series A led by Peak XV Partners announced on 29 September 2025, and its llms.txt file says 1,000,000+ candidates interviewed. Tracy's home page says 100K+ interviews completed. Both are company figures. ## Where does Tracy win? Every applicant interviewed, whenever they click. Each Tracy interview is its own session, started when the candidate clicks the link and ready in under 10 seconds, with no recruiter and no calendar in the loop. A candidate who drops can rejoin within 10 minutes, and a reminder goes out after a day for an unfinished interview. Alex also interviews around the clock, so on volume alone the two are level; what differs is what each interview holds and where it runs. An interview built from the CV and the role. With CV check on, Tracy adds 3 to 5 questions from each CV and the job's must-haves, on top of the questions it generates from the job description, which you can edit or replace with your own. Two applicants for one role do not get the same script. Alex describes its Interview Guide and a separate Resume Screens product; the pages we checked do not say that the interview questions change with each candidate's resume, so ask. Evidence behind each result. Tracy records both sides and anchors a transcript with timestamps to the recording, then writes an assessment for every question with its reasoning, so a reviewer can check each point against what was said. As the Tracy home page puts it, you read the words, then you decide. Alex's pages describe transcripts, role-fit scores, fraud flags and an audit trail for every interview. Its Proctor page says a flag opens to the timestamps and the transcript behind it, but we found no statement that a role-fit score links back to a passage. If that link matters to your reviewers, ask Alex to show you one finished report and trace a score to the answer behind it. Run the same test on Tracy. See [what reviewers should see in an AI interview report](/blog/ai-interview-reports). Follow-ups in the person's own field. Both tools ask follow-ups. Tracy's follow-up depends on the answer before it, so a vague answer gets a sharper question, and it waits for the end of a thought (up to about 2 seconds on hard technical questions) rather than cutting the person off, and its home page says it can introduce a coding task and probe the trade-offs behind an answer. Alex describes follow-ups guided by its Interview Guide, and its blog says it assesses coding knowledge through conversational dialogue. Both list coding: Alex says so in its launch post and blog, and Tracy can introduce a coding task and probe the trade-offs. For engineering or expert screens, give both the same brief and compare the second question. The [guide to AI technical interviews](/blog/ai-technical-interviews) has a task you can use. Fits the stack. Tracy connects through an API and an MCP server, so a customer's own system can request an interview and hand the candidate a link, and enterprise teams can ask for a customised setup. Alex's 33+ official ATS integrations go deeper on write-back, as above. Starting without a call. Tracy needs no card to start. The people you interview install nothing. Alex's route in is a 30-minute demo, and its pages state no free trial. Integrity checks are not a Tracy win. Tracy captures signals during the interview, assesses them together afterwards and gives the evidence to the reviewer beside the assessment. Tracy's home page lists tab changes and fullscreen, face and gaze, and timing and screens, such as a long delay before an answer or an external screen. They are signals, not a verdict, and they point to moments worth a look in the recording. Alex's Proctor also reads signals such as a second voice, off-screen glances and tab switching, rates the likelihood of assistance, and Identity adds an ID check. Alex's flags do not decide the outcome either; the team configures how a flagged interview is handled. Our [guide to AI interview cheating detection](/blog/ai-interview-cheating-detection) covers what to ask any vendor. ## How does pricing compare? Tracy: the first 3 results are free. After that a result costs 1 credit for every started 20 minutes, at $1 per credit. Credits never expire. Alex: the pricing page leads to a form for a 30-minute demo, and the ROI page says Alex is "priced against headcount, not software." Its blog is where a number appears: a post dated 19 Feb 2026 says "Alex pricing starts at $36K annually for staffing firms, with custom enterprise pricing based on volume, integrations, and deployment complexity." We found no per-interview price, plan tiers or free trial published. The $36K is a yearly contract and Tracy's $1 is one credit, so the two do not line up until you have a quote. [Alex (formerly Apriora) pricing](/compare/alex-apriora-pricing) lists what to ask for in a quote, and [AI interview cost per candidate](/blog/ai-interview-cost) puts published prices across the category on one scale. ## Which should your team choose? Choose Alex when: - candidates need WhatsApp, SMS or phone interviews, or you hire across many languages; - you require government ID and live selfie checks, or an application fraud-risk score; - you want outreach, interviewing and scheduling from one vendor; - procurement needs SOC 2 Type II, ISO 27001 and ISO 42001 evidence. Choose Tracy AI Interviewer when: - you want every applicant interviewed in depth, at any hour, with no scheduling; - the questions should follow each candidate's CV and answers, including a coding task; - your reviewers want a recording, transcript, written reasoning per question and integrity signals to check; - you want to try an interview this week without a sales call, paying per result. If it is close, run one real first-round brief through both and read the two reports side by side. Weighing more than two options? [Alex (formerly Apriora) alternatives](/compare/alex-apriora-alternatives) sets Tracy beside five others. For the wider market, see [the best AI interview software for hiring teams](/blog/best-ai-interview-software). ## Common questions ### Are Alex and Apriora the same product? Yes. The legal entity is Apriora Inc., which does business as Alex. The product and brand are now Alex, and apriora.ai redirects to alex.com. ### What is the main difference between Tracy and Alex? Alex describes itself as an AI recruiter that handles outreach, screening, interviewing, verification and scheduling. Tracy AI Interviewer does one job: it runs a two-way video interview and writes an assessment for each question with its reasoning, next to the recording and transcript. ### Does Alex publish its pricing? Not on its pricing page, which leads to a form for a 30-minute demo. Its blog says annual contracts start at $36K, for staffing firms in the 19 Feb and 2 Mar 2026 posts; a 17 Feb post names no buyer type. We found no plan tiers or free trial. Tracy publishes its price on its pricing page: the first 3 results are free, then $1 per credit. The units differ, so a like-for-like comparison needs an Alex quote. ### Which has stronger cheating and fraud checks? Alex lists three layers: Verify for application fraud risk, Proctor for help during the interview, and Identity, which checks a government ID against a live selfie. Tracy shows signals on video interviews as moments worth a look in the recording, not a verdict, and its pages describe no ID check. If identity proof is a requirement, Alex is the stronger fit. ### Can either tool make the hiring decision? No. Tracy says it does not reject anyone and a person decides. Alex says no candidate is auto-rejected and hiring teams can override its outputs. --- **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=tracy-vs-alex&utm_content=demo)** to run one of your first-round briefs through Tracy, or [try Tracy AI Interviewer free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=tracy-vs-alex&utm_content=try-free) with your first 3 results on us. ## Sources All sources retrieved 3 Nov 2026. - Tracy AI Interviewer, [home page](https://tracyinterview.ai/): follow-ups that depend on the answer before, coding task, integrity signals list, "you read the words, then you decide", the 100K+ interviews figure, API connection. - Tracy AI Interviewer, [pricing](https://tracyinterview.ai/pricing): free results, credit price, billing by started 20 minutes, no-answer policy, no seats, enterprise customised setup and API and MCP connections. - Tracy AI Interviewer, [product page](https://tracyinterview.ai/product): two-way video interview, one language chosen at setup, CV check, rejoin within 10 minutes, reminder after a day, waiting for the end of a thought, written assessment per question, integrity signals, "Tracy does not reject anyone. A person decides." - Tracy AI Interviewer, [Workable guide](https://tracyinterview.ai/help/workable): chat-based ATS connector, results stay in Tracy. - Tracy AI Interviewer, [Zapier guide](https://tracyinterview.ai/help/zapier): a Result Ready trigger carries the link to the result. - Alex, [platform](https://www.alex.com/platform): positioning, ATS count, language counts (26+ in text, 30 in a counter), SOC 2, ISO 27001 and ISO 42001 wording, human-decision statement. - Alex, [AI Interviewer](https://www.alex.com/product/ai-interviewer): channels, follow-ups, transcripts, scores and audit trail. - Alex, [Proctor](https://www.alex.com/product/proctor): signals, likelihood rating, flags open to timestamps and transcript, flags do not decide the outcome. - Alex, [Identity](https://www.alex.com/product/identity): government ID plus live selfie. - Alex, [llms.txt](https://www.alex.com/llms.txt): 30+ languages and 1,000,000+ candidates interviewed, both company figures. - Alex, [Verify](https://www.alex.com/product/verify): application fraud-risk band. - Alex, [Coordinator](https://www.alex.com/product/coordinator): scheduling after the AI screen. - Alex, [pricing](https://www.alex.com/pricing): redirects to a 30-minute demo form. - Alex, [blog: JobDiva vs Alex](https://www.alex.com/blog/jobdiva-vs-alex): 19 Feb 2026, pricing starts at $36K annually for staffing firms; the same figure appears in Alex posts dated 17 Feb and 2 Mar 2026. - Alex, [blog: HireVue vs Alex](https://www.alex.com/blog/hirevue-vs-alex): 16 Nov 2025, technical question interviews that assess coding knowledge through conversational dialogue. - Alex, [phone interviews](https://www.alex.com/product/phone): no webcams, apps or internet required. - Alex, [ROI calculator](https://www.alex.com/roi-calculator): "Priced against headcount, not software." - Alex, [We raised $20M](https://www.alex.com/blog/we-raised-20m-to-help-ai-hire-more-humans): funding, published 29 September 2025. - Y Combinator, [Alex (fka Apriora) launch post](https://www.ycombinator.com/launches/KY6-alex-fka-apriora-your-ai-interviewer-to-identify-the-best-candidates-faster): the former name and the interview types listed, including coding. The launch post is dated March 2024. - [apriora.ai](https://apriora.ai): redirects to alex.com. - Tracy AI Interviewer, [Connect page](https://tracyinterview.ai/connect): MCP and API connections. # Tracy vs HeyMilo: AI Interviews Compared (Nov 2026) > Tracy vs HeyMilo compared on interview depth, channels, evidence, ATS write-back and price. See which AI interviewer fits your hiring. - URL: https://tracyinterview.ai/compare/tracy-vs-heymilo - Published: 2026-11-04 - Updated: 2026-11-11 - Author: Tracy team Facts checked on 3 Nov 2026 against each vendor's own pages ([see Sources](#sources)). HeyMilo publishes no price for its screening and interview platform, so the pricing section compares what each company says it charges for, not dollar totals. HeyMilo does publish prices for a separate product, HeyMilo Red, and the pricing section says what those cover. > **In short** > - Pick **Tracy AI Interviewer** when you want to interview everyone who applies, not only a shortlist. Each applicant gets their own two-way video or audio interview when they click the link, at any hour, with no scheduling. Tracy can read the CV, ask questions tailored to it and to the role, follow up on what the person says, and flag integrity signals for a reviewer. > - Pick **HeyMilo** when you screen hourly or staffing volume and want resume, SMS, form, phone and video steps in one flow, written back to your ATS. > - Both keep a recruiter out of the first round and give reviewers evidence from the candidate's own words. Tracy's difference is depth for each applicant. HeyMilo's is breadth of channels and ATS write-back. HeyMilo and Tracy both run interviews that candidates take on their own time, and both hand a recruiter something to check against the candidate's own words. They are built for different jobs. HeyMilo describes itself as an AI recruiter for high-volume hiring that screens and interviews every applicant over voice, video, SMS and forms. Tracy AI Interviewer is an interviewer: each applicant gets a full spoken interview on video or audio, built from the job and, when you turn it on, from the candidate's CV. ## Tracy vs HeyMilo at a glance | Buying question | Tracy AI Interviewer | HeyMilo | |---|---|---| | Main job | A full two-way interview for each applicant, built from the role and the CV, with a recording, transcript and written assessment | Screening and interviewing every applicant across voice, video, SMS and forms | | Interview format | Video or audio, chosen per interview, in the browser | Video, phone call or browser audio, SMS and web form, plus resume analysis | | CV or resume | Optional CV check: 3 to 5 questions tailored to the CV and the role's must-haves | Resume analysis as a step in the funnel | | Follow-up questions | Listens for specifics and asks for the missing detail | Asks follow-ups in real time; the recruiter sets the maximum per question | | Technical tasks | Can introduce a coding task and probe the trade-offs | Interview types listed in its docs include an AI scenario assessment; a coding interview is not described | | Evidence in the report | A written assessment per question with its reasoning, the transcript with timestamps and the recording | Per-question score with rationale, timestamped quotes, transcript and recording | | Integrity checks | Face, gaze, response timing and browser signals (tab changes, leaving fullscreen, an external screen) flagged with evidence for review, not a verdict | Trust score, AI-answer and impersonation checks, plus the HeyMilo Red proctoring API | | ATS | API and MCP connections to the tools you already use; no direct ATS write-back described | 20+ ATS partners with scores, reports and transcripts written back | | Languages | Nine | 20+ | | Published price | $1 per credit, first 3 results free | None for the platform, quoted through a demo form; HeyMilo Red is priced separately (Free $0, Growth $59 a month) | ## Who is each one for? HeyMilo suits recruiting teams with a large applicant pool for roles where the first job is to sort people quickly: staffing, BPO and high-volume corporate hiring, in its own words. It also suits teams that want the screen to sit inside the ATS they already run. Tracy suits a team that wants every applicant interviewed in depth, whether that is a dozen specialists or hundreds of applicants for one role: skilled, technical and expert roles where the second question carries the signal, and volume roles where a form or a chat would miss it. It also lets a team start without a sales call, since the first three results are free. ## Where does HeyMilo win? These are the areas where HeyMilo is ahead, by its own pages. ### A wider screening funnel HeyMilo runs resume scoring, SMS, forms, phone and video agents in one posting. A recruiter can send a browser link for an audio interview or let the AI call candidates directly, and candidates who pass the screen can pick a time on the recruiter's Google, Outlook or Apple calendar. Tracy is not a phone line or a text chatbot. It can read the candidate's CV to tailor its questions, but the pages we checked describe no SMS, form or calendar step. ### ATS write-back HeyMilo's integrations page lists 20+ ATS partners, including Workday, Greenhouse, iCIMS, Lever, Bullhorn, Avionte and JobDiva (its compare index page says 15+ native integrations). Structured scores, PDF reports, transcripts and recordings sync back to the candidate record. Tracy runs from a chat in Claude or ChatGPT beside an ATS connector such as Workable's; Workable shares each candidate's name and email, and the results stay in Tracy. The pages we checked describe no direct ATS write-back for Tracy; it connects through its API and MCP server. ### A deeper integrity layer HeyMilo says every interview produces a trust score built from voice authentication, proctoring signals and AI-generated answer detection. Video adds candidate presence, identity and session trust integrity checks. Its docs say flags are for review and do not auto-reject candidates or change scores. On 24 September 2026 it also launched HeyMilo Red, a standalone proctoring API for interviews recorded on other platforms. ### An assessment of working with AI In June 2026 HeyMilo added an AI scenario assessment: a timed task with a built-in AI assistant, scored on prompting, judgment and deliverable quality. If you hire roles that work alongside an AI assistant, that is a different signal from an interview, and the pages we checked describe no equivalent for Tracy. ### More control over scoring, and a managed rollout The recruiter writes what a score of 1 and a score of 5 look like for each question, weights the questions and can override any AI score. HeyMilo also sells Enterprise Services, with a dedicated customer success manager and a forward deployed engineer. Its pages state SOC 2 Type I and Type II and monthly bias audits by Warden AI. For security and procurement review, ask Tracy for its documentation during a demo and put the answers beside HeyMilo's. ### Reach HeyMilo runs interviews in 20+ languages. Tracy runs in nine: English, German, Spanish, French, Indonesian, Japanese, Korean, Dutch and Vietnamese. HeyMilo also says more than one million candidates have been screened on the platform, a company statement from 16 June 2026. ## Where does Tracy win? Each claim below is stated against HeyMilo's own wording. ### Every applicant gets an interview, whenever they click Each Tracy interview is its own session, started when the candidate clicks the link and ready in under 10 seconds, with no recruiter in the loop. Applicants do not queue behind a calendar, and the interview is available at any hour. A candidate who drops can rejoin the same session within 10 minutes, and a reminder and re-invitation go out after a day if an interview failed. On volume alone the two are level, because HeyMilo also screens every applicant without a recruiter present. What differs is what each interview contains. ### An interview built from the CV and the role Tracy generates the questions from the job description, and you can edit them or bring your own. With CV check on, it asks 3 to 5 further questions based on the candidate's CV and the job's must-have requirements, so two applicants for one role do not get the same script. You also set the style on a scale from quick to deep and from technical to behavioural, and choose the skills to evaluate. HeyMilo analyses resumes as a separate step in its funnel; ask it whether the interview questions themselves change for each candidate. ### Follow-ups that wait for the answer HeyMilo asks follow-ups in real time and lets the recruiter cap how many. In its comparison with Alex, HeyMilo writes that follow-up questions can only come from the recruiter's scoring rubric. That is a sound design for consistent screening at volume. Tracy's approach is different: it listens for specifics and asks for the missing detail in whatever the person said, which matters most when the answers are technical and the best follow-up is not on a rubric you wrote in advance. It also does not cut people off. It waits for a pause that suits the question, up to about 2 seconds on hard technical ones, and checks that the thought is finished before it replies. ### Technical tasks Tracy can introduce a coding task and probe the trade-offs behind an answer. HeyMilo's interview types are resume, form, SMS, voice or video, and the AI scenario assessment; its pages do not describe a coding interview. That is not the same as saying it cannot run one, so ask. For a closer look at what a technical screen needs, see our guide to [AI technical interviews](/blog/ai-technical-interviews). ### Evidence a reviewer can check After each interview Tracy writes an assessment for every question with the reasoning behind it, leaving out warm-ups, notice period and salary questions, and shows the transcript with timestamps beside the recording. For roles where spoken language matters it adds a language evaluation: fluency, coherence, grammar, pronunciation and vocabulary. HeyMilo's score report carries a per-question rationale grounded in what the candidate said, timestamped quotes, the transcript and the recording. On this dimension the two are close, so judge the sample reports side by side. Our guide to [AI interview reports](/blog/ai-interview-reports) lists what a reviewer should look for. ### Integrity signals, with the evidence During the interview Tracy captures signals: the candidate's face and where they are looking, response timing, looking away from the screen, tab switching, leaving fullscreen, an external screen, and whether the candidate asks for a question to be repeated or gives a suspiciously polished answer. Nothing is decided during the interview. When it ends, the signals are assessed together and the evidence goes to the reviewer beside the assessment. They are signals, not a verdict, and Tracy publishes no detection rate. HeyMilo's integrity layer is broader, as above. Our piece on [cheating detection in AI interviews](/blog/ai-interview-cheating-detection) explains how to use signals like these. ### It connects to the tools you already use Tracy connects through an API and an MCP server, so interviews can be created and read from the tools a team already works in, and a customer's own system can request an interview and hand the candidate a link. Enterprise teams can ask for a customised setup, such as the avatar or the voice. HeyMilo's ATS write-back is deeper, as above; ask each vendor how results reach the system your reviewers use. ## How does pricing compare? | Pricing question | Tracy AI Interviewer | HeyMilo | |---|---|---| | Public price | $1 per credit | None for the screening and interview platform; HeyMilo Red, a separate proctoring product, is priced from $0 | | What a unit buys | One credit per started 20 minutes of a session with a result | Credits pay for interviews, resume screens and SMS; the rate is not public | | Seats | None | Not stated | | Free start | First 3 results free, no card needed to start | No free trial stated for the platform; a new workspace needs an onboarding key from the HeyMilo team. HeyMilo Red has a free tier and a 14-day trial of its Growth plan | | Session with no answers | Never charged | Not stated | | Enterprise | No enterprise tier is listed on the pricing page | Through your account manager; no public figure | Here is the arithmetic for Tracy. A team that runs 100 interviews of up to 20 minutes pays for 97 results after the 3 free ones: 97 credits at $1, which is $97. If every session ran 21 to 40 minutes, each paid result costs two credits, so the same 100 interviews come to 194 credits, or $194. We cannot make the same calculation for HeyMilo's platform, because the rate is not public, and we have not used third-party estimates. HeyMilo Red is different. On red.heymilo.ai/pricing it lists Free at $0 (about 300 minutes of video analysis a month, card required to start) and Growth at $59 a month (about 3,000 minutes, then $0.04 per minute), and its launch post adds a 14-day free trial. But Red scans recorded interviews for integrity findings and charges by the minute analysed. It prices a scan, not an interview, so it cannot sit beside a Tracy credit. Before you compare totals, ask HeyMilo what one interview costs at your length and volume, what the subscription minimum is, what a resume screen or SMS adds, what happens past the credit allowance, and whether enterprise services are priced separately. For more per-candidate figures across vendors, see [AI interview cost per candidate](/blog/ai-interview-cost). ## Which should you choose? Choose HeyMilo when: - you hire hourly, staffing or other high-volume roles and want SMS, forms and resume screening ahead of the interview; - results must write back to Bullhorn, Avionte, JobDiva, Workday or another ATS on your list; - you need to assess how candidates work with an AI assistant; - you want a vendor-run enterprise rollout with a named customer success manager and engineer; - you want a standalone proctoring API, or a calendar step after the screen. Choose Tracy AI Interviewer when: - you want every applicant interviewed in depth, at any hour, with no scheduling; - the questions should follow each candidate's CV and answers, and the role may be technical; - reviewers need a recording, a transcript, written reasoning per question and integrity signals to check; - you want to start with 3 free results and see the price before a sales call. To test either one, run the same role through both. Write one answer that sounds plausible but leaves out a constraint, and compare the next question each interviewer asks, the evidence in the result and the minutes a reviewer needs to reach a decision. If you are comparing more than two tools, our [best AI interview software](/blog/best-ai-interview-software) list and the [HeyMilo alternatives](/compare/heymilo-alternatives) page cover the wider field. ## Common questions ### What is the main difference between Tracy and HeyMilo? HeyMilo screens applicants across voice, video, SMS and forms and syncs results to an ATS, so it is built for breadth. Tracy AI Interviewer gives each applicant a full two-way video or audio interview built from the role and, when you turn it on, the CV. It follows up on what the person says and shows integrity signals with evidence, so it is built for depth. ### Does HeyMilo publish its pricing? Not for its screening and interview platform. Its pricing page leads to a demo form, and its documentation says plan options and pricing appear once billing is set up for your workspace. Its separate proctoring product, HeyMilo Red, publishes Free at $0 and Growth at $59 a month, priced by minutes of recorded video analysed. Tracy publishes $1 per credit for the interview itself, with the first 3 results free. ### Can HeyMilo run coding or technical interviews? HeyMilo's pages do not describe a coding interview. Its interview types include voice, video, SMS, forms, resume analysis and an AI scenario assessment, which scores how candidates prompt an AI assistant. Ask HeyMilo directly if coding matters. Tracy can introduce a coding task and probe the trade-offs behind an answer. ### Which has better cheating detection? HeyMilo's integrity layer is broader: a trust score, AI-answer and impersonation checks, and a standalone proctoring API. Tracy captures signals such as tab switching, gaze, response timing and an external screen, and shows the evidence to a reviewer after the interview as signals, not a verdict. ### Does HeyMilo work with my ATS? HeyMilo lists 20+ ATS partners on its integrations page, including Workday, Greenhouse, iCIMS, Lever and Bullhorn, and syncs scores, reports, transcripts and recordings back to the candidate record. Tracy connects through its API and MCP server, including chat connectors such as Workable's, and its results are read in Tracy; the pages we checked describe no direct ATS write-back. --- **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=tracy-vs-heymilo&utm_content=demo) with your own role and compare the second question, or [try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=tracy-vs-heymilo&utm_content=try-free) with your first 3 results.** ## Sources All sources retrieved 3 Nov 2026. - Tracy AI Interviewer, [home page](https://tracyinterview.ai/): interview formats, follow-ups, coding task, bring your own questions, reports backed by the participant's words, languages of the product. - Tracy AI Interviewer, [pricing](https://tracyinterview.ai/pricing): credit price, billing unit, free results, no-answer policy, video or audio, no account for participants, recording notice, languages. - Tracy AI Interviewer, [Workable guide](https://tracyinterview.ai/help/workable): chat-based connector, data shared, results stay in Tracy. - HeyMilo, [home page](https://www.heymilo.ai): positioning, 20+ languages, ATS write-back, resume scoring. - HeyMilo, [pricing page (leads to a demo form)](https://www.heymilo.ai/pricing): no public price, demo form. - HeyMilo, [billing documentation](https://docs.heymilo.ai/deep-dive/settings-and-admin/billing): pricing shown after billing set-up, credits, enterprise pricing through the account manager. - HeyMilo, [interviewing candidates](https://www.heymilo.ai/interviewing-candidates): real-time follow-ups, transcript-backed scoring. - HeyMilo, [configuring your AI interviewer](https://docs.heymilo.ai/guides/build-your-interviewer/configuring-your-ai-interviewer.md): maximum follow-ups, evaluation criteria per question. - HeyMilo, [compare with Alex](https://www.heymilo.ai/compare/alex-ai): HeyMilo's statement that follow-ups come from the scoring rubric. - HeyMilo, [compare index](https://www.heymilo.ai/compare): the 15+ native integrations figure. - HeyMilo, [interview types](https://docs.heymilo.ai/guides/build-your-interviewer/interview-types.md): the interview types listed in its documentation. - HeyMilo, [AI scenario assessment](https://www.heymilo.ai/product-feature/ai-scenario-assessment): timed scenario with a built-in AI assistant. - HeyMilo, [cheating detection](https://www.heymilo.ai/product-feature/cheating-detection): trust score and flagged behaviours. - HeyMilo, [interview integrity documentation](https://docs.heymilo.ai/deep-dive/ai-settings/interview-integrity.md): video checks, flags are for review. - HeyMilo, [HeyMilo Red launch post](https://blog.heymilo.ai/heymilo-red-interview-proctoring-api), published 24 Sep 2026: standalone proctoring API, Red's Free and Growth prices, 14-day trial. - HeyMilo Red, [pricing](https://red.heymilo.ai/pricing): Free $0, Growth $59 a month, $0.04 per minute after the allowance, card required to start. - HeyMilo, [responsible AI](https://www.heymilo.ai/responsible-ai): score rationale, timestamped quotes, SOC 2 wording, monthly bias audits. - HeyMilo, [integrations](https://www.heymilo.ai/integrations): 20+ ATS partners. - HeyMilo, [interview scheduling](https://www.heymilo.ai/interview-scheduling): calendar booking after the screen. - HeyMilo, [enterprise services](https://www.heymilo.ai/product-feature/enterprise-services): customer success manager and forward deployed engineer. - HeyMilo, [funding announcement](https://blog.heymilo.ai/heymilo-raises-6-million), published 16 Jun 2026: more than one million candidates screened, company statement. - HeyMilo, [self-serve onboarding documentation](https://docs.heymilo.ai/guides/get-started/self-serve-onboarding.md): onboarding key from the HeyMilo team. - HeyMilo, [scoring documentation](https://docs.heymilo.ai/guides/build-your-interviewer/how-scoring-works.md): rubric scoring and recruiter override. # AI Interview Cost: What You Pay Per Candidate in 2026 > Compare published AI interview costs from Tracy, Ribbon, Spark Hire, Willo and Hireflix, including subscriptions, limits and review time. - URL: https://tracyinterview.ai/blog/ai-interview-cost - Published: 2026-11-03 - Updated: 2026-11-11 - Author: Tracy team ![Cover graphic introducing ai interview cost.](/blog/images/ai-interview-cost-cover.png) Fact-checked on 2 Nov 2026 against five vendors' current pricing or product pages ([see Sources](#sources)). > **In short** > - Tracy AI Interviewer starts at **$1 per completed result up to 20 minutes** after three free results. > - Ribbon starts at **$499 a month billed annually for 100 interviews**, with published overages. > - Willo lists **$59 per live role** for occasional hiring; Hireflix advertises annual plans from **$75 a month** with unlimited interviews. > - The useful number is not the plan's lowest possible cost. Calculate subscription, overage, unused capacity, setup and reviewer minutes per completed interview. **On this page** - [How do AI interview tools charge?](#how-do-ai-interview-tools-charge) - [What do published AI interview prices look like?](#what-do-published-ai-interview-prices-look-like) - [How should you calculate cost per candidate?](#how-should-you-calculate-cost-per-candidate) - [Which hidden costs matter?](#which-hidden-costs-matter) - [Which pricing model fits your hiring pattern?](#which-pricing-model-fits-your-hiring-pattern) - [Questions](#questions) - [Sources](#sources) ## How do AI interview tools charge? Most products use one of four models: 1. **Per completed result.** You buy credits or interviews and pay when a candidate produces a result. 2. **Monthly capacity.** The plan includes a number of interviews, seats or active roles, then charges overage. 3. **Per live role.** You pay while a role is open, regardless of the number of responses inside the plan limit. 4. **Annual platform subscription.** Video interviews sit inside an ATS or hiring suite with other features. The models are not interchangeable. Ribbon's price includes sourcing and outreach. Spark Hire's current Recruit pricing includes applicant tracking features. Tracy's published price covers the interview result. Compare the work included before comparing the dollar amount. ![Three parts of an interview cost calculation: fees, usage and reviewer time.](/blog/images/ai-interview-cost-process.png) ## What do published AI interview prices look like? | Product | Published starting price | Model | Important limit or condition | |---|---:|---|---| | Tracy AI Interviewer | $1 per credit | Completed result length | One credit for each started 20 minutes; three results free | | Ribbon advertised entry package | $499/month billed annually | Monthly interview capacity | 100 interviews; current overage needs a quote | | Spark Hire Recruit Pro | $335/month billed annually | ATS subscription | Annual contract; one-way video included | | Willo Lite | $59/live role | Per open role | Intended for occasional hiring | | Hireflix Small | $75/month billed yearly | Flat subscription | Vendor advertises unlimited interviews | Ribbon's dedicated pricing page returned 404 in a browser on 2 Nov 2026. The entry figure above comes from its [comparison with BrightHire](https://www.ribbon.ai/blog/ribbon-vs-brighthire), published 23 Aug 2026; confirm current terms with Ribbon. These products do different jobs. Hireflix and Willo center on one-way recorded responses. Ribbon combines recruiting operations with interviews. Spark Hire bundles video into an ATS. Tracy runs a two-way spoken interview with follow-ups. Start with [one-way vs two-way interviews](/blog/one-way-vs-two-way-video-interviews?utm_source=blog&utm_medium=seo&utm_campaign=ai-interview-cost&utm_content=internal-link) if the required format is not settled. ## How should you calculate cost per candidate? Use completed interviews, not invited candidates or plan capacity: `cost per completed interview = (subscription + overage + monthly setup allocation + reviewer cost) / completed interviews` Illustrative calculation using Ribbon's advertised entry price: a team pays $499 and completes 60 of the included 100 interviews. The subscription portion is $8.32 per completed interview. If a recruiter spends eight minutes reviewing each result at an employment cost of $45 an hour, review adds $6 per result. The working total becomes $14.32 before implementation and integration costs. This is a calculation example, not a current Ribbon quote. Run the same calculation for your current process. Include scheduling, reminders, the screen itself, note-taking and manager review. Ribbon's own calculator separates platform cost from recruiter time and labels the result as an estimate. The detailed plan arithmetic is in [Ribbon pricing](/compare/ribbon-pricing?utm_source=blog&utm_medium=seo&utm_campaign=ai-interview-cost&utm_content=internal-link). Tracy's pricing is length-based, so a 35-minute completed interview uses two credits and costs $2 at list price. ## Which hidden costs matter? ### Reviewer time A one-way video may be inexpensive to collect but expensive to watch. Measure the full review path, including replay, notes and manager handoff. For an AI-generated report, measure correction time and how often the reviewer opens the recording. ### Unused capacity Annual subscriptions assume steady demand. Divide by actual monthly completions during both peak and quiet periods. Ask whether unused interviews carry forward. ### Roles, seats and integrations Some plans limit active roles or team seats. ATS mapping, SSO, support and data migration may require a higher tier or implementation fee. ### Candidate drop-off An incomplete interview consumes outreach and recruiter attention without producing a result. Track completion by device, role and candidate group. Do not turn a vendor's completion claim into your forecast. ### Contract scope Pricing pages change. Put the interview allowance, overage, support, retention, implementation and renewal terms into the written order form. ## Which pricing model fits your hiring pattern? | Hiring pattern | Model to test first | |---|---| | Occasional skilled roles | Per completed result | | Many open roles with predictable volume | Monthly capacity | | A few campaigns with many applicants | Per live role or flat one-way subscription | | Recruiting-suite replacement | Annual platform subscription | | Seasonal volume | Per result, flexible capacity or a written downgrade path | Price should follow the interview job. Teams hiring engineers and experts may pay more for a result that contains useful follow-ups. Teams checking basic eligibility may need a shorter, cheaper format. See [best AI interview software](/blog/best-ai-interview-software?utm_source=blog&utm_medium=seo&utm_campaign=ai-interview-cost&utm_content=internal-link) for the product comparison and [AI interview reports](/blog/ai-interview-reports?utm_source=blog&utm_medium=seo&utm_campaign=ai-interview-cost&utm_content=internal-link) for the evidence test. ![A checklist of the main actions for ai interview cost.](/blog/images/ai-interview-cost-takeaways.png) ## Common questions ### How much does an AI interview cost? Published prices range from per-result credits to monthly or annual subscriptions. Tracy starts at $1 for a completed result up to 20 minutes. Ribbon starts at $499 a month billed annually for 100 interviews. ### Which AI interview tools publish prices? Tracy, Ribbon, Spark Hire, Willo and Hireflix publish at least one price. HireVue and several enterprise AI interviewers require a quote. ### What is the cheapest model for occasional hiring? Pay-per-result or pay-per-live-role models avoid a year-round subscription. Compare Tracy's credits with Willo's $59 live-role plan for the format you need. ### What cost do buyers usually miss? Reviewer time is the common hidden cost. A low subscription price can still be expensive if hiring managers must watch every recording or correct weak summaries. ### Should cost per interview include unused capacity? Yes. Divide the actual monthly subscription by completed interviews, not by the plan maximum. --- **[Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=ai-interview-cost&utm_content=try-free)** or **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=ai-interview-cost&utm_content=demo)** ## Sources All sources retrieved 2 Nov 2026. - Tracy AI Interviewer, [pricing](https://tracyinterview.ai/pricing): credits, result-length billing, free results and no-answer policy. - Ribbon, [Ribbon vs BrightHire](https://www.ribbon.ai/blog/ribbon-vs-brighthire), 23 Aug 2026: advertised $499/month annual-billing starting price and 100 interviews. The pricing page returned 404 in a browser on 2 Nov 2026; overages and higher tiers need a quote. - Spark Hire, [pricing](https://www.sparkhire.com/pricing/): current Recruit prices, annual contract and one-way video inclusion. - Willo, [2026 platform comparison](https://www.willo.video/blog/one-way-video-interview-platforms): current Lite and Enterprise prices and one-way format. - Hireflix, [2026 pricing guide](https://hireflix.com/blog/one-way-video-interview-software-pricing/): published annual subscription prices and included usage. # HireVue Reviews: What Buyers and Candidates Say (Nov 2026) > HireVue reviews summarized from G2 and Capterra: ease of use, scale, flexibility, one-way interview concerns, integrations and price. - URL: https://tracyinterview.ai/compare/hirevue-reviews - Published: 2026-11-03 - Author: Tracy team ![Cover graphic introducing hirevue reviews.](/blog/images/hirevue-reviews-cover.png) Fact-checked on 2 Nov 2026 against 255 G2 reviews, 51 Capterra reviews and HireVue's current product pages ([see Sources](#sources)). > **In short** > - HireVue has a **4.1/5 rating from 255 G2 reviews** and **4.5/5 from 51 Capterra reviews**. Do not combine the scores because the sites use different review sets. > - Reviewers commonly praise ease of use, flexibility and the ability to support high-volume hiring. > - The visible concern themes include impersonal one-way interviews, integration issues, audio problems, app glitches and cost. > - HireVue is broader than its older one-way-video reputation. Its current suite includes a two-way AI Interviewer, assessments and live video. **On this page** - [What do HireVue ratings show?](#what-do-hirevue-ratings-show) - [What do buyers like about HireVue?](#what-do-buyers-like-about-hirevue) - [What complaints appear in HireVue reviews?](#what-complaints-appear-in-hirevue-reviews) - [What do candidates experience?](#what-do-candidates-experience) - [Who is HireVue best for?](#who-is-hirevue-best-for) - [What should you test before buying?](#what-should-you-test-before-buying) - [Questions](#questions) - [Sources](#sources) ## What do HireVue ratings show? | Review source | Rating | Review count | Details visible on 2 Nov 2026 | |---|---:|---:|---| | G2 | 4.1/5 | 255 | 55% five-star, 32% four-star; 190 enterprise reviewers | | Capterra | 4.5/5 | 51 | Ease of use 4.4; customer service 4.3 | The ratings show broad satisfaction, especially among enterprise users. They do not answer whether HireVue fits a particular workflow. G2 groups reviews across video interviewing, assessments, scheduling and AI recruiting, so a team buying the new AI Interviewer should confirm which product each review discusses. ![Three steps for interpreting reviews: check the source, read both sides and run a demo.](/blog/images/hirevue-reviews-process.png) ## What do buyers like about HireVue? G2's visible review summary names video interviewing, scheduling flexibility, time savings, setup ease and hiring efficiency among the positive themes. A validated mid-market reviewer described regular use during high-volume hiring and praised the interface, implementation, support, assessments and ATS connections. Candidate-facing reviews also mention practice recordings and the ability to complete an interview from home. That flexibility matters for distributed hiring, shift workers and applicants who cannot take a recruiter call during business hours. The broader product is another reason enterprises choose HireVue. It combines one-way video, live interviewing, assessments and a voice AI Interviewer. A buyer can standardize several stages without assembling separate products. ## What complaints appear in HireVue reviews? G2's current theme list includes integration issues, inefficient recruiting processes, candidate management, audio issues and app glitches. Capterra surfaces a candidate review describing one-way recording as unnatural and anxiety-inducing, with concerns about automated assessment. The format matters here. A one-way video presents preset questions and records each response. HireVue's newer AI Interviewer runs a two-way voice conversation with follow-up questions. A complaint about talking to a camera does not automatically describe the AI Interviewer, and praise for the assessment suite may not describe the candidate's video experience. Price also appears as a concern on G2's pricing summary. HireVue's own pricing page does not publish a standard plan amount, so buyers need a quote. Our [HireVue pricing guide](/compare/hirevue-pricing?utm_source=blog&utm_medium=seo&utm_campaign=hirevue-reviews&utm_content=internal-link) explains what is public and what must be confirmed. ## What do candidates experience? HireVue's one-way format gives candidates a deadline, a set of prompts and the ability to record without scheduling a live call. That can widen access, but the employer controls think time, retakes and the amount of explanation given before the interview. The AI Interviewer is different. HireVue says it asks probing follow-up questions in a two-way voice conversation, evaluates responses against role-specific criteria and explains why a response meets them. Recruiters make the final call. Candidate communication should name the format. Tell people whether they will record preset answers, speak with an AI interviewer or join a person. Greenhouse's 2026 candidate research found that lack of disclosure was a common reason people left AI-assisted hiring processes. See [one-way vs two-way interviews](/blog/one-way-vs-two-way-video-interviews?utm_source=blog&utm_medium=seo&utm_campaign=hirevue-reviews&utm_content=internal-link) for the full distinction. ## Who is HireVue best for? HireVue has the clearest fit for employers that need several of these capabilities together: - high-volume video interviewing; - technical, game-based or other assessments; - structured rating guides and enterprise administration; - live video and scheduling; - a conversational voice AI interviewer; - enterprise security or government procurement requirements. A smaller team that only needs a deep first conversation may prefer a focused product with public usage pricing. Compare [Tracy vs HireVue](/compare/tracy-vs-hirevue?utm_source=blog&utm_medium=seo&utm_campaign=hirevue-reviews&utm_content=internal-link) and [best AI interview software](/blog/best-ai-interview-software?utm_source=blog&utm_medium=seo&utm_campaign=hirevue-reviews&utm_content=internal-link) before deciding. ## What should you test before buying? Use the same two roles throughout the pilot. Include one role with clear eligibility questions and another where expertise appears only after a follow-up. Then check: 1. Candidate completion by device and time of day. 2. Whether the AI Interviewer asks for missing detail instead of accepting a polished first answer. 3. Whether reviewers can trace scores to the response. 4. ATS writeback, duplicate records and error recovery. 5. The time hiring managers spend watching video or correcting summaries. 6. Notice, consent, accommodation and human-review paths. For an evidence-first review workflow, see [what an AI interview report should contain](/blog/ai-interview-reports?utm_source=blog&utm_medium=seo&utm_campaign=hirevue-reviews&utm_content=internal-link). ![A checklist of the main actions for hirevue reviews.](/blog/images/hirevue-reviews-takeaways.png) ## Common questions ### Is HireVue well reviewed? Yes. G2 shows 4.1 out of 5 from 255 reviews, while Capterra shows 4.5 from 51 reviews. Ratings describe different reviewer groups and should not be combined. ### What do reviewers like about HireVue? Review themes include ease of use, scheduling flexibility, time savings, interview practice and support for high-volume hiring. ### What do reviewers dislike about HireVue? The recurring concerns visible in the review summaries include the impersonal feel of one-way video, integration issues, audio problems, app glitches and price. ### Does HireVue only offer one-way interviews? No. HireVue also sells live video, assessments and an AI Interviewer that runs a two-way voice conversation with follow-up questions. ### Who should choose HireVue? Large employers that want video interviewing, assessments, enterprise controls and an AI interviewer in one suite have the clearest fit. --- **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=hirevue-reviews&utm_content=demo)** or **[Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=hirevue-reviews&utm_content=try-free)** ## Sources All sources retrieved 2 Nov 2026. - G2, [HireVue seller profile](https://www.g2.com/sellers/hirevue) and [reviews](https://www.g2.com/products/hirevue/reviews): rating, review count, reviewer mix and visible themes. - Capterra, [HireVue reviews](https://www.capterra.com/p/132193/Team-Build/reviews/): rating, count, category scores and candidate review excerpt. - HireVue, [AI Interviewer](https://www.hirevue.com/platform/ai-interviewer): two-way voice format, follow-ups, criteria and recruiter decision. - HireVue, [video interviewing](https://www.hirevue.com/platform/video-interviewing): one-way and live video product scope. - HireVue, [pricing](https://www.hirevue.com/pricing): quote-based pricing page. # Ribbon Alternatives for AI Interviews (Nov 2026) > Compare six Ribbon alternatives for deeper interviews, enterprise suites, multilingual screening, hourly hiring and one-way video. - URL: https://tracyinterview.ai/compare/ribbon-alternatives - Published: 2026-11-03 - Updated: 2026-11-11 - Author: Tracy team ![Cover graphic introducing ribbon alternatives.](/blog/images/ribbon-alternatives-cover.png) Fact-checked on 2 Nov 2026 against each vendor's own product or pricing page ([see Sources](#sources)). > **In short** > - **Tracy AI Interviewer** is the Ribbon alternative for teams that already have applicants and want a deeper spoken interview with adaptive follow-ups. > - **HireVue** fits enterprises that want assessments, video interviewing and an AI interviewer in one suite. > - **HeyMilo** covers more candidate channels; **Alex** combines interviews and scheduling; **Sapia.ai** centers on chat-based high-volume hiring; **Spark Hire** fits one-way and live video. > - Stay with Ribbon when sourcing, talent rediscovery and outreach are part of the reason you bought it. **On this page** - [Why do teams consider Ribbon alternatives?](#why-do-teams-consider-ribbon-alternatives) - [The six best Ribbon alternatives](#the-six-best-ribbon-alternatives) - [How do the alternatives compare?](#how-do-the-alternatives-compare) - [How should you choose?](#how-should-you-choose) - [Questions](#questions) - [Sources](#sources) ## Why do teams consider Ribbon alternatives? Ribbon covers more than interviews. It sources candidates, searches existing ATS talent, runs outreach, conducts interviews and syncs results back to recruiting systems. That breadth is a strength until a buyer needs one narrower job done differently. The alternatives below make sense for one of four reasons: - the interview needs deeper follow-ups for skilled or technical work; - the company wants an established enterprise assessment suite; - candidates need more channels or a different interview format; - the team only needs video interviewing and does not want the rest of a recruiting platform. If price is the main concern, read the current [Ribbon pricing breakdown](/compare/ribbon-pricing?utm_source=blog&utm_medium=seo&utm_campaign=ribbon-alternatives&utm_content=internal-link) first. Changing products can cost more than choosing the correct Ribbon plan. ![Three steps for choosing an alternative: name the job, shortlist tools and test a brief.](/blog/images/ribbon-alternatives-process.png) ## The six best Ribbon alternatives ### 1. Tracy AI Interviewer for adaptive interview depth Tracy runs spoken video or audio interviews in the browser. It follows what the participant says, asks for missing detail, and organizes each result around fields the reviewer chose. Technical interviews can include a coding task and questions about the candidate's trade-offs. Choose Tracy when your team already has candidates and the quality of the second question matters more than sourcing or outreach. It is a smaller product surface than Ribbon. That is a benefit for an interview-first purchase and a limitation for a team replacing several recruiting tools. Tracy publishes per-result pricing: the first three results are free, then $1 per credit. One credit covers each started 20 minutes. See [Tracy vs Ribbon](/compare/tracy-vs-ribbon?utm_source=blog&utm_medium=seo&utm_campaign=ribbon-alternatives&utm_content=internal-link) for the direct comparison. ### 2. HireVue for an enterprise hiring suite HireVue's AI Interviewer conducts voice interviews with probing follow-ups and evaluates responses against role-specific criteria. The wider platform includes assessments, one-way video and live interviewing. HireVue does not publish a standard price. Choose HireVue when procurement, assessment science, government requirements or a single enterprise suite outweigh a self-serve start. Read [Tracy vs HireVue](/compare/tracy-vs-hirevue?utm_source=blog&utm_medium=seo&utm_campaign=ribbon-alternatives&utm_content=internal-link) or [HireVue reviews](/compare/hirevue-reviews?utm_source=blog&utm_medium=seo&utm_campaign=ribbon-alternatives&utm_content=internal-link) before building a shortlist. ### 3. HeyMilo for channels and languages HeyMilo supports voice, video, SMS and forms. Its current multilingual guide documents 14 or more languages across those channels, with per-question scoring and recruiter-readable reports. Its Workable integration writes scores, summaries, recording links and reports into candidate profiles. Choose HeyMilo when the candidate channel or language list drives the project. Ask for a pilot in your second-largest language and verify that the same scoring and consent flow works there. Pricing is quote-based on the pages checked. ### 4. Alex for scheduled, multichannel interviews Alex runs live two-way interviews over video, phone or chat and asks follow-up questions based on the candidate's answers. Candidates schedule a time and can complete the interview on any device. The employer receives the recording, transcript and notes; Alex says humans make the hiring decision. Choose Alex when scheduling and several interview channels belong in the same workflow. Pricing is not published on the product pages checked. ### 5. Sapia.ai for chat-based high-volume hiring Sapia.ai centers on a mobile-first chat interview and sells to employers with high hiring volume. Its pricing page uses a per-hire model and says the product is designed for companies making more than 500 hires a year. Choose Sapia.ai when an untimed chat format and per-hire commercial model fit the candidate population. Pick a spoken interviewer when voice, video or live follow-up is essential to the role. ### 6. Spark Hire for one-way and live video Spark Hire combines one-way video with live interviews and a broader applicant-tracking product. Its current pricing page lists Recruit Pro from $335 a month billed annually, with one-way video and AI video review among the included recruiting features. Choose Spark Hire when hiring managers want to watch candidate recordings and keep live video in the same product. A one-way recording cannot ask a question based on the answer it just heard. Our [one-way vs two-way guide](/blog/one-way-vs-two-way-video-interviews?utm_source=blog&utm_medium=seo&utm_campaign=ribbon-alternatives&utm_content=internal-link) explains the trade-off. ## How do the alternatives compare? | Product | Best for | Interview format | Published price? | What Ribbon still does better | |---|---|---|---|---| | Tracy AI Interviewer | Skilled and technical follow-ups | Two-way video or audio | Yes | Sourcing and outreach | | HireVue | Enterprise suite and assessments | AI voice, one-way and live video | No | Faster public self-serve entry | | HeyMilo | Multichannel, multilingual screening | Voice, video, SMS and forms | No | Public package detail | | Alex | Scheduled interviews across channels | Video, phone and chat | No | Sourcing and talent rediscovery | | Sapia.ai | High-volume chat interviews | Text chat | Per-hire model, quote required | Spoken voice and video workflow | | Spark Hire | One-way and live video | Recorded and live video | Yes | Adaptive AI interview plus outreach | No alternative wins every row. Ribbon remains the strongest fit in this group when a buyer wants sourcing, rediscovery, outreach and interviews under one subscription. ## How should you choose? Use one live role and test the product against work your team already does. Record: 1. Minutes from invitation to a usable result. 2. Minutes a recruiter spends reviewing and correcting it. 3. Whether each score can be traced to what the candidate said. 4. Completion and restart rates by device and candidate group. 5. Manual work left in the ATS. Include an answer that needs a follow-up, a candidate who asks for another format and an integration error. The polished demo path rarely reveals the work your team will inherit. The broader market comparison is in [best AI interview software](/blog/best-ai-interview-software?utm_source=blog&utm_medium=seo&utm_campaign=ribbon-alternatives&utm_content=internal-link), while [AI interview reports](/blog/ai-interview-reports?utm_source=blog&utm_medium=seo&utm_campaign=ribbon-alternatives&utm_content=internal-link) covers the evidence reviewers should expect. ![A checklist of the main actions for ribbon alternatives.](/blog/images/ribbon-alternatives-takeaways.png) ## Common questions ### What is the best Ribbon alternative for technical interviews? Tracy AI Interviewer is the focused choice when role-specific follow-ups, coding tasks and evidence behind the report matter more than sourcing and outreach. ### Which Ribbon alternative is best for a full enterprise suite? HireVue combines its AI Interviewer with assessments, one-way video, live interviewing and enterprise controls. ### Which alternatives publish prices? Tracy and Spark Hire publish prices. HireVue, HeyMilo, Alex and Sapia.ai require a quote for the products compared here. ### Which alternative works for multilingual and multichannel screening? HeyMilo supports voice, video, SMS and forms and documents screening in 14 or more languages. ### Should I leave Ribbon if I use its sourcing and outreach? Probably not without a replacement plan. Most alternatives in this guide cover the interview rather than Ribbon's full sourcing, rediscovery and outreach workflow. --- **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=ribbon-alternatives&utm_content=demo)** or **[Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=ribbon-alternatives&utm_content=try-free)** ## Sources All sources retrieved 2 Nov 2026. - Ribbon, [product](https://www.ribbon.ai/) and [comparison with BrightHire](https://www.ribbon.ai/blog/ribbon-vs-brighthire): platform scope and advertised entry package. The dedicated pricing page returned 404 on 2 Nov 2026; confirm current terms with Ribbon. - Tracy AI Interviewer, [product](https://tracyinterview.ai/) and [pricing](https://tracyinterview.ai/pricing): formats, follow-ups, reports and credit pricing. - HireVue, [AI Interviewer](https://www.hirevue.com/platform/ai-interviewer) and [pricing](https://www.hirevue.com/pricing): interview behavior, wider suite and quote model. - HeyMilo, [product](https://www.heymilo.ai/): candidate channels and report format. - Alex, [AI Interviewer](https://www.alex.com/product/ai-interviewer): formats, scheduling, follow-ups and human decision. - Sapia.ai, [pricing](https://sapia.ai/pricing/): per-hire model and company-size guidance. - Spark Hire, [pricing](https://www.sparkhire.com/pricing/): current Recruit plans and video interview features. # Ribbon Pricing: Plans, Overage and Real Cost (Nov 2026) > Ribbon advertises $499/month for 100 interviews in its comparison article. See what is confirmed, what needs a quote and how to compare costs. - URL: https://tracyinterview.ai/compare/ribbon-pricing - Published: 2026-11-03 - Updated: 2026-11-11 - Author: Tracy team ![Cover graphic introducing ribbon pricing.](/blog/images/ribbon-pricing-cover.png) Checked on 2 Nov 2026. Ribbon's pricing page returned a real 404 in a browser. Its [comparison with BrightHire](https://www.ribbon.ai/blog/ribbon-vs-brighthire) still advertises an annual-billing entry package of $499/month for 100 interviews and a seven-day trial. Confirm current terms with Ribbon before budgeting. > **In short** > - Ribbon's own comparison article advertises **$499/month, billed annually, for 100 interviews**. Its dedicated pricing page is currently unavailable. > - Higher-tier prices, seats, active-role limits and overage rates need a current written quote. > - At 100 completed interviews, that advertised entry price works out to **$4.99 per interview**, before any other charges or reviewer time. > - Tracy AI Interviewer charges $1 per credit instead of a monthly subscription. One credit covers each started 20 minutes of a completed result. **On this page** - [How much does Ribbon cost?](#how-much-does-ribbon-cost) - [What is included in each Ribbon plan?](#what-is-included-in-each-ribbon-plan) - [What does Ribbon cost at different volumes?](#what-does-ribbon-cost-at-different-volumes) - [Which costs sit outside the subscription?](#which-costs-sit-outside-the-subscription) - [How does Tracy pricing compare?](#how-does-tracy-pricing-compare) - [Questions](#questions) - [Sources](#sources) ## How much does Ribbon cost? The accessible first-party source is Ribbon's comparison with BrightHire, published on 23 August 2026. It advertises a starting package; it does not establish that every previously advertised plan remains available today. | Item | What the accessible source confirms | |---|---| | Advertised starting price | $499/month, billed annually | | Included interviews | 100 per month | | Trial | Seven days advertised | | Higher tiers and Enterprise | Request current prices and limits | | Seats, roles and overage | Not confirmed by this source; request a quote | The starting figure implies $5,988 over twelve months before additional charges. Confirm the billing schedule, commitment, cancellation terms and any taxes with Ribbon. ![Three steps for calculating Ribbon pricing: choose a plan, estimate usage and add overages.](/blog/images/ribbon-pricing-process.png) ## What is included in each Ribbon plan? Ribbon describes a platform covering sourcing, talent rediscovery, outreach, voice or video interviews, structured scoring, ATS connections, analytics and API access. Those product capabilities do not prove what is included in the entry package. Ask Ribbon to itemize the modules, seats, active roles, usage, support and integration work in the proposed contract. That scope matters when comparing Ribbon with an interview-only product. The fee covers sourcing and outreach capabilities as well as interviews. Read [Tracy vs Ribbon](/compare/tracy-vs-ribbon?utm_source=blog&utm_medium=seo&utm_campaign=ribbon-pricing&utm_content=internal-link) for the workflow comparison or [Ribbon alternatives](/compare/ribbon-alternatives?utm_source=blog&utm_medium=seo&utm_campaign=ribbon-pricing&utm_content=internal-link) for other product shapes. ## What does Ribbon cost at different volumes? Only the 100-interview entry allowance is confirmed by the accessible comparison article. Larger-volume pricing cannot be calculated without current overage rates or a higher-tier quote. | Completed interviews in a month | Calculation using the advertised entry package | |---:|---| | 60 | $499 / 60 = $8.32 per completed interview | | 100 | $499 / 100 = $4.99 per completed interview | | 500 | Request a quote for capacity and overage | | 1,000 | Request a quote for capacity and overage | These are illustrative calculations using the advertised $499 starting price, not a vendor quote. A team with seasonal hiring should model a quiet month and a peak month using the actual terms it receives. Our [AI interview cost guide](/blog/ai-interview-cost?utm_source=blog&utm_medium=seo&utm_campaign=ribbon-pricing&utm_content=internal-link) compares published models across vendors. ## Which costs sit outside the subscription? With the pricing page unavailable, request a written quote for the exact workflow. Ask about: - implementation work and data migration; - ATS field mapping beyond the listed package; - extra seats or active roles; - interview volume that carries forward, if any; - data retention and storage terms; - support response times; - the cost of changing plans during a seasonal peak. The [Ribbon ROI calculator](https://www.ribbon.ai/calculator) estimates recruiter time saved and lets a buyer enter platform and one-time costs. It labels the result as an estimate rather than a cash-savings forecast. Use your own review minutes, salary and completion data. ## How does Tracy pricing compare? Tracy AI Interviewer uses credits rather than a monthly package: | Completed result length | Credits | Published price | |---|---:|---:| | Up to 20 minutes | 1 | $1 | | 21 to 40 minutes | 2 | $2 | | 41 to 60 minutes | 3 | $3 | The first three results are free. Credits do not expire, and a session with no answers is not charged. Tracy does not bundle sourcing, outreach, seats or active roles into the published price. At 100 results of 20 minutes or less, Tracy's standard rate is $100 before free results or any volume agreement. Ribbon's comparison article advertises $499 for 100 interviews with annual billing, but the package and current terms need confirmation. Compare the written quote with Tracy's usage model and the workflow your team needs. For teams still comparing formats, [one-way and two-way interviews](/blog/one-way-vs-two-way-video-interviews?utm_source=blog&utm_medium=seo&utm_campaign=ribbon-pricing&utm_content=internal-link) can change the review-time calculation more than the subscription price. ![A checklist of the main actions for ribbon pricing.](/blog/images/ribbon-pricing-takeaways.png) ## Common questions ### How much does Ribbon cost? Ribbon's comparison article advertises a starting price of $499/month billed annually for 100 interviews. The dedicated pricing page is unavailable; request current higher-tier prices and terms. ### How many interviews are included? The comparison article advertises 100 interviews per month at the starting price. Ask Ribbon to confirm other allowances. ### What happens when I exceed the limit? The accessible source does not confirm current overage rates. Request the rate, hard limits and any rollover policy in writing. ### Does Ribbon offer a free trial? Ribbon's comparison article advertises a seven-day trial. Confirm current eligibility and terms before signing up. ### How does Tracy pricing compare? Tracy has no published monthly plan. The first three results are free, then each credit costs $1 and covers each started 20 minutes of a completed result. --- **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=ribbon-pricing&utm_content=demo)** or **[Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=ribbon-pricing&utm_content=try-free)** ## Sources All sources retrieved 2 Nov 2026. - Ribbon, [Ribbon vs BrightHire](https://www.ribbon.ai/blog/ribbon-vs-brighthire), 23 Aug 2026: advertised $499/month annual-billing starting price, 100 interviews and seven-day trial. The dedicated pricing page returned 404 in a browser on 2 Nov 2026; higher tiers and overages are not confirmed. - Ribbon, [ROI calculator](https://www.ribbon.ai/calculator): calculation method and buyer-entered platform costs. - Tracy AI Interviewer, [pricing](https://tracyinterview.ai/pricing): credits, result-length billing, free results, expiry and no-answer policy. # Tracy vs Ribbon: AI Interviews Compared (Nov 2026) > Tracy vs Ribbon compared on interview depth, sourcing, outreach, evidence and price. See which AI interviewer fits your hiring workflow. - URL: https://tracyinterview.ai/compare/tracy-vs-ribbon - Published: 2026-11-03 - Updated: 2026-11-11 - Author: Tracy team ![Cover graphic introducing tracy vs ribbon.](/blog/images/tracy-vs-ribbon-cover.png) Checked on 2 Nov 2026 against first-party sources ([see Sources](#sources)). Ribbon's dedicated pricing page is unavailable; its comparison article still advertises the entry package quoted below. Confirm current terms with Ribbon. > **In short** > - Choose **Tracy AI Interviewer** when the interview itself is the product you need: spoken video or audio, follow-up questions based on each answer, and a report tied to what the candidate said. > - Choose **Ribbon** when you want one system for sourcing, talent rediscovery, outreach, interviews and ATS-connected recruiting operations. > - Tracy starts at $1 for a completed result of up to 20 minutes. Ribbon's comparison article advertises $499/month billed annually for 100 interviews; current overage rates need a quote. **On this page** - [Tracy vs Ribbon at a glance](#tracy-vs-ribbon-at-a-glance) - [How do their interviews differ?](#how-do-their-interviews-differ) - [Where does Ribbon win?](#where-does-ribbon-win) - [Where does Tracy win?](#where-does-tracy-win) - [How does pricing compare?](#how-does-pricing-compare) - [Which should your team choose?](#which-should-your-team-choose) - [Questions](#questions) - [Sources](#sources) ## Tracy vs Ribbon at a glance | Buying question | Tracy AI Interviewer | Ribbon | |---|---|---| | Main job | Run a spoken interview and return answers with supporting evidence | Run sourcing, rediscovery, outreach, interviews and connected recruiting operations | | Interview format | Two-way video or audio in the browser | Voice or video interviews, plus email, SMS and WhatsApp outreach | | Follow-up depth | Follows what the participant says and asks for missing detail | Structured interview questions with customizable scoring criteria | | Reviewer evidence | Structured fields, direct quotes, recording and transcript | Ranked summary, recording, transcript and configured scoring | | Published entry price | Three results free, then $1 per credit | $499 a month billed annually for 100 interviews | | Best fit | Skilled, technical and expert interviews where the second question matters | Teams that want the work before and around the interview in one platform | ![A three-step comparison: define the need, run the same brief and inspect the evidence.](/blog/images/tracy-vs-ribbon-process.png) ## How do their interviews differ? Tracy starts from a brief. You describe who you want to interview and what you need to learn, then share a link. The participant joins in a browser with no account. Tracy asks the prepared questions, follows an answer when a useful detail is missing, and returns structured fields with direct quotes. The [Tracy product page](https://tracyinterview.ai/?utm_source=blog&utm_medium=seo&utm_campaign=tracy-vs-ribbon&utm_content=internal-link) describes video and audio interviews, technical tasks and reports backed by the participant's own words. Ribbon describes a wider operating layer. Its AI recruiter can contact applicants, run voice or video interviews, score against a configurable rubric, sync results to an ATS and flag integrity signals. Ribbon's public pages also cover sourcing, talent rediscovery and multichannel outreach. The practical difference appears after an applicant gives a vague first answer. A team evaluating Tracy should test whether the next question reaches the work behind the claim. A team evaluating Ribbon should test that too, then test the handoffs from sourcing and outreach through the ATS. For the broader category, see [what an AI interviewer is](/blog/what-is-an-ai-interviewer?utm_source=blog&utm_medium=seo&utm_campaign=tracy-vs-ribbon&utm_content=internal-link) and the difference between [one-way and two-way video interviews](/blog/one-way-vs-two-way-video-interviews?utm_source=blog&utm_medium=seo&utm_campaign=tracy-vs-ribbon&utm_content=internal-link). ## Where does Ribbon win? Pick Ribbon when the recruiting team wants to find and contact candidates before interviewing them. Ribbon lists sourcing across a large profile index, ATS talent rediscovery, email and SMS outreach, WhatsApp outreach, interviews, analytics and API access. Tracy does not claim to be a sourcing platform. Ribbon's comparison article advertises an annual-billing starting package with 100 interviews per month. Its dedicated pricing page returned 404 during this check, so seats, active roles, higher tiers and overages need confirmation in a written quote. Staffing firms that want sourcing, contact and screening in one workflow should put Ribbon first on the shortlist. Agencies that already have candidates and want to send clients a deeper interview record should test Tracy alongside it. Our [AI interview tools for staffing agencies](/blog/ai-interview-tools-for-staffing-agencies?utm_source=blog&utm_medium=seo&utm_campaign=tracy-vs-ribbon&utm_content=internal-link) covers that use case. ## Where does Tracy win? Tracy is the cleaner choice when the team already has an applicant pool and cares most about what happens inside the first conversation. It supports video or audio, asks adaptive follow-ups, and can introduce a coding task before probing the candidate's trade-offs. The report keeps direct quotes next to the fields the reviewer needs. That narrower scope also makes the buying model easy to test. Three results are free. After that, one credit costs $1 and covers each started 20 minutes of a session that produces a result. Credits do not expire, and a session with no answers is not charged. There is no seat price in the published plan. For a technical hiring team, run the same role through both products. Include an answer that sounds plausible but omits a constraint. Compare the next question, the resulting evidence and the minutes a reviewer needs to reach a decision. The [AI technical interview guide](/blog/ai-technical-interviews?utm_source=blog&utm_medium=seo&utm_campaign=tracy-vs-ribbon&utm_content=internal-link) gives a task you can use. ## How does pricing compare? The accessible [Ribbon comparison article](https://www.ribbon.ai/blog/ribbon-vs-brighthire) supports the advertised entry package. It does not confirm current higher-tier prices or overages: | Pricing item | What is confirmed | |---|---| | Advertised entry price | $499/month, billed annually | | Entry allowance | 100 interviews per month | | Higher tiers, Enterprise and overage | Request a current quote | Tracy charges by completed result length. Up to 20 minutes costs one $1 credit; 21 to 40 minutes costs two credits; 41 to 60 minutes costs three. Read the full calculation in [AI interview cost per candidate](/blog/ai-interview-cost?utm_source=blog&utm_medium=seo&utm_campaign=tracy-vs-ribbon&utm_content=internal-link) or the plan-by-plan detail in [Ribbon pricing](/compare/ribbon-pricing?utm_source=blog&utm_medium=seo&utm_campaign=tracy-vs-ribbon&utm_content=internal-link). The sticker price does not settle the choice. Ribbon includes recruiting work that Tracy does not attempt. Tracy lets a team buy interview results without paying for seats, sourcing or outreach. Compare the cost of the workflow you will use, including the time left for a recruiter after the result arrives. ## Which should your team choose? Choose Ribbon when: - sourcing and talent rediscovery are part of the same purchase; - the team wants email, SMS or WhatsApp outreach beside interviews; - predictable monthly interview capacity matters; - ATS-connected recruiting operations are the main project. Choose Tracy AI Interviewer when: - candidates are already entering your funnel; - role-specific follow-ups carry more signal than a broad recruiting suite; - reviewers need direct quotes and the recording behind each result; - you want per-result pricing without a seat commitment. Before buying either, run two real roles and include awkward cases: a vague answer, a candidate who needs an alternative format, a restart and a low-confidence result. Review the evidence with the people who will make the hiring decision. ![A checklist of the main actions for tracy vs ribbon.](/blog/images/tracy-vs-ribbon-takeaways.png) ## Common questions ### What is the main difference between Tracy and Ribbon? Tracy AI Interviewer focuses on spoken video or audio interviews with adaptive follow-ups and summaries backed by the participant's own words. Ribbon covers a wider recruiting workflow that includes sourcing, rediscovery, outreach, interviews and ATS-connected operations. ### Which is better for staffing agencies? Ribbon is the stronger fit when an agency wants sourcing and outreach in the same platform. Tracy fits teams that already have candidates and want a deeper spoken interview before submitting them. ### Which product publishes pricing? Tracy publishes $1 per credit, with one credit covering each started 20 minutes of a completed result. Ribbon's comparison article advertises an annual-billing starting price of $499/month for 100 interviews. Its pricing page is unavailable, so current overages and higher tiers need confirmation. ### Can either product make the final hiring decision? No. Tracy says the decision stays with the customer. Ribbon says ranked summaries support recruiter review and the final decision remains with the human recruiter. ### Does Ribbon have a free trial? Ribbon's comparison article advertises a seven-day trial; confirm current terms. Tracy includes the first three interview results free. --- **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=tracy-vs-ribbon&utm_content=demo)** or **[Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=tracy-vs-ribbon&utm_content=try-free)** ## Sources All sources retrieved 2 Nov 2026. - Tracy AI Interviewer, [product page](https://tracyinterview.ai/): interview formats, adaptive follow-ups, technical tasks and evidence in reports. - Tracy AI Interviewer, [pricing](https://tracyinterview.ai/pricing): free results, credit price, session-length billing and no-answer policy. - Ribbon, [product page](https://www.ribbon.ai/): recruiting workflow, structured scoring, ATS sync and human decision statement. - Ribbon, [comparison with BrightHire](https://www.ribbon.ai/blog/ribbon-vs-brighthire): advertised $499/month annual-billing starting package, 100 interviews and seven-day trial; higher tiers and overages need confirmation. - Ribbon, [Ribbon vs BrightHire](https://www.ribbon.ai/blog/ribbon-vs-brighthire): product scope and buyer guidance, published 23 Aug 2026. # AI Interview for Staffing Agencies: 7 Tools Compared (Nov 2026) > Compare seven AI interview tools for staffing agencies. See how Tracy helps recruiters submit faster with candidate answers clients can check. - URL: https://tracyinterview.ai/blog/ai-interview-tools-for-staffing-agencies - Published: 2026-11-02 - Updated: 2026-11-11 - Author: Tracy team Fact-checked on 29 Oct 2026 against 15 sources ([see Sources](#sources)). > **In short** > - **Submit faster and with proof:** Tracy AI Interviewer lets applicants complete a browser-based first interview after hours, then gives recruiters a summary tied to the candidate's own words, video and transcript. > - **Go deeper on skilled roles:** Tracy follows up on vague answers, so clients can check how a candidate reasons. Recruiters can work from existing tools and pay per result. > - Choose HeyMilo for specific Bullhorn or JobDiva workflows, or a multilingual phone-screen tool for high-volume temp roles. **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=ai-interview-tools-for-staffing-agencies&utm_content=demo)** or **[Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=ai-interview-tools-for-staffing-agencies&utm_content=try-free)** **On this page** - [Why is Tracy the best AI interview for staffing agencies?](#why-is-tracy-the-best-ai-interview-for-staffing-agencies) - [How do AI recruiting tools for staffing agencies compare?](#how-do-ai-recruiting-tools-for-staffing-agencies-compare) - [When should an agency pick another tool?](#when-should-an-agency-pick-another-tool) - [What does an AI interview for staffing agencies cost per candidate?](#what-does-an-ai-interview-for-staffing-agencies-cost-per-candidate) - [How did we rank the tools?](#how-did-we-rank-the-tools) - [Questions](#questions) - [Sources](#sources) ## Why is Tracy the best AI interview for staffing agencies? An AI interview for staffing agencies has to make you faster and more credible at once. In Robert Half's survey of more than 2,000 US hiring managers, 65% said AI-enhanced resumes make skills harder to verify, so a fast submission also needs proof behind it. Tracy is built for both. ![Timeline: a req opens at 7 pm and the interview link goes to every applicant, applicants join in the browser overnight and Tracy follows up on their answers, and by 8 am the client submission goes out with each candidate's own words, video and transcript.](/blog/images/staffing-speed-to-submit.svg) *A req that opens in the evening can have an evidence-backed submission on the client's desk the next morning.* ### Applicants interviewed overnight, so you submit first Send the interview link to every applicant when the req opens. Each person opens it in the browser and joins with no sign-up, whenever it suits them, so the req you open at 7 pm can have finished interviews by morning without a recruiter on a single call. Tracy reports a 72% candidate completion rate across its interviews, so most applicants you invite finish the interview and reach your shortlist with answers. The interview goes deep enough to submit on. Tracy runs natural two-way video interviews with role-specific questions and follows up on vague answers, so your recruiter reads how a data engineer handled late events instead of a yes to "do you know Spark". For engineers it can introduce a coding task and probe the trade-offs behind an answer, as our [AI technical interviews](/blog/ai-technical-interviews?utm_source=blog&utm_medium=seo&utm_campaign=ai-interview-tools-for-staffing-agencies&utm_content=internal-link) guide shows. Tracy's interview integrity detection flags tab changes, abnormal eye tracking, live face detection, average delay before an answer, external screens and other unusual patterns, so your recruiter can check for a stand-in or a hidden helper before your agency's name goes on the submission. See [how to detect AI cheating in interviews](/blog/ai-interview-cheating-detection?utm_source=blog&utm_medium=seo&utm_campaign=ai-interview-tools-for-staffing-agencies&utm_content=internal-link). ### A summary you can forward, so the client trusts the submission ![A Tracy result report for a data analyst submitted to a client. Four fields, such as tools used daily and working with stakeholders, are marked Stated, each with the candidate's own words as evidence. The recording and timestamped transcript sit above them.](/blog/images/lib-result-traced-analyst.png) *The summary you forward to the client, each point in the candidate's own words (illustration with sample data).* A hiring manager who gets "4.2 out of 5, strong communicator" has nothing to check. Tracy turns each interview into a readable summary "with the participant's own words attached", with every answer organised around the fields you define and the video and transcript included. Direct quotes give context, and the AI explains each answer score. You paste the strongest answers into your submission, and the client reads what the candidate said. Your recruiter makes the call on who to submit. Example: part of a Tracy summary for a senior data engineer req, written up as a client submission. Illustration with test data, not a real candidate. | Field | What the candidate said | | --- | --- | | Pipeline at scale | "We moved the nightly batch to streaming. Late events broke the daily totals, so I added a watermark and a backfill job." | | Follow-up: trade-off | Asked what the watermark cost: "About 40 minutes of delay on the dashboard. Finance agreed, because the numbers stopped changing after they'd read them." | | Availability | "Four weeks' notice. Looking for contract, remote." | A fixed-question phone screen stops at the first row. ### Inside the tools you already use, so nothing new to roll out Tracy works inside tools your recruiters already use, such as Slack, WhatsApp and Workable, so they send interviews and read summaries where they already work. Tracy drafts the questions for you, or you add the questions that matter, so each client's req gets its own interview. Tracy is SOC 2 Type 2 compliant, with the report on request at tracy@qode.world, which answers the security review your enterprise clients will run. ## How do AI recruiting tools for staffing agencies compare? Rows are the four things Tracy is built around, plus price. Rival cells come from each vendor's own page, retrieved 29 Oct 2026. | | Tracy AI Interviewer | HeyMilo | Alex | Tenzo | Carv | Ribbon | Honeit | | --- | --- | --- | --- | --- | --- | --- | --- | | **Interviews like your best expert** | Follows up on vague answers; coding task with trade-off questions | "Scored two-way interviews" on video | "Personalized interviews with each candidate" | "Phone or video interviews 24/7 in any language" | Voice screen that "asks follow-up questions" | "Structured screening interviews by voice" | Recruiter-run calls with AI notes | | **Thousands of interviews, zero new hires** | Applicants join with no sign-up, whenever it suits them | Voice, video, SMS and forms | 30 languages; 1,000,000 candidates interviewed | 40+ languages, 24/7 | 99 languages | 10 languages | Priced per recruiter seat | | **Work with your stack** | ChatGPT, Claude, Gemini, Mistral, Grok, Slack, WhatsApp, Zapier, Workable, HubSpot | Bullhorn, Avionté, JobDiva, Greenhouse, Workday and more | "33+" ATS integrations, not named | "57+ integrations", list on request | Bullhorn, JobAdder, Avionte shown | "60+ ATS" | 1 ATS on Core, unlimited on Pro | | **Every score, traced to the answer** | Summary with the candidate's own words attached and each answer score explained by the AI, plus video and transcript | Report "with per-question rationale" | Not described | Not described | Notes and transcripts logged to your ATS | "Every interview is scored and ranked" | AI notes and scorecards | | **Price** | Published per credit; enterprise custom based on volume | Not published | Not published | Not published | Not published | $499 a month for 100 interviews, billed annually | $79/seat/month; Pre-Screener from $1.50/credit | ## When should an agency pick another tool? Tracy fits agencies that place skilled or technical people and need to prove depth to the client. These tools fit other desks better. - **HeyMilo.** Its staffing page promises "Faster candidate screening + submission", it names Bullhorn, Avionté and JobDiva as integrations, and each interview produces "a scored candidate report with per-question rationale". No public price. Pick HeyMilo instead if you want scored reports written into Bullhorn or JobDiva; pick Tracy if the client wants the candidate's own words behind the submission. - **Carv.** Voice AI that dials applicants, "Speaks 99 languages fluently", and logs notes and transcripts to your ATS; its staffing page shows Bullhorn and JobAdder. No public price. Pick Carv instead if you fill temp, logistics or contact center shifts by the hundred and a call that confirms availability is all the client needs. - **Ribbon.** Built for "warehouse, retail, hospitality and healthcare roles", by voice. Its price is $499 a month for 100 interviews, billed annually, per Ribbon's own comparison page (its pricing page didn't load on 29 Oct). Pick Ribbon instead if you place hourly workers and want a fixed monthly cost before you talk to sales. - **Honeit.** "An integrated talent delivery platform for recruiting, RPO, TA and staffing teams", with recruiter-led calls, AI notes and candidate presentations at $79 per recruiter seat a month. Pick Honeit instead if your recruiters want to keep running every screen call themselves. - **Alex.** Lists RPO and staffing ("submit shortlists faster"), and reaches candidates by phone, video, SMS and WhatsApp in 30 languages. No public price. Pick Alex instead if you run RPO programs and want one tool across every channel. - **Tenzo.** "Screens, interviews, and schedules candidates 24/7 in 40+ languages", with "40+ signals" for fraud and SOC 2 Type 2. No public price. Pick Tenzo instead if scheduling in many languages is your main bottleneck. ## What does an AI interview for staffing agencies cost per candidate? Agency volume swings by client, and many screens go nowhere, so cost per candidate matters more than the plan name. - **Tracy AI Interviewer:** $1 per credit; one credit covers up to 20 minutes, so a first-round screen is one credit. A session with no answers is never charged and credits never expire, so credits bought for a busy month carry over to a quiet one. For agency volume, enterprise pricing is custom, based on your volume. - **Ribbon:** $499 a month for 100 interviews, billed annually, per Ribbon's own comparison page. That's about $5 an interview if you use all 100. - **Honeit:** the AI Pre-Screener costs $1.50 to $2 a credit, plus $79 or $129 per recruiter seat each month. - **HeyMilo, Alex, Tenzo and Carv** don't publish prices. Ask each whether you pay for a no-show, and whether seats or a monthly minimum apply. ## How did we rank the tools? We picked AI interviewers that list staffing or RPO as a use and took only what each vendor publishes on its own pages. The ranking weights speed to submit, fit with many clients' systems, cost per candidate and a report the client can read, with the report weighted most for skilled placements. For in-house teams at large employers, see [HireVue alternatives](/compare/hirevue-alternatives?utm_source=blog&utm_medium=seo&utm_campaign=ai-interview-tools-for-staffing-agencies&utm_content=internal-link) and our list of the [best AI interview software for hiring teams](/blog/best-ai-interview-software?utm_source=blog&utm_medium=seo&utm_campaign=ai-interview-tools-for-staffing-agencies&utm_content=internal-link). ## Common questions ### Is interview AI staffing software worth it for a small agency? Yes, once applicants outnumber recruiter hours. A desk of three recruiters can't phone-screen 400 applicants for one req in a day. With Tracy every applicant can take the interview that evening, and you pay for results, not seats. ### How do agencies use AI in the recruitment process without losing the client? Put the AI interview at the first round and keep your recruiters on the shortlist and the client conversation. Every applicant gets the same first interview with Tracy, and your recruiter decides who to submit from the answers. ### Do AI interviewers integrate with Bullhorn? HeyMilo names Bullhorn as an integration, and Carv shows it among the systems it syncs with. Tracy connects through Zapier, Workable and HubSpot and works inside Slack and WhatsApp. Ask any vendor whether results write back to the candidate record or arrive as a link. ### Can I send the AI interview report to my client? Yes. Tracy's summary has the candidate's own words attached, with direct quotes, the interview video and transcript, so you forward the answers that support your submission. ### Will candidates think worse of my agency for using an AI interview? Candidates rate Tracy 4.7 out of 5 on average. Tell them up front that an AI runs the first interview and that a recruiter reviews every result. Our guide on how to screen candidates when there are too many applicants covers what candidates say they want from an AI interview. --- **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=ai-interview-tools-for-staffing-agencies&utm_content=demo)** or **[Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=ai-interview-tools-for-staffing-agencies&utm_content=try-free)** ## Sources All retrieved 29 Oct 2026. 1. Tracy AI Interviewer, [pricing](https://tracyinterview.ai/pricing): "$1 per credit, in USD. Credits never expire." / "You pay for results, not seats." / "1 credit for each started 20 minutes, charged when the result is ready" / "A session with no answers is never charged." 2. Tracy AI Interviewer, [tracyinterview.ai](https://tracyinterview.ai): "Tracy AI Interviewer runs natural two-way video interviews, follows up on vague answers, digs into the interesting ones" / "Each person opens a link in their browser and joins with no sign-up." / "People join in their browser whenever it suits them." / "Tracy AI Interviewer can ask role-specific questions, introduce a coding task, and probe the trade-offs behind an answer" / "Each interview becomes a readable summary with the participant's own words attached" / "Direct quotes for context" / "your own repeatable interview, with every answer organized around the fields you define." / "Your questions are drafted for you." and "add the questions that matter" 3. Qode, [qode.world/tracy](https://qode.world/tracy): "4.7/5 average candidate rating across interviews" / "72% candidate completion rate for Tracy interviews" 4. Qode, [qode.world/pricing](https://qode.world/pricing): "Interview integrity detection: Tab changes, Abnormal eye tracking, Live face detection, Average delay time before answer, External screens, Other unusual patterns" / "Video & chapter breakdown", "Transcript", "Answer summary & score", "Explain answer score by AI Agent" / Enterprise: "Custom based on volume". 5. Qode, 28 Oct 2026: enterprise pricing is custom, based on volume. 6. Qode, 29 Oct 2026: Tracy works inside ChatGPT, Claude, Gemini, Mistral, Grok, Slack, WhatsApp, Zapier, Workable and HubSpot; SOC 2 Type 2, report on request at tracy@qode.world. 7. HeyMilo, [heymilo.ai](https://www.heymilo.ai): "Screen and interview every applicant with structured AI voice, video, SMS, and forms." / "Faster candidate screening + submission" / "connects to major ATS platforms including Greenhouse, Bullhorn, Avionté, Ashby, Workday, SmartRecruiters, and JobDiva" / "Every interview produces a scored candidate report with per-question rationale" / "Scored two-way interviews" / pricing: the only call to action is "Book a Demo". 8. Alex, [alex.com](https://www.alex.com): "Handle more requisitions, screen every inbound applicant, and submit shortlists faster." / "Let Alex call thousands of candidates instantly" / "Engage candidates with intelligent text conversations" / "Conduct seamless interviews with your global candidates" (WhatsApp) / "33+ officially supported ATS integrations" / "1,000,000 candidates interviewed" / "30 languages supported" / "Alex conducts personalized interviews with each candidate" / no pricing on the page. 9. Tenzo, [tenzo.ai](https://tenzo.ai): "screens, interviews, and schedules candidates 24/7 in 40+ languages" / "Phone calls, WhatsApp, video, Email." / "AI powered phone or video interviews 24/7 in any language" / "Tenzo AI uses 40+ signals to route out mis-representation, cheating, and stolen identities." / "57+ integrations out of the box. Contact our team for the full list." / "SOC 2 Type 2 & GDPR Compliant" / pricing: "No commitment. Just tell us a bit about your business." 10. Carv, [Voice AI](https://www.carv.com/use-cases/voice-ai): "Run your phone screens on autopilot." / "Speaks 99 languages fluently." / "The agent dials out, has a real conversation, and qualifies the candidate." / "asks follow-up questions" / "everything is automatically logged into your ATS the moment the call ends" 11. Carv, [Staffing](https://www.carv.com/industries/staffing): "syncs every tidbit of information with your ATS automatically." (page shows Bullhorn, JobAdder, Avionte, Workday, Greenhouse and SmartRecruiters logos; no pricing on the page) 12. Ribbon, [ribbon.ai](https://www.ribbon.ai): "Ribbon screens, interviews and ranks candidates for warehouse, retail, hospitality and healthcare roles" / "conducts structured screening interviews by voice" / "Every interview is scored and ranked." / "Syncs with 60+ ATS out of the box." / "Works on any phone, in 10 languages" 13. Ribbon, [Ribbon vs BrightHire](https://ribbon.ai/blog/ribbon-vs-brighthire): "Public annual-billing packages start at $499/month for 100 interviews" (ribbon.ai/pricing did not load on 29 Oct 2026) 14. Honeit, [pricing](https://www.honeit.com/pricing): "Honeit is an integrated talent delivery platform for recruiting, RPO, TA and staffing teams." / "VoIP and video calls, AI notes and candidate presentations for recruiters who move fast." / "AI notes and structured scorecards" / "per recruiter seat / month, billed quarterly" ($79 Core, $129 Pro) / "API and unlimited ATS integrations" (Pro) / "1 credit = 1 AI pre-screen" / "100 credits = $200" / "1,000 credits = $1,500" 15. Robert Half, [survey release, 10 Mar 2026](https://press.roberthalf.com/2026-03-10-Robert-Half-survey-67-of-HR-leaders-report-AI-generated-applications-are-slowing-hiring): "65% of hiring managers say AI-enhanced resumes make skills harder to verify." / "The survey contains responses from more than 2,000 hiring managers in the United States." # Best AI Interview Software for Hiring Teams (Nov 2026) > The best AI interview software for hiring teams, ranked on follow-up depth, scale, stack fit and traced scores. 8 tools, facts checked 29 Oct 2026. - URL: https://tracyinterview.ai/blog/best-ai-interview-software - Published: 2026-11-02 - Updated: 2026-11-13 - Author: Tracy team > **In short** > - **Best for skilled and technical first-round hiring among the eight tools reviewed: Tracy AI Interviewer.** It interviews every applicant, probes vague answers and lets reviewers check each score against the candidate's own words. > - Tracy meets all four criteria in our comparison: adaptive depth, scale, stack fit and evidence-backed scores. > - Choose HireVue for a full enterprise suite or FedRAMP, CodeSignal for a test-first technical workflow, and Ribbon for hourly roles. **On this page** - [What should hiring teams look for in the best AI interview software?](#what-should-hiring-teams-look-for-in-the-best-ai-interview-software) - [Why is Tracy the best AI interview software for skilled roles?](#why-is-tracy-the-best-ai-interview-software-for-skilled-roles) - [How do the 8 best AI interview tools compare?](#how-do-the-8-best-ai-interview-tools-compare) - [Which of the other AI interview tools fits your team?](#which-of-the-other-ai-interview-tools-fits-your-team) - [How did we rank them?](#how-did-we-rank-them) - [Questions](#questions) - [Sources](#sources) ## What should hiring teams look for in the best AI interview software? AI-written applications now read alike, so the CV no longer separates the engineer who built the system from the one who described it well. The interview has to do that, for every applicant, before a hiring manager spends an hour on a call. We judged each tool on four things: 1. Interviews like your best expert. Does it ask the follow-up that gets to the real story, in the candidate's own field? 2. Thousands of interviews, zero new hires. Can every applicant take it without anyone scheduling it? 3. Work with your stack. Does it fit the tools you run, with your own questions? 4. Every score, traced to the answer. Can the reviewer check a score against what the candidate said, and does a person decide? ## Why is Tracy the best AI interview software for skilled roles? Tracy follows up on vague answers and digs into the interesting ones, "a real follow-up, not a rigid script", so a rehearsed or AI-fed first answer doesn't carry a candidate through. It can "introduce a coding task, and probe the trade-offs behind an answer", so you see how an engineer reasons about their own code. Our guide to [AI technical interviews](/blog/ai-technical-interviews?utm_source=blog&utm_medium=seo&utm_campaign=best-ai-interview-software&utm_content=internal-link) shows a full session. **Example: a Tracy follow-up chain for a data analyst.** An illustration with test data, not a real candidate. 1. **Tracy:** Tell me about a dataset you cleaned that changed a decision. 2. **Candidate:** Churn data for a subscription product. About 8% of rows had missing cancel dates, so I dropped them. 3. **Tracy:** Who were those 8%? Were they missing at random? 4. **Candidate:** No. Most were annual plans that hadn't renewed yet. Dropping them made churn look higher. 5. **Tracy:** What did you do instead, and how did the number change? 6. **Candidate:** I treated them as still active until their renewal date. Monthly churn went from 6.1% to 4.8%, and the team stopped a discount campaign. The second answer passes most checklists. The fourth and sixth show the analyst understands bias in missing data and can tie it to a business decision. ![A Tracy result report for a customer support lead. Four fields, such as handling escalations and start date, are marked Stated, each with the candidate's own words as evidence. The recording and timestamped transcript sit above them.](/blog/images/lib-result-traced-support-lead.png) *Every point in a Tracy report is the candidate's own words, next to the recording and transcript (illustration with sample data).* Each report has a summary with "the participant's own words attached" and direct quotes for context, plus the interview video, transcript and chapter breakdown. The AI explains each answer score, so a hiring manager can check it against the answer before deciding. Tracy's interview integrity detection flags tab changes, abnormal eye tracking, live face detection, average delay before an answer, external screens and other unusual patterns, so a suspicious session gets a second look. See [how to spot AI cheating in interviews](/blog/ai-interview-cheating-detection?utm_source=blog&utm_medium=seo&utm_campaign=best-ai-interview-software&utm_content=internal-link). Candidates join in their browser with no sign-up, "whenever it suits them", so you can [screen every candidate in depth](/blog/how-to-screen-candidates?utm_source=blog&utm_medium=seo&utm_campaign=best-ai-interview-software&utm_content=internal-link) without adding interviewers, and candidates rate Tracy 4.7 out of 5 on average. Tracy works inside tools your team already uses, such as Slack and Workable. The price is $1 per credit; one credit covers up to 20 minutes. You pay for results, not seats, and a session with no answers is never charged. At high volume, enterprise pricing is custom, based on your volume. **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=best-ai-interview-software&utm_content=demo)** or **[Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=best-ai-interview-software&utm_content=try-free)** ## How do the 8 best AI interview tools compare? ![Grid of 8 AI interview tools marked yes, partly, no or not stated on four pillars. Tracy AI Interviewer, in the first row, is the only tool marked yes on all four.](/blog/images/best-ai-interview-software-scores.svg) *Tracy is the only tool here marked yes on all four pillars. Marks are our reading of each vendor's own pages, 29 Oct 2026.* | Tool | Best for | What stands out | Published price | |---|---|---|---| | 1. Tracy AI Interviewer | Skilled and technical roles | Follows up on vague answers; an assessment per question, in the candidate's own words | Yes | | 2. HireVue | One enterprise suite | Voice AI Interviewer plus assessments and live video; FedRAMP | No | | 3. CodeSignal | Teams already on its assessments | AI interviews on the same credits as tests | Yes | | 4. HeyMilo | Many languages and channels | Voice, video, SMS and forms in 20+ languages | No | | 5. Talview | High-stakes identity checks | Proctoring and deepfake detection | No | | 6. Sapia.ai | Employers hiring 500+ a year | Priced per hire | No | | 7. Ribbon | Hourly and frontline roles | Phone voice screens, 60+ ATS | Yes, via its own blog | | 8. Spark Hire | One-way and live video only | No AI follow-up; 40+ ATS integrations | Yes | Every rival fact below links to the vendor's own page in the Sources. ## Which of the other AI interview tools fits your team? ### 2. HireVue HireVue's AI Interviewer runs voice interviews with "probing follow-up questions", scores skill answers on a "Custom 1-5 rubric", and puts scores in your ATS. The reviewer gets a score that "explains why". Around it sit assessments, one-way video and live interviews. It holds ISO/IEC 27001:2022 and FedRAMP authorization, and doesn't publish prices (see [what HireVue costs](/compare/hirevue-pricing?utm_source=blog&utm_medium=seo&utm_campaign=best-ai-interview-software&utm_content=internal-link)). Choose HireVue for one vendor from application to offer, or for US government work. If you need the interview step done in depth beside your suite, Tracy fits there. Read [Tracy vs HireVue](/compare/tracy-vs-hirevue?utm_source=blog&utm_medium=seo&utm_campaign=best-ai-interview-software&utm_content=internal-link), or [HireVue alternatives](/compare/hirevue-alternatives?utm_source=blog&utm_medium=seo&utm_campaign=best-ai-interview-software&utm_content=internal-link) if you're leaving it. ### 3. CodeSignal CodeSignal's AI interviewers probe "if answers are thin". The reviewer gets "an in-depth report with an overall score, skill evaluation, and key insights", with "cheating & fraud flags". The Build plan is $79 a month billed annually (60 annual credits) or $99 billed monthly (5 monthly credits), shared with its assessments. It's the natural pick if your engineers already run CodeSignal assessments. If you want the coding task inside a conversation that asks about trade-offs, Tracy does that. ### 4. HeyMilo HeyMilo interviews "with structured AI voice, video, SMS, and forms", "24/7 in 20+ languages". Its reviewer gets "a scored candidate report with per-question rationale", and HeyMilo provides "a trust score that indicates the likelihood of cheating". Pricing isn't published. HeyMilo suits teams hiring in many languages or running SMS screens. For skilled roles where the follow-up decides, Tracy goes deeper. Agencies should also read [AI interview tools for staffing agencies](/blog/ai-interview-tools-for-staffing-agencies?utm_source=blog&utm_medium=seo&utm_campaign=best-ai-interview-software&utm_content=internal-link). ### 5. Talview Talview's interviewer runs "conversational screening with real-time follow-ups", and "every score shows response behind it". It adds "continuous liveness and face match", a "secondary-camera view of the desk and the room" and a "secure browser with code jams, whiteboarding and structured evaluations". Pricing is unit-based, with "a minimum commitment of approximately 150 to 200 candidates". Talview fits proctored, high-stakes tests. If you want pay-per-result pricing with no minimum, Tracy charges only for interviews that produce answers. ### 6. Sapia.ai Sapia.ai prices "on a per-hire basis", and is "designed to best serve companies that hire more than 500 people per year." Its pages didn't describe what the reviewer sees. Sapia.ai fits large employers that want cost tied to hires. Smaller skilled teams get a published price with Tracy. ### 7. Ribbon Ribbon "screens, interviews and ranks candidates for warehouse, retail, hospitality and healthcare roles" by voice, on "any phone, in 10 languages", and "syncs with 60+ ATS". Reviewers get "live interview captions and recordings". Pricing is $499 a month for 100 interviews, billed annually, per Ribbon's own comparison page; its pricing page didn't load on 29 Oct 2026. Ribbon is built for hourly hiring by phone. For engineers, analysts and specialists, Tracy's follow-ups fit better. ### 8. Spark Hire Spark Hire includes "unlimited live and one-way video interviews" from $249 a month billed annually, and lists "40+ ATS integrations". A one-way video has no follow-up, so the reviewer watches each recording. Spark Hire suits teams that want to watch every candidate themselves. If you'd rather read answers with follow-ups already asked, Tracy saves that viewing time. ## How did we rank them? We selected eight tools covering conversational interviews, technical assessments, full hiring suites and one-way video. We read each vendor's own product and pricing pages on 29 Oct 2026 and took every fact from them. For skilled roles we prioritized depth of follow-up, then evidence behind scores, scale and stack fit. Price transparency helped distinguish otherwise similar options; HireVue ranks above CodeSignal because it combines a conversational interviewer with a broader hiring suite. These categories serve different buyers, so use each tool's best-for description before comparing rank. For expert vetting, the same test applies; see our [interview questions for AI trainers and domain experts](/blog/ai-trainer-interview-questions?utm_source=blog&utm_medium=seo&utm_campaign=best-ai-interview-software&utm_content=internal-link). ## Common questions ### What is the best AI interviewer for technical roles? We rank Tracy first, because it follows up on vague answers and can probe the trade-offs behind a coding task. CodeSignal is the pick if your team already runs its assessments. ### How much does an AI interview platform cost? Tracy, CodeSignal and Spark Hire publish prices, and Ribbon states one on its own blog. HireVue, HeyMilo, Talview and Sapia.ai quote on request. ### Which AI interview tools for recruiters work with an ATS? HireVue puts scores in the ATS, HeyMilo names seven ATS platforms, Ribbon says 60+ and Spark Hire 40+. Tracy works inside named tools, including Workable and HubSpot. ### Do AI interview tools detect cheating? Several flag it. Tracy flags tab changes, abnormal eye tracking, live face detection, average delay before an answer, external screens and other unusual patterns. HeyMilo, Talview, CodeSignal and HireVue also describe cheating or fraud features. A person should review any flag. ### Who makes the hiring decision? Your team does. A Tracy report gives each answer a written assessment with its reasoning and the candidate's own words, and your reviewer chooses who goes to the next round. HireVue says "Recruiters always make the final call." Tell applicants an AI runs the interview and that a person decides. ### Is a one-way video tool enough? It is if your team wants to watch every recording. If you need follow-up questions, choose a tool that runs a two-way interview. --- **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=best-ai-interview-software&utm_content=demo)** or **[Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=best-ai-interview-software&utm_content=try-free)** ## Sources All retrieved 29 Oct 2026. - Tracy, [tracyinterview.ai](https://tracyinterview.ai): "Tracy AI Interviewer runs natural two-way video interviews, follows up on vague answers, digs into the interesting ones"; "A real follow-up, not a rigid script."; "Tracy AI Interviewer can ask role-specific questions, introduce a coding task, and probe the trade-offs behind an answer"; "Each interview becomes a readable summary with the participant's own words attached"; "Direct quotes for context"; "Each person opens a link in their browser and joins with no sign-up."; "People join in their browser whenever it suits them." - Tracy, [pricing](https://tracyinterview.ai/pricing): "$1 per credit, in USD." "A result costs 1 credit for every 20 minutes of the session, started minutes included." "You pay for results, not seats." "A session with no answers is never charged." - Qode, [qode.world/tracy](https://qode.world/tracy): "4.7/5 average candidate rating across interviews". - Qode, [qode.world/pricing](https://qode.world/pricing): "Interview integrity detection: Tab changes, Abnormal eye tracking, Live face detection, Average delay time before answer, External screens, Other unusual patterns". "Video & chapter breakdown", "Transcript", "Answer summary & score", "Explain answer score by AI Agent". Enterprise: "Custom based on volume". - Qode, 28 Oct 2026: enterprise pricing is custom, based on volume. - Qode, 29 Oct 2026: Tracy works inside ChatGPT, Claude, Gemini, Mistral, Grok, Slack, WhatsApp, Zapier, Workable and HubSpot. - HireVue, [AI Interviewer](https://www.hirevue.com/platform/ai-interviewer): "AI agents ask and record candidate responses, including probing follow-up questions, in a natural two-way conversation." "AI evaluates each response against your role-specific criteria, highlights whether it meets them, and explains why." Also "Custom 1-5 rubric", "Recruiters always make the final call", "Recruiters view scores in the ATS and quickly decide the next steps". - HireVue, [AI Interviewer launch announcement](https://www.hirevue.com/press-release/hirevue-launches-ai-interviewer), 18 Jun 2026: "dynamic, two-way voice interviews with every applicant 24/7". - HireVue, [Pricing](https://www.hirevue.com/pricing): "From mid-market to enterprise our platform is built for your specific hiring needs." - HireVue, [Enterprise](https://www.hirevue.com/solutions/enterprise): "Hirevue is certified to ISO/IEC 27001:2022 for Information Security Management"; "authorization by FedRAMP PMO." Products: "Match and Apply", "AI Interviewer", "Assessments", "Video Interviewing". - HireVue, [Video Interviewing](https://www.hirevue.com/platform/video-interviewing): "One-way asynchronous interviewing giving candidates 24/7 flexibility"; "Meet in Hirevue, Zoom, Teams, or Webex"; "Built-in cheating mitigation, advanced fraud detection, and ID verification". - CodeSignal, [AI Interviewer](https://codesignal.com/ai-interviewer/): "If answers are thin, it probes and asks follow up questions based on your criteria." "Hiring teams receive an in-depth report with an overall score, skill evaluation, and key insights." "Structured scores, cheating & fraud flags, and skill reports you can act on." - CodeSignal, [Pricing](https://codesignal.com/pricing/): Build "$79 /mo", "Billed annually", 60 annual credits; "$99 /mo", "Billed monthly", 5 monthly credits; "a credit is used when a candidate starts an assessment, AI interview, or live interview." - HeyMilo, [heymilo.ai](https://www.heymilo.ai/): "Screen and interview every applicant with structured AI voice, video, SMS, and forms." "Operates 24/7 in 20+ languages." "Every interview produces a scored candidate report with per-question rationale, competency breakdowns". "We provide a trust score that indicates the likelihood of cheating." "We utilize an active proctoring layer, an AI-classifier, and more to detect cheating." FAQ names Greenhouse, Bullhorn, Avionté, Ashby, Workday, SmartRecruiters and JobDiva. - Talview, [talview.com](https://www.talview.com/): "Conversational screening with real-time follow-ups, multilingual by design." "Never poses as human recruiter; every score shows response behind it." "Continuous liveness and face match, not a one-time check at login." "Secondary-camera view of the desk and the room." "Secure browser with code jams, whiteboarding and structured evaluations." "Detect deepfakes, AI teleprompters, fake credentials, AI generated claims, proxies, and hiring fraud." - Talview, [Pricing](https://www.talview.com/pricing): "Our pricing model is unit-based." "A minimum commitment of approximately 150 to 200 candidates is typically required." - Sapia.ai, [Pricing](https://sapia.ai/pricing/): "Sapia.ai's pricing structure works on a per-hire basis." "Our technology is designed to best serve companies that hire more than 500 people per year." - Ribbon, [ribbon.ai](https://www.ribbon.ai/): "Ribbon screens, interviews and ranks candidates for warehouse, retail, hospitality and healthcare roles." "Works on any phone, in 10 languages." "Syncs with 60+ ATS out of the box." "Live interview captions and recordings." - Ribbon, [Ribbon vs BrightHire](https://ribbon.ai/blog/ribbon-vs-brighthire), 23 Aug 2026: "Public annual-billing packages start at $499/month for 100 interviews". Ribbon's pricing page (ribbon.ai/pricing) returned 404 on 29 Oct 2026. - Spark Hire, [Pricing](https://www.sparkhire.com/pricing/): "Starting at $249 / month", billed annually; "Unlimited live and one-way video interviews"; "Access 40+ ATS integrations". # HireVue Alternatives for Hiring Teams (Nov 2026) > HireVue alternatives compared: Tracy AI Interviewer, Ribbon, HeyMilo, Alex, Paradox, Sapia.ai, Spark Hire and Willo, by interview depth and price. - URL: https://tracyinterview.ai/compare/hirevue-alternatives - Published: 2026-11-02 - Updated: 2026-11-11 - Author: Tracy team > **In short** > - **Best HireVue alternative for skilled roles: Tracy AI Interviewer.** It interviews every applicant, follows up beyond rehearsed answers and gives reviewers evidence in the candidate's own words at a published per-result price. > - Choose a frontline tool such as Ribbon or Paradox for hourly hiring, or Spark Hire or Willo for simple one-way video. > - **Choose Tracy when the follow-up can change the hiring decision.** **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=hirevue-alternatives&utm_content=demo)** or **[Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=hirevue-alternatives&utm_content=try-free)** ## Contents - [Why is Tracy the HireVue alternative for skilled hiring?](#why-is-tracy-the-hirevue-alternative-for-skilled-hiring) - [Why do teams look for HireVue alternatives?](#why-do-teams-look-for-hirevue-alternatives) - [Which HireVue alternatives fit which team?](#which-hirevue-alternatives-fit-which-team) - [What do the other HireVue alternatives do well?](#what-do-the-other-hirevue-alternatives-do-well) - [How do the HireVue alternatives compare?](#how-do-the-hirevue-alternatives-compare) - [How do you choose AI interview software to replace HireVue?](#how-do-you-choose-ai-interview-software-to-replace-hirevue) - [Questions](#questions) - [Sources](#sources) ## Why is Tracy the HireVue alternative for skilled hiring? Tracy is built for roles where the second and third questions decide: engineers, specialists and domain experts. Every fact about the other tools comes from their own pages, read on 29 Oct 2026. **Interviews like your best expert.** Tracy runs natural two-way video interviews and follows up on vague answers, so a candidate who rehearsed the first answer still has to explain how they did the work. HireVue's AI Interviewer scores answers on formats such as a "Custom 1-5 rubric", which suits jobs with clear right answers. Tracy suits jobs where you need to hear how the person reasons. ![A candidate in a Tracy interview in the browser. Tracy's caption asks a follow-up about the answer just given: "You refunded before the invoice was fixed. How did you stop the same error reaching other customers?" A side panel quotes that answer.](/blog/images/lib-live-interview-support-follow-up.png) *Tracy's next question builds on the answer the candidate just gave (illustration with sample data).* Tracy's own site gives the example of a backend engineer asked to design a rate limiter, with follow-ups on concurrency, failure modes and testing. Here is how that chain can run. Example: an illustration, not a real candidate. 1. Tracy: "Design a rate limiter for a public API. Walk me through it." 2. The candidate describes a token bucket per API key, stored in Redis. 3. Tracy: "Two requests for the same key arrive on different servers at the same moment. What happens?" 4. The candidate explains the race and moves the check into an atomic script. 5. Tracy: "Redis goes down for a minute. Do you let traffic through or block it?" 6. The candidate chooses to fail open for paying customers and closed for anonymous traffic, and says why. A candidate who memorised "token bucket" passes step 2 and struggles by step 5. Tracy can also "introduce a coding task, and probe the trade-offs behind an answer", so your engineers spend their hour only on people who can reason about code; [AI technical interviews](/blog/ai-technical-interviews?utm_source=blog&utm_medium=seo&utm_campaign=hirevue-alternatives&utm_content=internal-link) shows a full one. Each interview reports integrity signals (tab changes, abnormal eye tracking, live face detection, average delay before an answer, external screens and other unusual patterns), so you can check that the person who passed did it alone. [AI interview cheating detection](/blog/ai-interview-cheating-detection?utm_source=blog&utm_medium=seo&utm_campaign=hirevue-alternatives&utm_content=internal-link) explains each signal. **Every score, traced to the answer.** Each interview becomes a summary "with the participant's own words attached" and direct quotes for context, next to the video and transcript. The AI explains each answer score, so a reviewer can check a strong answer or catch a thin one and defend the shortlist. Candidates rate Tracy 4.7 out of 5 on average, and they join in the browser with no sign-up, whenever it suits them, so the deeper interview doesn't turn good applicants away. You pay for results, not seats, and a session with no answers is never charged, so you can [screen every candidate in depth](/blog/how-to-screen-candidates?utm_source=blog&utm_medium=seo&utm_campaign=hirevue-alternatives&utm_content=internal-link) instead of only those whose CV stood out. Tracy also works for [expert vetting](/blog/ai-trainer-interview-questions?utm_source=blog&utm_medium=seo&utm_campaign=hirevue-alternatives&utm_content=internal-link) and [accelerator interviews](/blog/accelerator-interview?utm_source=blog&utm_medium=seo&utm_campaign=hirevue-alternatives&utm_content=internal-link). ## Why do teams look for HireVue alternatives? HireVue is a full suite: an AI Interviewer, one-way and live video, and security credentials up to FedRAMP. Teams look elsewhere for narrower reasons. - **The price is hidden.** HireVue's pricing page offers a demo and an ROI calculator. Buyers on Vendr report a $20,900 median a year, and Vendr's buyer guide puts the maximum at about $145,000, on contracts that renew unless you give 60 days' notice. Our [HireVue pricing](/compare/hirevue-pricing) guide sets every figure against per-interview pricing. - **It's hard to change.** A G2 reviewer quoted in Willo's review wrote that "HireVue can be difficult to adapt when workflows change." Willo also notes HireVue "is not available in Zapier's app library." - **The AI interview is a scored voice screen.** HireVue describes "two-way voice interviews" scored against your criteria, which tells you whether expected points were covered more than how the person thinks. ## Which HireVue alternatives fit which team? ![A two by two map of HireVue alternatives. Across: quote only to published price. Up: fixed questions to follow-ups in the candidate's field. Tracy AI Interviewer sits at top right. HireVue sits top left. HeyMilo, Alex, Sapia.ai and Paradox sit bottom left. Ribbon sits just below the middle on the published-price side. Spark Hire and Willo sit bottom right.](/blog/images/hirevue-alternatives-matrix.svg) *Tracy pairs follow-ups in the candidate's field with a price you can see before a sales call.* | Tool | Best for | | --- | --- | | Tracy AI Interviewer | Skilled and technical roles, expert vetting and founder applications, where the follow-up decides | | Ribbon | Warehouse, retail, hospitality and healthcare roles screened by phone | | HeyMilo | High-volume teams that want voice, video, SMS and forms in 20+ languages | | Alex | Teams that want phone, video, SMS and WhatsApp interviews in 30 languages | | Paradox | Frontline hiring with self-serve interview scheduling in 100+ languages | | Sapia.ai | Employers hiring 500+ people a year who want to pay per hire | | Spark Hire | Small and mid-sized teams that want an ATS with one-way video at a published price | | Willo | Small teams that want one-way video paid by the role | ## What do the other HireVue alternatives do well? ### Ribbon Ribbon "screens, interviews and ranks candidates for warehouse, retail, hospitality and healthcare roles." It runs voice interviews that work "on any phone, in 10 languages, with follow-ups", and "Syncs with 60+ ATS out of the box". Pricing is $499 a month for 100 interviews, billed annually, per Ribbon's own comparison page; its pricing page didn't load for us on 29 Oct 2026. It advertises a 7-day free trial. - Pick Ribbon instead if you fill frontline roles at volume and candidates are easiest to reach by phone. ### HeyMilo HeyMilo is "The AI Recruiter for teams swamped by applications." It screens by "structured AI voice, video, SMS, and forms", "Operates 24/7 in 20+ languages", gives a "trust score that indicates the likelihood of cheating", and names Bullhorn and JobDiva among its ATS connections. Pricing is by demo. - Pick HeyMilo instead if you need many channels and languages, or you're an agency on Bullhorn; see [AI interview tools for staffing agencies](/blog/ai-interview-tools-for-staffing-agencies). ### Alex Alex, from Apriora Inc., interviews by phone, video, SMS and WhatsApp, lists 30 languages and "33+ officially supported ATS integrations", and "scores every application for fraud risk across email, phone, LinkedIn, location and more." Pricing is by demo. - Pick Alex instead if your candidates answer a phone call or WhatsApp sooner than a video link. ### Paradox Paradox's assistant, Olivia, offers "Mobile-first automation for frontline workers and managers", lets candidates "choose the best time slot that works for everyone", and replies "in 100+ languages." Paradox for Workday handles "Candidate communication directly through SMS on Workday." Pricing is by demo. - Pick Paradox instead if you hire hourly staff at scale and scheduling is the bottleneck. ### Sapia.ai Sapia.ai's pricing "works on a per-hire basis": you "interview and assess as many people as you want, without limits", and it "is designed to best serve companies that hire more than 500 people per year." - Pick Sapia.ai instead if you hire 500+ people a year and would rather pay per hire than per interview. ### Spark Hire Spark Hire offers "Unlimited live and one-way video interviews" and helps "surface the right candidates early with AI scoring". "More than 10,000 organizations have used Spark Hire." Video interviews start at $249 a month billed annually, or $299 a month on the monthly plan. - Pick Spark Hire instead if a recorded answer to set questions tells you enough and you want a visible price. ### Willo Willo sells one-way video by the role. Willo Lite is "£49 / live role", and you "Pay only for the live roles you use each month", with 150 assessed candidates per role and AI summaries. Willo Enterprise starts "From £2,999 / year", with "50+ native integrations" and Willo Intelligence tools that include follow-up questions. - Pick Willo instead if you have a few open roles at a time and want to pay only in months you hire. ## How do the HireVue alternatives compare? The rows are Tracy's four pillars, plus price. Rival cells come from each vendor's own pages, read on 29 Oct 2026. | | Tracy AI Interviewer | HireVue | Multi-channel screens (Paradox, Ribbon, HeyMilo, Alex) | Pay per hire (Sapia.ai) | One-way video (Spark Hire, Willo) | | --- | --- | --- | --- | --- | --- | | Interviews like your best expert | Natural two-way video or audio interviews; follows up on vague answers; coding tasks that probe the trade-offs behind an answer | Two-way voice interview with "probing follow-up questions", scored on a rubric | Voice, phone, SMS or chat screens in 10 to 100+ languages; Ribbon adds follow-ups | Interview format not stated on its pricing page | Recorded answers to set questions; Willo Enterprise adds follow-up question tools | | Thousands of interviews, zero new hires | People join "whenever it suits them", with no sign-up; you pay for results, not seats | 24/7 voice interviews on an annual contract | Built for frontline volume (HeyMilo: "1.4M+ candidates screened") | Unlimited interviews, paid per hire | Unlimited interviews (Spark Hire); 150 candidates per role (Willo Lite) | | Work with your stack | ChatGPT, Claude, Gemini, Mistral, Grok, Slack, WhatsApp, Zapier, Workable, HubSpot | ATS; not in Zapier, per Willo | Ribbon 60+ ATS; Alex 33+; HeyMilo names 7; Paradox inside Workday | Not stated on its pricing page | Spark Hire 40+ ATS; Willo 50+ on Enterprise | | Every score, traced to the answer | Summary with the candidate's own words and direct quotes, an AI explanation of each answer score, plus video and transcript | Scores in the ATS, with an explanation of each | HeyMilo trust score; Alex fraud-risk score | Not stated on its pricing page | Willo AI summaries; Spark Hire AI scoring | | Price | $1 per credit; one credit covers up to 20 minutes. Enterprise: custom, based on your volume | By quote; $20,900 median reported | Ribbon $499 a month for 100 interviews, billed annually, per its own comparison page; others by demo | Per hire, by quote | From $249 a month (Spark Hire); from £49 per live role (Willo) | ## How do you choose AI interview software to replace HireVue? 1. **Where are your candidates?** Hourly applicants on phones point to Paradox, Ribbon, HeyMilo or Alex. Professionals at a laptop can take a video interview, which opens up Tracy, Spark Hire and Willo. 2. **What decides the hire?** If a recorded answer to a set question is enough, one-way video is cheapest. If the follow-up decides, you need an interviewer that asks one, and that's Tracy. 3. **How steady is your volume?** Volume that moves with hiring cycles fits per-result pricing (Tracy) or per-role pricing (Willo) better than an annual contract. 4. **What can you test first?** Tracy charges per result and its credits never expire, and Ribbon has a 7-day trial. Run one role next to your current first round and compare what your reviewers would have decided. For the head-to-head, read [Tracy vs HireVue](/compare/tracy-vs-hirevue?utm_source=blog&utm_medium=seo&utm_campaign=hirevue-alternatives&utm_content=internal-link); for a ranking by segment, the [best AI interview software for hiring teams](/blog/best-ai-interview-software?utm_source=blog&utm_medium=seo&utm_campaign=hirevue-alternatives&utm_content=internal-link). ## Common questions ### What is the best HireVue alternative? For skilled and technical roles where follow-ups matter, Tracy AI Interviewer. For hourly hiring with scheduling, Paradox. For low-cost one-way video, Spark Hire or Willo. ### HireVue vs Tracy: which asks better follow-ups? HireVue's AI Interviewer asks "probing follow-up questions" by voice and scores answers against your criteria. Tracy follows up on vague answers on video or audio, can set a coding task and probe the trade-offs behind an answer, and attaches the candidate's own words to the summary, which matters most for technical and specialist roles. ### Which HireVue alternatives publish their prices? Tracy, Spark Hire and Willo show a price on their own pricing pages. Ribbon states a starting price on its own comparison page. HeyMilo, Alex, Paradox and Sapia.ai sell by demo or quote. ### Is there a HireVue alternative without an annual contract? Yes. Tracy charges per result and credits never expire. Willo Lite is paid per live role each month, and Spark Hire has a monthly video interview plan. ### Can I switch from HireVue mid-contract? Check your renewal date first; a buyer note on Vendr says HireVue needs 60 days' notice to opt out. You can run an alternative on one role next to HireVue before then. --- **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=hirevue-alternatives&utm_content=demo)** or **[Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=hirevue-alternatives&utm_content=try-free)** ## Sources All retrieved 29 Oct 2026. 1. HireVue, Pricing. https://www.hirevue.com/pricing. "Need help choosing the right plan for your business?" 2. HireVue, AI Interviewer. https://www.hirevue.com/platform/ai-interviewer. "including probing follow-up questions, in a natural two-way conversation." "Custom 1-5 rubric." "AI evaluates each response against your role-specific criteria, highlights whether it meets them, and explains why." "Recruiters view scores in the ATS." 3. HireVue, Enterprise. https://www.hirevue.com/solutions/enterprise. "authorization by FedRAMP PMO." 4. HireVue, AI Interviewer launch announcement, 18 Jun 2026. [Official announcement](https://www.hirevue.com/press-release/hirevue-launches-ai-interviewer). "conducts dynamic, two-way voice interviews with every applicant 24/7." 5. Vendr, HireVue. https://www.vendr.com/marketplace/hirevue. Median buyer pays "$20,900" per year. "The auto-renewal terms for HireVue include a requirement for a 60-day notice to opt out of automatic renewal." 6. Vendr, HireVue buyer guide. https://www.vendr.com/buyer-guides/hirevue. "The maximum price for HireVue software is around $145,000." 7. Willo, HireVue Review (2026). https://www.willo.video/blog/hirevue-review. "HireVue can be difficult to adapt when workflows change" (G2 reviewer, as quoted). "HireVue is not available in Zapier's app library." 8. Ribbon. https://www.ribbon.ai. "Ribbon screens, interviews and ranks candidates for warehouse, retail, hospitality and healthcare roles." "Works on any phone, in 10 languages, with follow-ups that actually make sense." "Syncs with 60+ ATS out of the box." "7-day free trial." 9. Ribbon, [Ribbon vs BrightHire](https://www.ribbon.ai/blog/ribbon-vs-brighthire), 23 Aug 2026: advertises annual-billing packages starting at $499/month for 100 interviews. Its dedicated pricing page returned 404 in a browser on 2 Nov 2026; confirm current terms with Ribbon. 10. HeyMilo. https://www.heymilo.ai. "The AI Recruiter for teams swamped by applications." "structured AI voice, video, SMS, and forms." "Operates 24/7 in 20+ languages." "1.4M+ candidates screened." "We provide a trust score that indicates the likelihood of cheating." "connects to major ATS platforms including Greenhouse, Bullhorn, Avionté, Ashby, Workday, SmartRecruiters, and JobDiva." 11. Alex. https://www.alex.com. "Copyright © 2026 Apriora Inc. dba Alex." "30" "Languages supported." "33+ officially supported ATS integrations." "Let Alex call thousands of candidates instantly" "Engage candidates with intelligent text conversations" "Conduct seamless interviews with your global candidates" "Alex scores every application for fraud risk across email, phone, LinkedIn, location and more." 12. Paradox. https://www.paradox.ai. "Mobile-first automation for frontline workers and managers." "Let candidates choose the best time slot that works for everyone." "Respond instantly in 100+ languages." "Candidate communication directly through SMS on Workday." 13. Sapia.ai, Pricing. https://sapia.ai/pricing. "Sapia.ai's pricing structure works on a per-hire basis." "You can interview and assess as many people as you want, without limits." "Our technology is designed to best serve companies that hire more than 500 people per year." 14. Spark Hire, Pricing. https://www.sparkhire.com/pricing/. Video interviews monthly "Starting at" "$299" "/ month"; annual "Starting at" "$249" "/ month" "Billed annually". "Unlimited live and one-way video interviews." "Access 40+ ATS integrations." 15. Spark Hire. https://www.sparkhire.com/. "Spark Hire helps you surface the right candidates early with AI scoring." "More than 10,000 organizations have used Spark Hire." 16. Willo, Pricing. https://www.willo.video/pricing. "£49 / live role." "Pay only for the live roles you use each month." "Includes 150 assessed candidates per live role." Lite: "AI summaries, benchmarks & insights". "From £2,999 / year." "50+ native integrations." Enterprise Willo Intelligence: "Follow-up questions". 17. Tracy AI Interviewer. https://tracyinterview.ai. "Tracy AI Interviewer runs natural two-way video interviews, follows up on vague answers, digs into the interesting ones." "Choose video or audio for the interview. Each person opens a link in their browser and joins with no sign-up." "People join in their browser whenever it suits them." "Tracy AI Interviewer can ask role-specific questions, introduce a coding task, and probe the trade-offs behind an answer." "Ask a backend engineer to design a rate limiter, then follow up on concurrency, failure modes, and testing." "Each interview becomes a readable summary with the participant's own words attached." "Direct quotes for context." 18. Tracy AI Interviewer, Pricing. https://tracyinterview.ai/pricing. "$1 per credit, in USD. Credits never expire." "You pay for results, not seats." "A result costs 1 credit for every 20 minutes of the session." "A session with no answers is never charged." 19. Qode, Tracy. https://qode.world/tracy. "4.7/5 average candidate rating across interviews." 20. Qode, Pricing. https://qode.world/pricing. "Interview integrity detection: Tab changes, Abnormal eye tracking, Live face detection, Average delay time before answer, External screens, Other unusual patterns." "Video & chapter breakdown", "Transcript", "Answer summary & score", "Explain answer score by AI Agent". Enterprise: "Custom based on volume". 21. Qode, 28 Oct 2026. Enterprise pricing is custom, based on volume. 22. Qode, 29 Oct 2026. Integrations: ChatGPT, Claude, Gemini, Mistral, Grok, Slack, WhatsApp, Zapier, Workable, HubSpot. SOC 2 Type 2, report on request at tracy@qode.world. 23. HireVue, Video Interviewing. https://www.hirevue.com/platform/video-interviewing. Retrieved 29 Oct 2026. "One-way asynchronous interviewing giving candidates 24/7 flexibility." "Meet in Hirevue, Zoom, Teams, or Webex." # HireVue Pricing: Reported Costs vs Tracy (Nov 2026) > HireVue pricing isn't public. Buyers report a $20,900 median a year. See what Tracy AI Interviewer costs per interview instead. - URL: https://tracyinterview.ai/compare/hirevue-pricing - Published: 2026-11-02 - Updated: 2026-11-13 - Author: Tracy team > **In short** > - **HireVue does not publish a standard price; Tracy AI Interviewer charges per interview result.** Tracy costs $1 per credit, one credit covers up to 20 minutes, unused credits do not expire and a session with no answers is not charged. > - At list price, 2,500 interviews of 20 minutes or less cost $2,500, versus the reported $20,900 HireVue median cited below. High-volume Tracy pricing is custom. **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=hirevue-pricing&utm_content=demo)** or **[Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=hirevue-pricing&utm_content=try-free)** ## Contents - [What does Tracy cost next to HireVue pricing?](#what-does-tracy-cost-next-to-hirevue-pricing) - [What does a Tracy interview give you for the money?](#what-does-a-tracy-interview-give-you-for-the-money) - [What is the real HireVue cost per year?](#what-is-the-real-hirevue-cost-per-year) - [Why is HireVue pricing not public?](#why-is-hirevue-pricing-not-public) - [What does HireVue's AI Interviewer add for the price?](#what-does-hirevues-ai-interviewer-add-for-the-price) - [How many Tracy interviews does a HireVue budget buy?](#how-many-tracy-interviews-does-a-hirevue-budget-buy) - [Who should pick HireVue instead?](#who-should-pick-hirevue-instead) - [How do you get a HireVue quote you can compare?](#how-do-you-get-a-hirevue-quote-you-can-compare) - [Questions](#questions) - [Sources](#sources) ## What does Tracy cost next to HireVue pricing? HireVue pricing is sold by quote, on an annual contract. Tracy AI Interviewer publishes its price: $1 per credit; one credit covers up to 20 minutes. You pay for results, not seats. A session with no answers is never charged, so applicants who don't show up cost nothing. Credits never expire, so a quiet quarter doesn't waste budget. Here is what that means at the volumes a skilled-hiring team runs, with each interview at 20 minutes or less. | Interviews a year | Tracy at list price | | --- | --- | | 1,000 | $1,000 | | 2,500 | $2,500 | | 5,000 | $5,000 | At 5,000 interviews a year, list price is still well below HireVue's reported median. Enterprise pricing is custom, based on your volume, so a team at that scale should ask for a quote before comparing. ![Bar chart of annual cost in US dollars. Tracy at list price: 1,000 interviews a year $1,000, 2,500 $2,500, 5,000 $5,000. HireVue buyer-reported: marketplace low $9,794, median $20,900, high $41,056; buyer guide average $49,855 and maximum about $145,000.](/blog/images/hirevue-pricing-cost-chart.svg) *At list price, 2,500 Tracy interviews a year cost less than HireVue's median contract.* **Thousands of interviews, zero new hires.** People join in their browser with no sign-up, whenever it suits them, so you can interview every applicant in depth during a hiring spike without adding a recruiter. For skilled and technical roles, where AI-written applications look alike, that lets you stop picking the shortlist from CVs alone. Candidates rate Tracy 4.7 out of 5 on average, so the extra interview doesn't turn good applicants away. Our guide on how to [screen every candidate in depth](/blog/how-to-screen-candidates?utm_source=blog&utm_medium=seo&utm_campaign=hirevue-pricing&utm_content=internal-link) shows how teams set it up. ## What does a Tracy interview give you for the money? You're paying to learn how the candidate did the work. Tracy runs natural two-way video interviews and follows up on vague answers, so a rehearsed first answer is followed by a question about what the person did. Each interview becomes a summary with "the participant's own words attached" and direct quotes for context, and the AI explains each answer score, so a reviewer can check the score against what was said. ![A Tracy result report for a sales support role, held in Japanese. The same sections as any report: a summary, the recording with a timestamped transcript, and each question with what the candidate said and quotes as evidence.](/blog/images/lib-report-overview-sales-ja.png) *The same report for an interview held in Japanese: the recording, the transcript and every answer with the candidate's own words (illustration with sample data).* Example: part of a Tracy report for a data engineer. It's an illustration with test data, not a real candidate. | Field | What the candidate said | | --- | --- | | Pipeline design | "We moved the nightly batch to streaming because finance needed numbers by 9am." | | Follow-up: why not a faster batch? | "The source system locked tables during the run, so a faster job still waited on the lock." | | Follow-up: what broke? | "Late events. We added a two-hour window and a daily reconciliation job." | The first answer could come from a good CV. The two follow-ups show whether the person did the work. For engineers, Tracy can "introduce a coding task, and probe the trade-offs behind an answer", so your senior engineers spend their hour only on people who can reason about code; [AI technical interviews](/blog/ai-technical-interviews?utm_source=blog&utm_medium=seo&utm_campaign=hirevue-pricing&utm_content=internal-link) walks through one. Each interview reports integrity signals (tab switching, leaving fullscreen, an external screen, face and gaze, response timing and repeated requests for a question), so a reviewer can check that the person who passed did it alone. ## What is the real HireVue cost per year? HireVue's pricing page says the platform is built "from mid-market to enterprise" and offers a demo and an ROI calculator, with no figures. The numbers below come from buyers who shared what they paid. | What it measures | Figure | Source | | --- | --- | --- | | Marketplace median | $20,900 a year | Vendr marketplace | | Marketplace low and high | $9,794 to $41,056 a year | Vendr marketplace | | Buyer guide average | $49,855 (5 deals, 3 buyers) | Vendr buyer guide | | Buyer guide maximum | about $145,000 | Vendr buyer guide | | Essentials package starting point | about $35,000 a year | Willo's HireVue review (Willo is a HireVue rival) | | Renewal | auto-renews, 60 days' notice to opt out | Vendr buyer note | The samples are small. The buyer guide rests on 5 deals, and the marketplace page shows no sample size. Sources also disagree: one cost site lists a median of $124,927 "across 5 Vendr purchases", six times Vendr's own marketplace median. Read $20,900 as a rough midpoint. The $145,000 comes from the buyer guide and sits well above the marketplace high of $41,056. ## Why is HireVue pricing not public? HireVue sells a suite, and each contract is built from modules. Its site lists Match and Apply, AI Interviewer, Assessment Builder, Assessments, Video Interviewing, and Cheating + Fraud Mitigation. Vendr lists three packages (HireVue: Essential, HireVue: Enterprise and HireVue: Premium) with no price for any of them. When the modules, roles and support level all move the total, a quote is the normal way to sell. Willo's review notes that "Teams report widely varying final costs for similar configurations." Ask about three costs before you sign: 1. Implementation. One cost site says "HireVue implementation is not included in the license cost." Get the setup fee in writing. 2. Volume limits. Ask how many interviews the contract covers and what one more costs. 3. Renewal. Put the 60-day notice date in your calendar the day you sign. ## What does HireVue's AI Interviewer add for the price? Part of what a HireVue contract pays for is the AI Interviewer and the research behind it. HireVue's launch announcement in June 2026 described "dynamic, two-way voice interviews with every applicant 24/7". Its product page says the agents ask "probing follow-up questions" and score answers on one of three formats: "Meets or Does Not Meet", a "Custom 1-5 rubric", or "Hirevue validated competencies". HireVue reports "1,300+ Validation studies for construct validity", and recruiters "view scores in the ATS". That suits roles with clear right answers and teams that want a score in their ATS. Tracy suits roles where you need to hear how the person reasons. The contract also carries credentials not every buyer needs, such as FedRAMP authorization. ## How many Tracy interviews does a HireVue budget buy? At the $1 list price, with each interview at 20 minutes or less: | HireVue figure (reported) | Annual cost | Tracy interviews it would pay for | | --- | --- | --- | | Marketplace low | $9,794 | 9,794 | | Marketplace median | $20,900 | 20,900 | | Buyer guide average | $49,855 | 49,855 | | Buyer guide maximum | about $145,000 | 145,000 | Example: a team hires 40 engineers and specialists a year and interviews 50 applicants per role. That's 2,000 interviews, or $2,000 a year at list price, which is $50 per hire. The same team on HireVue's median contract would pay about $523 per hire. A 30-minute interview uses two credits, so set the length to fit the role. A HireVue contract can include assessments and live interview tools, which Tracy doesn't sell. If the AI interview is the part you need, the per-interview comparison is the one that counts. For the full feature view, see [Tracy vs HireVue](/compare/tracy-vs-hirevue?utm_source=blog&utm_medium=seo&utm_campaign=hirevue-pricing&utm_content=internal-link). ## Who should pick HireVue instead? - You hire for US federal agencies and need FedRAMP authorization. - You want assessments, one-way video, live interviews in Zoom, Teams or Webex, and AI interviews under one contract. - You want HireVue's validated competency scores to land in your ATS. For everyone else, Tracy fits next to the tools you have. It works inside ChatGPT, Slack, Workable and other tools your team already uses, and it is SOC 2 Type 2 compliant, so your security review has a report to check (on request at tracy@qode.world). Comparing several tools? See our [HireVue alternatives](/compare/hirevue-alternatives?utm_source=blog&utm_medium=seo&utm_campaign=hirevue-pricing&utm_content=internal-link) and the [best AI interview software for hiring teams](/blog/best-ai-interview-software?utm_source=blog&utm_medium=seo&utm_campaign=hirevue-pricing&utm_content=internal-link). ## How do you get a HireVue quote you can compare? Get these answers in writing: 1. The annual price, split by module. 2. The interviews, roles or users it covers, and the cost of going over. 3. The setup fee. 4. The renewal date and notice period. 5. The price of the AI Interviewer on its own. Divide the last figure by the interviews you expect to run in a year, and you have a cost per interview to set against Tracy's $1 a credit. Vendr's buyer guide lists the usual levers: a multi-year term, bundling, your growth plans and signing near the supplier's quarter end. Agencies running interviews for several clients should ask whether pricing is per client; our guide to [AI interview tools for staffing agencies](/blog/ai-interview-tools-for-staffing-agencies?utm_source=blog&utm_medium=seo&utm_campaign=hirevue-pricing&utm_content=internal-link) covers that case. ## Common questions ### Is HireVue free? No. HireVue's pricing page offers a demo and an ROI calculator, with no free plan or trial listed. Tracy publishes its price and charges per result. ### What is HireVue pricing per month? HireVue doesn't publish a monthly price. Spread over 12 months, the $20,900 median contract reported on Vendr works out to about $1,742 a month. ### Can I cancel a HireVue contract at any time? A buyer note on Vendr says HireVue contracts renew automatically unless you opt out with 60 days' notice. Check the terms in your own contract. ### How much would 1,000 interviews cost with Tracy? $1,000 at list price, if each runs 20 minutes or less. Larger teams can ask for enterprise pricing, which is custom, based on volume. ### What counts as a charged interview on Tracy? A session that produces a result. A session with no answers is never charged, and one credit covers up to 20 minutes. --- **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=hirevue-pricing&utm_content=demo)** or **[Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=hirevue-pricing&utm_content=try-free)** ## Sources All retrieved 29 Oct 2026. 1. HireVue, Pricing. https://www.hirevue.com/pricing. "From mid-market to enterprise our platform is built for your specific hiring needs." 2. HireVue, Enterprise. https://www.hirevue.com/solutions/enterprise. Modules listed: "Match and Apply", "AI Interviewer", "Assessment Builder", "Assessments", "Video Interviewing", "Cheating + Fraud Mitigation". Security: "authorization by FedRAMP PMO." 3. HireVue, AI Interviewer. https://www.hirevue.com/platform/ai-interviewer. "including probing follow-up questions, in a natural two-way conversation." Scoring formats: "Meets or Does Not Meet", "Custom 1-5 rubric", "Hirevue validated competencies". "1,300+ Validation studies for construct validity." "Recruiters view scores in the ATS and quickly decide the next steps." 4. HireVue, Video Interviewing. https://www.hirevue.com/platform/video-interviewing. "Meet in Hirevue, Zoom, Teams, or Webex." 5. HireVue, AI Interviewer launch announcement, 18 Jun 2026. [Official announcement](https://www.hirevue.com/press-release/hirevue-launches-ai-interviewer). "conducts dynamic, two-way voice interviews with every applicant 24/7." 6. Vendr, HireVue marketplace page. https://www.vendr.com/marketplace/hirevue. Median buyer pays "$20,900" per year; low "$9,794"; high "$41,056". Products: "HireVue: Enterprise", "HireVue: Essential", "HireVue: Premium". "The auto-renewal terms for HireVue include a requirement for a 60-day notice to opt out of automatic renewal." 7. Vendr, HireVue buyer guide. https://www.vendr.com/buyer-guides/hirevue. "The maximum price for HireVue software is around $145,000." Avg Contract Value "$49,855"; Deals handled 5; Unique Purchasers 3. Levers: "Multi-Year Commitments", "Economies of Scale", "Expected Growth", "End-of-Quarter Signatures". 8. Truffle, Screening software pricing. https://www.hiretruffle.com/compare/screening-software-pricing. "No published pricing. Vendr's marketplace puts the median buyer at $20,900 a year (low $9,794, high $41,056)." 9. Willo, HireVue Review (2026). https://www.willo.video/blog/hirevue-review. "pricing starts around $35,000 per year for the Essentials package, with Enterprise plans requiring custom quotes." "Teams report widely varying final costs for similar configurations." 10. CostBench, HireVue hidden costs. https://costbench.com/software/ai-recruiting/hirevue/hidden-costs. "HireVue implementation is not included in the license cost." "The median HireVue contract is $124,927/yr across 5 Vendr purchases." 11. Tracy AI Interviewer, Pricing. https://tracyinterview.ai/pricing. "$1 per credit, in USD. Credits never expire." "You pay for results, not seats." "A result costs 1 credit for every 20 minutes of the session, started minutes included." "A session with no answers is never charged." 12. Tracy AI Interviewer. https://tracyinterview.ai. "Tracy AI Interviewer runs natural two-way video interviews, follows up on vague answers, digs into the interesting ones." "Tracy AI Interviewer can ask role-specific questions, introduce a coding task, and probe the trade-offs behind an answer." "Each interview becomes a readable summary with the participant's own words attached." "Direct quotes for context." "Each person opens a link in their browser and joins with no sign-up." "People join in their browser whenever it suits them." 13. Qode, Tracy. https://qode.world/tracy. "4.7/5 average candidate rating across interviews." 14. Tracy AI Interviewer, product page: Integrity. https://tracyinterview.ai/product#integrity. "Tracy captures signals during the interview: the face and where the person is looking, response timing, tab switching, leaving fullscreen, an external screen and repeated requests for a question." 15. Qode, 28 Oct 2026. Enterprise pricing is custom, based on volume. 16. Qode, 29 Oct 2026. Integrations (ChatGPT, Claude, Gemini, Mistral, Grok, Slack, WhatsApp, Zapier, Workable, HubSpot) and SOC 2 Type 2, report on request at tracy@qode.world. # Tracy vs HireVue: AI Interviews Compared (Nov 2026) > Tracy vs HireVue for skilled hiring: compare adaptive follow-ups, evidence behind scores, pricing and when a full hiring suite makes sense. - URL: https://tracyinterview.ai/compare/tracy-vs-hirevue - Published: 2026-11-02 - Updated: 2026-11-13 - Author: Tracy team ![A candidate in a Tracy interview in the browser. Tracy asks a follow-up about the answer he just gave, shown as a live caption.](/blog/images/tracy-interview-follow-up.png) > **In short** > - **Choose Tracy AI Interviewer for skilled and technical hiring when the first interview must show how someone thinks.** It interviews every applicant, follows up on vague answers and ties each score to the candidate's own words. > - Choose HireVue when you want one enterprise suite from application to offer or need FedRAMP. Choose Tracy for deeper, verifiable first-round screening with published pay-per-result pricing. HireVue is a full hiring suite. Its AI Interviewer holds a two-way voice conversation, scores answers against a rubric and sends scores to your ATS. Tracy interviews on video, follows up inside each person's field and backs its summary with their own words, so reviewers can check the evidence before deciding. Pick HireVue if you want one vendor from application to offer, automatic scores in your ATS, or FedRAMP. Pick Tracy if you want follow-up questions that go deep in each person's field, cheating signals in every interview, a setup built around the process you already run, and no upfront cost. ## Who each is for HireVue is built for large employers that want one platform for the whole hiring process, with validated assessments and a long list of security certifications, including FedRAMP for US government work. Tracy is built for teams that screen skilled people at volume, where the second and third questions decide: engineers, domain experts, specialists, and [founders applying to a program](/blog/ai-interviews-for-applications?utm_source=blog&utm_medium=seo&utm_campaign=tracy-vs-hirevue&utm_content=internal-link). ## Side by side | | Tracy at Qode | HireVue | |---|---|---| | What it is | An AI interviewer | A hiring suite: AI Interviewer, video interviewing, assessments, automation and scheduling, Match and Apply | | The AI interview | Two-way video. Follow-up questions inside the person's own field, built on each answer, including live coding for engineers | A voice interview in "a natural two-way conversation", with "probing follow-up questions" | | What the reviewer gets | Each answer in the person's own words, linked to its moment in the recording, organized by your fields. Scoring is optional | Scores against your criteria: meets or does not meet; a 1 to 5 rubric based on which expected points the answer mentions; validated competency scales. Scores go to your ATS with an explanation of why | | Who decides | Your team. Scoring is optional and follows your setup | "Recruiters always make the final call" | | Other interview formats | None | One-way on-demand video; live interviews in HireVue, Zoom, Teams or Webex | | Use cases | Hiring, technical interviews with live coding, expert vetting, founder applications | Hiring | | Security and compliance | SOC 2 Type 2 attested | ISO/IEC 27001:2022, ISO/IEC 27701:2019, an annual SOC 2 Type 2 audit, FedRAMP authorization | | Price | No upfront cost and no seats; you pay per result. $1 per credit; one credit covers up to 20 minutes. Enterprise pricing set to your volume | Not published; HireVue asks you to request a demo | ![HireVue's AI Interviewer scoring options: Meets or Does Not Meet, a custom 1 to 5 rubric, and HireVue-validated competency scales](/blog/images/hirevue-ai-interviewer-scoring.png) *HireVue's AI Interviewer page. Its example skill rubric scores an answer from 1 to 5 by how many expected points it mentions. Screenshot of hirevue.com/platform/ai-interviewer, taken 26 Oct 2026.* ## Where Tracy is the better pick **Your roles need judgment a checklist can't score.** HireVue's own product page shows its example skill rubric scoring an answer from 1 to 5 by how many expected points it mentions. A candidate who has rehearsed, or has an AI feeding them the right words, can hit every point. Tracy asks about what the person just said, such as why they chose that approach or what they would change. Ask a backend engineer to design a rate limiter, and Tracy follows up on concurrency, failure modes and testing, or hands them a coding task and probes the trade-offs. AI technical interviews shows a full example. **Cheating signals come built in.** HireVue's AI Interviewer, the part that asks follow-up questions, is voice-based. Its video interviews are one-way recordings or live calls with your own interviewers. HireVue does list a cheating and fraud mitigation offering, and says suspicious AI Interviewer cases can be flagged for human review. Tracy asks its follow-ups on video, and every interview comes with cheating signals for your reviewers to check. Gartner predicts 1 in 4 candidate profiles worldwide could be fake by 2028, and in Greenhouse's 2025 AI in Hiring Report, 39% of US hiring managers said they now run more in-person interviews to confirm candidates are real. How to detect AI cheating in interviews covers the warning signs. **Every point comes with the moment it was said.** A 3 out of 5 doesn't show what the person said. In a Tracy summary, each point is the person's own words, linked to that moment in the recording, so a reviewer can confirm a strong answer, or catch a thin one, in one click. AI interview summaries shows the layout. **It fits the process you already run.** Bring your own questions or start from a proven interview type, choose the fields your reviewers get back, and send a personal invite or a public link from wherever you already contact people. They join in the browser with no sign-up. Teams get a custom interview set up in days. HireVue is built to run the whole process, from application to offer, on its own platform. **There's no upfront cost.** You pay only when an interview produces a result. There are no seats to buy, a session with no answers is never charged, and enterprise teams get pricing set to their volume. HireVue doesn't publish prices; its pricing page asks you to request a demo. The HireVue pricing guide collects what is public. ## What a reviewer sees in Tracy ![How each tool hands over an interview. HireVue: the candidate answers by voice, the AI scores against a rubric, the score goes to the ATS, and a recruiter decides from scores. Tracy: the person answers on video, gets a summary by field, each point is quoted and linked to the recording, and a reviewer decides from their words.](/blog/images/hirevue-vs-tracy-reviewer-flow.svg) Here is part of a Tracy summary for a customer support lead role. It is an illustration with test data, not a real candidate. | Field | What the candidate said | |---|---| | Handling escalations | "Last month a customer threatened to cancel over a billing error. I refunded first and fixed the invoice after, because the refund was the part they cared about." | | Team managed | "Six agents across two shifts." | | Start date | "I can start on 1 December." | Click any quote and the recording jumps to that moment. Scoring is optional, and every point stays tied to the answer behind it. ![A Tracy result report for a customer support lead. Four fields, such as handling escalations and start date, are marked Stated, each with the candidate's own words as evidence. The recording and timestamped transcript sit above them.](/blog/images/tracy-interview-result-support-lead.png) *What the reviewer sees afterwards: every answer quoted, next to the recording and a transcript with the time of each line. Your team decides.* ## Pricing side by side HireVue doesn't publish its prices. Its pricing page says the platform is built "from mid-market to enterprise" and offers a demo and an ROI calculator. Tracy publishes its list price at tracyinterview.ai/pricing: $1 per credit; one credit covers up to 20 minutes. You pay per result, never per seat, and credits never expire. Enterprise teams get pricing set to their volume. ## Trying Tracy next to HireVue You don't have to replace HireVue to try Tracy. Pick one role or one program, run Tracy next to your current first round on the same applicants for two weeks, and compare what your reviewers would have decided. Tell applicants that an AI runs the interview; are AI interviews legal? covers what to disclose. ## Common questions ### Does HireVue have an AI interviewer? Yes. HireVue's AI Interviewer holds a two-way voice conversation with follow-up questions and scores each response against role-specific criteria. ### Does Tracy score candidates? No. Tracy doesn't score, rank or reject anyone. It organizes what each person said by your fields, backs every point with their own words, and someone on your team makes the decision. ### How much does HireVue cost? HireVue doesn't publish prices. Its pricing page asks you to request a demo. ### Is HireVue certified for government work? HireVue says it holds FedRAMP authorization, after an audit by a third-party assessment organization. ### Can I use Tracy for expert vetting or founder interviews? Yes. Tracy has templates for hiring, technical, expert and application interviews, and you can write your own questions. --- **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=tracy-vs-hirevue&utm_content=demo)** or **[Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=tracy-vs-hirevue&utm_content=try-free)** ## Sources All retrieved 26 Oct 2026. - HireVue, [Pricing](https://www.hirevue.com/pricing) - HireVue, [Enterprise](https://www.hirevue.com/solutions/enterprise) (security and compliance, product list) - HireVue, [AI Interviewer](https://hirevue.com/platform/ai-interviewer) - HireVue, [Video interviewing](https://www.hirevue.com/platform/video-interviewing) - HireVue, [AI Interviewer launch announcement](https://www.hirevue.com/press-release/hirevue-launches-ai-interviewer), June 2026 - Gartner, [Survey Shows Just 26% of Job Applicants Trust AI Will Fairly Evaluate Them](https://www.gartner.com/en/newsroom/press-releases/2025-07-31-gartner-survey-shows-just-26-percent-of-job-applicants-trust-ai-will-fairly-evaluate-them), 31 Jul 2025 - Greenhouse, [2025 AI in Hiring Report release](https://www.greenhouse.com/newsroom/an-ai-trust-crisis-70-of-hiring-managers-trust-ai-to-make-faster-and-better-hiring-decisions-only-8-of-job-seekers-call-it-fair), 19 Nov 2025 - Tracy at Qode, [tracyinterview.ai](https://tracyinterview.ai) and [pricing](https://tracyinterview.ai/pricing) # Admissions Interview Questions for Programme Teams > Admissions interview questions for programme teams: 15 questions with follow-ups, strong and weak answers, and how to interview every applicant. - URL: https://tracyinterview.ai/blog/admissions-interview-questions - Published: 2026-11-13 - Author: Tracy team > **In short** **Tracy AI Interviewer** gives every applicant a first interview, starting on click at any hour, and follows up on what each person says, so no application is judged on a form alone. > > Use the 15 admissions interview questions below, written for programme teams, each with a follow-up and what to listen for. > > Every result keeps a recording, a timestamped transcript and a written assessment per question; your team decides. > > Next step: create one interview for your next intake. An accelerator, fellowship or scholarship opens applications and 800 arrive for 30 places. The form says what people chose to write. It does not say who can explain a decision under a follow-up. Admissions interview questions have to find that, and to ask each applicant questions that test the same criteria. These 15 admissions interview questions are written for programme and admissions teams selecting people for accelerators, fellowships and scholarships. Applicants and parents also search for this; the questions below show what an interviewer is looking for, so they work as preparation too. Each has a follow-up and a note on what a strong answer contains. Tracy AI Interviewer gives every applicant a first interview. She asks each person, follows up on what they said and waits for the answer. See the [applications use case](/use-cases/applications) and [accelerator interviews](/blog/accelerator-interview) for the wider process. ## Fifteen admissions interview questions, grouped by what they test Pick one or two from each group. Ask the follow-up only after the first answer, write down your criteria before the first applicant joins, and ask every applicant the same core questions, then follow up on what each one says. ### Group 1: what they have done | Ask | Follow up with | Listen for | | --- | --- | --- | | 1. Tell me about something you built or organised that you are proud of. | What did you change after the first month, and why? | A specific project, their own part, a change they made on evidence. | | 2. Describe a time something you tried did not work. | What did you do the next day? | Names the failure without blaming. A concrete next step. | | 3. What is the hardest thing you have taught yourself? | How did you know you had learned it? | A method, a test of their own, not just time spent. | ### Group 2: how they think | Ask | Follow up with | Listen for | | --- | --- | --- | | 4. Pick a problem you care about. Why does it exist? | What would change your mind about that? | A cause beyond the obvious. Open to evidence that would change the view. | | 5. What is a belief you held a year ago and have since dropped? | What changed it? | A real reversal and the reason for it. | | 6. You have two weeks and no budget. What do you do first? | What do you leave out? | Picks one thing, gives a reason, accepts the cost of leaving things out. | ### Group 3: motivation and fit | Ask | Follow up with | Listen for | | --- | --- | --- | | 7. Why this programme, and why now? | What would you do if you were not accepted? | Specific to the programme. A plan that does not depend on it. | | 8. What do you want to be different in a year? | What are you doing about it this month? | A goal with an action beside it. | | 9. Which of your answers on the form would you change now? | Why? | Reflects honestly. Corrects or sharpens, does not defend. | ### Group 4: working with others | Ask | Follow up with | Listen for | | --- | --- | --- | | 10. Tell me about a team you were part of that struggled. | What was your part in it? | Owns a share. Describes the fix, not only the problem. | | 11. Describe a time you disagreed with someone and turned out to be wrong. | How did you find out? | Says it plainly. What they do now. | | 12. Who has helped you most, and what did they do? | What did you do for them? | Specific help. Gives back. | ### Group 5: judgement and evidence | Ask | Follow up with | Listen for | | --- | --- | --- | | 13. Tell me about a decision you made with too little information. | What would you want to know next time? | Names the gap and how they acted anyway. | | 14. How would you tell whether your idea is worth another year? | What result would make you stop? | A stop rule stated before the work. | | 15. What question do you wish we had asked? | Why that one? | Shows what they think matters. Useful for the next round. | ## A strong answer shows a choice, and a weak one shows an outcome The difference shows up in the second answer. These pairs are written by us to show how it sounds; they are not quotes from real applicants. | | Weak answer | Strong answer | | --- | --- | --- | | Question 1, the project | "It went really well and the group grew." | "After a month I saw quiet students never spoke, so I moved to small groups. Attendance held and more of them answered." | | Question 2, what did not work | "Nothing has really gone wrong for me." | "I sent a survey nobody answered. The next day I asked five people in person, and that told me what to change." | | Question 9, the changed answer | "I would not change anything." | "I wrote that I wanted to build an app. Having thought about it, the real goal is the problem, and an app may not be the answer." | Listen for choices. An applicant who explains why they changed something has shown how they think; one who lists results has shown what happened. ## One follow-up chain shows where the real answer starts This is an illustration with sample data, not a real interview. The applicant is fictional. It shows the shape of what Tracy runs: an opening question from your list, the answer, and a follow-up built on it. The first chain starts from question 1; two shorter ones follow. > **Tracy:** Tell me about something you built or organised that you are proud of. > > **Applicant (fictional):** I ran a weekly tutoring group for eleven students. > > **Tracy:** What did you change after the first month, and why? > > **Applicant:** I moved it to small groups, because quiet students never spoke. > > **Tracy:** How did you know the change worked? > > **Applicant:** More of them answered questions, and two started to help the others. The first answer could be on any form. By the third, you can hear whether the applicant ran it. Question 4, why a problem exists: > **Tracy:** Pick a problem you care about. Why does it exist? > > **Applicant (fictional):** Students skip the library because the opening hours do not fit their classes. > > **Tracy:** What would change your mind about that? > > **Applicant:** If most of them told me hours were fine and the real issue was the noise. Question 10, a team that struggled: > **Tracy:** Tell me about a team you were part of that struggled. > > **Applicant (fictional):** Our robotics team missed two deadlines in a row. > > **Tracy:** What was your part in it? > > **Applicant:** I kept the schedule and did not flag the slip until it was too late. ![Transcript card: applicant ran a tutoring group, Tracy asks what changed after the first month and why.](/blog/images/admissions-interview-questions-1.png) *Tracy asks a follow-up built on the applicant's own answer. The programme team hears the thinking behind the form. Illustration with sample data* A follow-up on the applicant's own answer shows what they changed and why. ## Give every applicant a first interview You cannot hold 800 calls. Here is what Tracy does instead. - **Every applicant starts on click.** Each person opens a link in a browser, with no account and no scheduling, at any hour, with no one from your team present. They are told on the join screen that the conversation is recorded and transcribed. See the [participant experience](/participant-experience). - **Follow-ups on what they say.** Tracy reads the CV when the CV check is on, asks in the applicant's own field and follows up on the answer. - **A dropped applicant can rejoin.** Within 10 minutes by default, set per customer. A reminder goes out after a day. - **A result you can share.** A recording and a timestamped transcript sit behind every result, with a written assessment per question, so a borderline applicant can be discussed with the team. See [AI interview reports](/blog/ai-interview-reports). - **Your questions.** Define what you want to assess before launch, then edit the draft questions Tracy makes from your brief. On screening time, a Director at a top-tier early-stage VC describes it: "I was effectively spending around six hours a day on screening. With Tracy, I can now review and assess those candidates in approximately 1.5 hours." It is $1 per credit, one credit per started 20 minutes of an interview that produced a result. The first 3 results are free with no card, and there are no seats. See [pricing](/pricing). ![Recreated Tracy result report for a fictional applicant: recording, transcript with the 4:01 answer, and each question with what was said and the quote.](/blog/images/admissions-interview-questions-2.png) *A recording and a timestamped transcript sit behind every Tracy result. The team checks what the applicant said and can discuss a borderline case. Illustration with sample data* Your team can open the recording and the transcript behind each answer, and share the page for a borderline applicant. ## Keep three decisions with your team 1. **Your criteria.** Write down what a strong answer contains before the first applicant joins, and use the same criteria for everyone; opening questions can differ with the CV. 2. **The decision.** A person on your team reads the evidence and decides who goes to the next round. Tracy publishes no bias audit, so your selection criteria and review stay yours. 3. **The final round.** A conversation with the people who will work with the cohort stays live. For 20 applicants, you can interview all of them yourself. The case for Tracy starts when applications outnumber your hours; see [AI interview tools for accelerators](/compare/ai-interview-tools-for-accelerators). Pilot on one intake: 30 applicants, six opening questions, one reviewer. ## Common questions ### What are good admissions interview questions? Ask about what the applicant has done, how they think, why this programme, how they work with others and how they judge evidence. Keep the follow-up beside each question, because that is where the real answer starts. ### What admission interview questions suit students? Use the first three groups: something they built, something that failed, and why this programme. Prepare one real example for each and be ready to say what you changed. See the weak and strong answers table above for what an answer should sound like. ### What are common college interview questions? Why this place, what you have done, how you handle a setback and what you want from the next years. The same follow-ups apply: what did you change, and how did you know it worked? ### Can Tracy interview every applicant to a programme? Yes. Each applicant starts on click at any hour, with no scheduling and no one from your team present. Your team gets a recording, transcript and assessment per question and decides. ### Does Tracy decide who is admitted? No. A person on your team reads the evidence and decides. --- Create one interview for your next intake from these questions, then read the recording and report; the button opens the Connect page. Or send us one programme and we will build the interview plan, with a recording and report you can check. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=admissions-interview-questions&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=admissions-interview-questions&utm_content=demo) ## Sources - [Tracy AI Interviewer home page](https://tracyinterview.ai/), retrieved 9 Nov 2026. - [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer Connect page](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. # Customer Success Manager Interview Questions for Hiring Managers > Customer success manager interview questions for hiring managers: 15 questions with follow-ups, strong and weak answers, and how to run them at volume. - URL: https://tracyinterview.ai/blog/customer-success-manager-interview-questions - Published: 2026-11-13 - Author: Tracy team > **In short** **Tracy AI Interviewer** interviews every customer success applicant on a real account story and follows up on what they actually did, so you hear judgement, not rehearsed answers. > > Use the 15 customer success manager interview questions below, written for the interviewer, each with a follow-up and what to listen for. > > Every result keeps a recording, a timestamped transcript and a written assessment per question; you decide. > > Next step: create an interview from your job description. You post a customer success role and a hundred people apply, most with the same phrases: relationship building, retention, stakeholder management. The CV does not say who has saved an account and who sat near one. Customer success manager interview questions have to ask for the account story and then ask what the person did next. These 15 customer success manager interview questions are written for the interviewer: a head of customer success, a hiring manager or a recruiter screening for the role. Each has a follow-up and a note on what a strong answer contains. They suit a first conversation of 20 to 30 minutes; a role play with your own team belongs in a later round. Tracy AI Interviewer runs that first conversation at volume. She asks each candidate in their own field, follows up on what they said and waits for the end of the thought. See [volume hiring](/use-cases/volume-hiring) for the wider process and [how to screen candidates](/blog/how-to-screen-candidates) for the steps before it. ## Fifteen customer success manager interview questions, grouped by what they test Pick one or two from each group. Ask the follow-up only after the first answer. Choose your opening questions for the role and write down what a strong answer contains before the first candidate joins. ### Group 1: retention and risk *Table 1: Retention and risk, with follow-ups and what a strong answer contains.* | Ask | Follow up with | Listen for | | --- | --- | --- | | 1. Tell me about an account that was close to leaving. | What did you change in the account after that? | A named account, a cause found by asking, a change that followed, the outcome. | | 2. How do you know an account is at risk before the customer says so? | Which signal has been wrong? | Usage, support tickets, a changed sponsor. Admits a false alarm. | | 3. Describe a customer you lost. | What would you do differently now? | Owns part of it. A lesson that shows in how they work today. | ### Group 2: onboarding and adoption *Table 2: Onboarding and adoption, with follow-ups and what a strong answer contains.* | Ask | Follow up with | Listen for | | --- | --- | --- | | 4. How do you run the first 30 days with a new customer? | Where do new customers most often stall? | A plan with milestones. A real stall point and what they did about it. | | 5. A customer bought the product but nobody uses it. What do you do? | Who do you speak to first? | Finds the cause before offering training. Reaches users, not only the buyer. | | 6. How do you decide which customers get your time? | Tell me about one you deprioritised. | A rule they can state. Comfortable with a trade-off. | ### Group 3: difficult conversations *Table 3: Difficult conversations, with follow-ups and what a strong answer contains.* | Ask | Follow up with | Listen for | | --- | --- | --- | | 7. A customer asks for a feature you will not build. What do you say? | How did that customer react? | Honest no, the reason, an alternative. Keeps the relationship. | | 8. Tell me about a conversation with an angry customer. | What did you say first? | Listens before explaining. A concrete opening line. | | 9. How do you handle a request from sales that conflicts with what the customer needs? | What did you give up? | Names both sides and a decision. Does not blame. | ### Group 4: commercial sense *Table 4: Commercial sense, with follow-ups and what a strong answer contains.* | Ask | Follow up with | Listen for | | --- | --- | --- | | 10. How do you prepare for a renewal? | When do you start, and why then? | Starts early, ties value to what the customer said they wanted at signing. | | 11. Tell me about an expansion you helped create. | How did you know the customer was ready? | Need first, offer second. No pressure story. | | 12. How do you show a customer the value they have received? | What if the numbers are not good? | Uses the customer's own measures. Does not hide a weak quarter. | ### Group 5: working inside the company *Table 5: Working inside the company, with follow-ups and what a strong answer contains.* | Ask | Follow up with | Listen for | | --- | --- | --- | | 13. How do you pass a customer problem to product or engineering? | Tell me about one that got fixed. | A clear write-up, evidence, follow-through. | | 14. How do you keep notes so a colleague could take over tomorrow? | What do you always write down? | History, decisions, owners, next dates. | | 15. What does a good week look like for you in this role? | What do you drop when it goes badly? | Realistic, customer-led, knows what to cut. | ## A strong answer tells you what changed, and a weak one stays general What matters comes in the second answer. We wrote these pairs to show how it sounds; none is a quote from a real candidate. *Table 6: Weak and strong answers to the same questions, written as examples.* | | Weak answer | Strong answer | | --- | --- | --- | | Question 1, the account near leaving | "I built a strong relationship with them." | "Usage dropped, so I called their sponsor. The team had never been trained. I ran two sessions and they renewed." | | Question 7, the feature request | "I'd tell them we'd pass it to product." | "I'd say we are not building it this year, explain why, and show them the closest thing that works today." | | Question 10, the renewal | "I'd send a reminder 60 days out." | "I'd start at signing: what did they say they would measure? At renewal I show that number, good or bad." | Listen for what changed. A story that ends with a customer's feeling tells you less than one that ends with a decision. ## One follow-up chain shows where the real answer starts This is an illustration with sample data, not a real interview. The candidate is fictional. It shows the shape of what Tracy runs: an opening question from your list, the answer, and a follow-up built on it. It starts from question 1. > **Tracy:** Tell me about an account that was close to leaving. > > **Candidate (fictional):** A client stopped logging in. I called their sponsor and found out why. > > **Tracy:** What did you change in the account after that call? > > **Candidate:** Their team had never been trained. I ran two short sessions and set a monthly check on usage. > > **Tracy:** How did you know it had worked? > > **Candidate:** Logins doubled within six weeks, and they renewed for a second year. The first answer is the one every candidate prepares. By the third you know whether the account was theirs. ![Transcript card: client stopped logging in, candidate called the sponsor, Tracy asks what changed in the account.](/blog/images/customer-success-manager-interview-questions-1.png) *Tracy asks a follow-up built on the candidate's own account story. The hiring manager hears what the candidate actually did. Illustration with sample data* A follow-up on a real account story shows what the candidate changed, not only how they felt. ## Run the interview for every applicant without losing the follow-up A live call for every applicant does not survive a hundred applications. - **Set up from a job description.** Tracy drafts the questions from the JD, which you edit, or you paste your own, from quick to deep and behavioural to technical. See the [product page](/product). - **Follow-ups on the story.** If the candidate names an account, Tracy asks what they changed and how they knew it worked. - **Start on click.** Each candidate opens a link in a browser, with no account and no calendar, at any hour, and is told on the join screen that the conversation is recorded and transcribed. See the [participant experience](/participant-experience). - **A result you can check.** A recording and a timestamped transcript sit behind every result, with a written assessment per question. See [AI interview reports](/blog/ai-interview-reports). It is $1 per credit, one credit per started 20 minutes of an interview that produced a result. The first 3 results are free with no card, and there are no seats. See [pricing](/pricing). ![Tracy result report for a fictional customer success manager: recording with timestamped transcript, and three questions with what they said and quoted evidence.](/blog/images/customer-success-manager-interview-questions-2.png) *A recording and timestamped transcript sit behind every Tracy result. You check what a candidate said to each question before inviting them on. Illustration with sample data* The hiring manager can read what the candidate said to each question, with the recording and transcript beside it. ## Keep three decisions with people 1. **Who moves forward.** The hiring manager reads the report and decides. A recruiter decides: they can move a candidate forward or invite them to interview again, and the history stays on the record. 2. **The calibration.** Have one customer success lead read five transcripts blind. If they disagree with the written assessments, change the questions first. 3. **The live round.** A role play on a hard customer call, and a conversation about team fit, stay with your people. For two or three hires a year, a call with the hiring manager is simpler; see [AI interview vs human interview](/blog/ai-interview-vs-human-interview). The case for Tracy starts when applicants outnumber the hours you have. To try it, build an interview from one job description, send it to your next applicants and read the first reports before you decide how far to go. ## Common questions ### What are good customer success manager interview questions? Ask about accounts near leaving, onboarding and adoption, difficult conversations, renewals and working with product. Ask for a real account and a follow-up on what the person changed. ### What customer success manager interview questions suit someone senior? Use the commercial and company groups: renewals, expansion and how they pass a problem to product. Ask for the numbers they used and the decision that followed. ### How do client success manager interview questions differ? Client success manager is often used for the same role. Use the same groups and change the cases to your market, for example a services client rather than a software subscriber. ### Can Tracy interview customer success applicants? Yes. Tracy runs a structured interview from your job description, follows up on what each person says, and returns a recording, transcript and assessment per question. A person decides. ### Does Tracy decide who to hire? No. Your team reads the evidence and decides. --- Create an interview for this exact role from these questions, then read the recording and report; the button opens the Connect page. Or send us one role and we will build the interview plan, with a recording and report you can check. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=customer-success-manager-interview-questions&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=customer-success-manager-interview-questions&utm_content=demo) ## Sources - [Tracy AI Interviewer home page](https://tracyinterview.ai/), retrieved 9 Nov 2026. - [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer Connect page](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. # How to Run Customer Interviews > How to run customer interviews: pick one question, write an open guide, let customers join when it suits them, and read across the answers. - URL: https://tracyinterview.ai/blog/how-to-run-customer-interviews - Published: 2026-11-13 - Author: Tracy team > **In short** Tracy AI Interviewer runs customer interviews from one link: say what you want to learn, she drafts the guide, customers answer when it suits them. > > To run them yourself: one question, six to eight open questions about the last time it happened, a follow-up on every short answer, then read for reasons that repeat. > > Surveys show what people chose, not why, and booking twenty calls takes two weeks; findings link back to the words and the moment in the recording. > > The first 3 results are free: run one question before you commit. How to run customer interviews comes down to four moves: decide what you need to learn, write a guide of open questions, give customers a way to answer without a calendar, and read across the answers for reasons that repeat. The hard part is the third move. Booking twenty calls takes about two weeks, and notes from different interviewers cannot be compared. [Tracy AI Interviewer](/use-cases/customer-research) removes the booking: customers join in the browser when it suits them, and she asks the follow-ups. ## What should a customer interview find out? A customer interview should answer one question, stated before you write a single prompt. Four kinds cover most work: - **Discovery:** what is the customer trying to get done, and how do they do it today? - **Concept test:** does this idea fit a problem they have, and what would they do instead? - **Win-loss:** why did they choose us, or someone else? - **Onboarding feedback:** where did the first week go wrong? Pick one. A guide that tries to cover all four produces thin answers to each. Write the question as a sentence you could put at the top of the report, for example "Why do trial users stop before they invite a colleague?" ## How to run customer interviews, step by step 1. **Name the question and the decision it feeds.** If no decision changes with the answer, do not interview yet. 2. **Choose who to hear from.** Pick people who did the thing: churned accounts, lost deals, new sign-ups. Do not ask people what they might do. 3. **Write the guide.** Six to eight open questions are enough. Ask about the last time something happened, not about habits in general. Put the question you most need answered first. 4. **Let people answer when it suits them.** A booked call favours customers with free time, which is a bias of its own. A link they can open in a browser reaches people who would never find an hour. 5. **Follow up on every short answer.** "It got fiddly" is a pointer. Ask what they were trying to do, and where it went wrong. 6. **Read across the interviews.** Look for reasons that repeat, quote the words people used, and keep the moment each quote came from, so a colleague can check it. 7. **Close the loop.** Tell the team what changed because of what you heard. ## Which customer interview questions work? The same few questions work for customer interview questions and user interview questions alike. The follow-up carries the value: it turns an opinion into an example. A starter set for customer and user interviews. Reword each to your product. | Question | What it finds | Follow-up | | --- | --- | --- | | Tell me about the last time you [did the task]. | The real workflow, not the ideal one | What did you do first? What happened next? | | What were you using before? | The alternative you compete with | What made you switch, and what almost stopped you? | | What did you try that did not work? | Failed workarounds and their cost | What did that cost you in time or money? | | Who else was involved in the decision? | The buying group | What did each of them care about? | | What would make you stop using it? | The churn trigger | When did you last think about that? | ![Tracy asks why a customer stopped using a report builder, the customer says it got fiddly, and Tracy asks for the last report.](/blog/images/how-to-run-customer-interviews-1.png) *Tracy follows a vague answer with a question about the last time it happened, and each line is timestamped. Illustration with sample data* A vague answer becomes a specific example when the interviewer asks for the last time it happened. ## How do you read across the interviews? Read the first three interviews slowly, then write down the reasons you heard in the customers' own words. Group the reasons that repeat, and count how many customers gave each. Keep the quote and the moment it came from beside every group, so a product manager who doubts a finding can check the exact answer. Treat a reason heard once as a lead for the next round, not a conclusion. This is where a shared guide pays off: when everyone answered the same questions, you compare like with like instead of weighing different conversations. ![Tracy findings: one theme about a report builder, with two short customer quotes and their timestamps.](/blog/images/how-to-run-customer-interviews-3.png) *Findings with the words and the moment behind each. Illustration with sample data* A finding comes with the customers' own words and the moment each was said. ## What mistakes spoil customer interviews? Four habits ruin otherwise good interviews: - **Leading questions.** "Would a faster export help?" invites a yes. Ask what they do after the export. - **Pitching.** Once you describe your idea, people react to it politely. Describe the problem instead. - **Hypotheticals.** "Would you pay for this?" measures manners. Ask what they paid for the last similar thing. - **Different guides for different people.** If each interviewer asks in their own order, the answers cannot be compared. Note-taking is the last trap. An interviewer who writes while listening drops the follow-up. A recording with a transcript lets you listen first and read later. ## How does Tracy run customer interviews? Tracy runs the method above from one link. You describe what you want to learn and she drafts the interview guide, or you supply your own questions. You send one link, and customers join in their browser when it suits them, with no account needed. Every customer gets the same guide, so answers compare. She follows up on what each person says, and you read across the interviews: each finding links back to the words and the moment in the recording. The [customer research page](/use-cases/customer-research) shows the sample report, and [how AI interviews work](/blog/how-do-ai-interviews-work) covers what a customer sees. ![Three steps: say what you want to learn, send one link, read across interviews with each finding linked to the recording.](/blog/images/how-to-run-customer-interviews-2.png) *Customer interviews with Tracy in three steps: the guide is drafted from what you want to learn, customers join from one link, and each finding links back to the recording. Source: Tracy customer research use case.* Three steps replace the booking: say what you want to learn, send one link, read across the interviews. ## How to pilot it Start with one question and one list of customers: churned accounts, last quarter's lost deals or new sign-ups. [Create the interview](/connect), send the link, and read the first results before you widen. Stop when the answers repeat. A person still decides what to build or change, and what to tell the team; Tracy does the asking and keeps the record. [Pricing](/pricing) is per interview, so a small pilot costs little. If you run check-ins or exit interviews instead, see the [people teams page](/use-cases/people-teams). ## Common questions ### How do you run user interviews? Choose one question, recruit people who did the thing, ask open questions about the last time it happened, follow up on short answers, and read across the interviews for repeated reasons. ### What are good customer interview questions? Ask about the last time they did the task, what they used before, what they tried that did not work and who else was involved. Follow each with a request for an example. ### How many customers should you interview? Enough that the reasons start to repeat. Start with a small first batch from one list, read it, and add more only if new reasons keep appearing. ### Can an AI interviewer run customer interviews? Tracy runs discovery, concept test and win-loss interviews from one link, using a guide drafted from what you want to learn or your own. A person reads the results and decides. --- See the method applied in one real interview: describe what you want to learn and Tracy drafts the guide. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=how-to-run-customer-interviews&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=how-to-run-customer-interviews&utm_content=demo) ## Sources - [Tracy AI Interviewer: Customer research](https://tracyinterview.ai/use-cases/customer-research), retrieved 9 Nov 2026. - [Tracy AI Interviewer: Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. # How to Write Interview Questions an AI Can Follow Up On > AI interview questions that earn a follow-up: how to write them, six rewrites of weak questions and how to set the depth. Includes a candidate's view. - URL: https://tracyinterview.ai/blog/interview-questions-for-ai-interviews - Published: 2026-11-13 - Author: Tracy team > **In short** **Tracy AI Interviewer** drafts questions from your job description, asks each person in their own field and follows up on what they say, so questions get answered in depth and not recited. > > A good AI interview question asks about one event, one decision and one part, so there is something to follow up on. > > A recording, a timestamped transcript and a written assessment for each question come back, and a person decides. > > Next step: send us one role and read the follow-ups Tracy asks. Most AI interview questions fail because nobody said anything that happened. A hiring manager pastes twelve questions into an interview tool. "Tell me about yourself." "What are your strengths?" "Where do you see yourself in five years?" The answers come back fluent, similar and impossible to tell apart. There is nothing in them to ask a second question about. AI interview questions have one extra job compared with a human script. Tracy AI Interviewer listens to the answer, waits for the end of the thought and asks about the part that matters. That works when the answer contains something specific. This page shows how to write questions that do, with rewrites of weak ones and a way to test your set before you send it. The [AI interviewer explained](/blog/what-is-an-ai-interviewer) covers what she does, and [how AI interviews work](/blog/how-do-ai-interviews-work) walks through a session. Candidates: skip to the last section. ## A good AI interview question has something to hold on to A follow-up needs an anchor: a fact, a choice or a number in the answer. Four features make an anchor likely: - **It asks about something that happened.** "Describe the last time a launch slipped" gets an event. "How do you handle deadlines?" gets a policy. - **It asks for a decision.** What did you choose, and what else could you have done? A decision has reasons, and reasons can be probed. - **It has one part.** "What was the problem, what did you do and what did you learn?" invites three short answers. One part gets one real one. - **It fits the person's own field.** A question about a model's grading rubric means nothing to a nurse. A question about a patient handover means nothing to a data engineer. Imagine the best answer you could expect. If you cannot think of a second question to ask it, the first question is the problem. ![Card: a weak question, a better question about a reversed decision, and Tracy's follow-up on how they knew it was wrong.](/blog/images/interview-questions-for-ai-interviews-1.png) *Tracy follows up on a question rewritten from a stalling one, so answers carry detail your reviewers can check. Illustration with sample data* Tracy asks a follow-up on a question rewritten from a stalling one. ## Write each question in four parts For each question, write four lines before you send it. For worked sets, see our [AI behavioral interview guide](/blog/ai-behavioral-interviews) and the question banks for [software engineers](/blog/software-engineer-interview-questions) and [AI trainers](/blog/ai-trainer-interview-questions). 1. **The question.** One part, about an event or a decision. 2. **A strong answer holds.** Two or three things, such as a named cost, a number or an owned mistake. 3. **Gaps to probe.** What the answer might leave out, which becomes the follow-up. 4. **A stopping point.** When you have enough and should move on. | Question | A strong answer holds | Probe if missing | | --- | --- | --- | | Describe the last project of yours that slipped. | A cause, a date, what they decided not to do | What did you tell your manager that week? | | Tell me about a decision you reversed. | What changed their mind, what it cost | How did you know the first decision was wrong? | | Walk me through how you would find why a report is wrong. | Where they start, what they rule out first | What would make you stop and ask someone? | You do not have to write every follow-up in advance. Tracy asks them from the answers. The probe column is for you, to check that a follow-up exists. ## Six weak questions and how to rewrite them Most weak questions fail in the same few ways. Here are six rewrites. | Weak question | Why it stalls | Rewrite | | --- | --- | --- | | What are your strengths? | A list of adjectives, nothing to probe | Tell me about something you did that a colleague could not have done as well. | | Where do you see yourself in five years? | A rehearsed answer | What would you want to be doing in 18 months that you cannot do now, and why? | | Are you good with deadlines? | Yes or no | Describe the tightest deadline you met. What did you cut? | | What would you do if a client was angry? | Hypothetical, so any answer is safe | Tell me about a client who was angry with you. What did you say first? | | Explain how a hash map works. | Recall, one right answer | You chose a hash map in a recent project. What else did you consider? | | Why do you want this job? | A script every candidate has | What in the job description did you read twice, and why? | Each rewrite swaps a general claim for something that happened, and a real event gives the follow-up somewhere to go. ## Avoid questions an AI cannot follow up well Some questions fail for any interviewer, human or AI: - **Yes or no questions.** They end the answer where it begins. - **Questions with several parts.** The person answers the easiest part. - **Trivia.** A definition tests memory. - **Leading questions.** "Don't you think testing matters?" teaches the answer. - **Questions about age, health, family plans or other personal characteristics.** They do not belong in a hiring interview. If you are unsure, ask your legal adviser for your region's rules. Keep the list short: fewer questions with deeper follow-ups. ## Set the depth to match the decision How deep Tracy goes is a setting. You choose a style from quick to deep, and from technical to behavioural. A first screen for a high-volume role can use the quick style. For a decision about an expert or an engineer we suggest fewer questions with deeper follow-ups. Pay is by length: $1 per credit, one credit per started 20 minutes, so 100 interviews of 20 minutes cost $100, and the first three results are free. See [pricing](/pricing). ## Where Tracy enters: from a job description to questions you can edit You do not have to write the set alone. Describe the role, or paste the job description, and Tracy drafts the questions. You edit them or bring your own, and set the skills and the weight each one carries for the role or role family. Each person takes the interview from a browser link at any hour, with no calendar and no account. She asks the question, listens, waits for the end of the thought and follows up on what the person said. A Director at a top-tier early-stage VC says Tracy "sometimes brings up exactly those questions I might have forgotten to ask, which has been a pleasant surprise." Afterwards your team reads a recording of both sides, a timestamped transcript and a written assessment for each question with its reasoning. A person decides. See [what the interview report shows](/blog/ai-interview-reports). Our [technical interviews](/use-cases/technical-interviews) and [people teams](/use-cases/people-teams) pages show where teams use her. ![Transcript card: Tracy asks about a project that slipped, the candidate cites a vendor API change, Tracy asks what they told their manager.](/blog/images/interview-questions-for-ai-interviews-2.png) *Tracy takes a rewritten event question and asks the follow-up, so answers carry detail your reviewers can check. Illustration with sample data* A question about an event gets a follow-up built from the answer. ## Test your questions before anyone else sees them A short test finds most problems: 1. **Answer each question aloud yourself.** If you cannot say a specific answer, the question is too general. 2. **Ask a colleague to answer one.** Write down the second question you want to ask. If none comes, rewrite it. 3. **Read the first five results.** Look for answers that all sound alike. That means the question is generic. 4. **Run a small first batch.** The first 3 results are free. You decide which questions are fair for the role, how the skills are weighted and what a strong answer is. Tracy asks and records. Prefer to start from a draft? Send us one role and we will build the interview plan in days. ## If you are a candidate: how to answer AI interview questions Searching for AI interview questions and answers as a candidate? You will not find a script that works, because the follow-up depends on what you say. What helps: - **Use a real event.** Pick one specific project, decision or mistake, with a date and a number if you have one. - **Say what you did.** "We" hides your part. Say "I" and name the step. - **Expect the follow-up.** It will ask about what you just said, and does not mean the first answer was wrong. - **Say what you do not know.** Honest and specific beats confident and vague. You need only a browser link, and you are told on the join screen that the conversation is recorded and transcribed. See [what the person interviewed sees](/participant-experience). ## Common questions ### What are good AI interview questions? Questions about something that happened or a decision the person made, with one part and a field the person works in. "Describe the last project of yours that slipped" gives an AI interviewer something to follow up on. "What are your strengths?" does not. ### Is there an AI interview questions generator? Yes. Describe the role or paste the job description into Tracy and she drafts the questions. You edit them or bring your own, and set the skills and weights. ### Are there AI interview questions and answers I can practise? You can practise question types, not scripts. Pick a real event and say what you did. The follow-up depends on your answer, so a memorised reply does not carry through. ### How do you write questions an AI interviewer can follow up on? Quote something from the answer and ask for the missing part: the reason, the cost, the number or what happened next. With Tracy, you write the first question and she asks the follow-up. --- Copy these questions into a Tracy interview and read the follow-ups: send us one role and we will build the interview plan in days, with a recording and report you can check. The first 3 results are free, with no card. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=interview-questions-for-ai-interviews&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=interview-questions-for-ai-interviews&utm_content=demo) ## Sources - [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 5 Nov 2026. - [Tracy AI Interviewer home page, customer testimonial](https://tracyinterview.ai/), retrieved 9 Nov 2026. # Interview Transcript Analysis: From Recording to Themes > Interview transcript analysis in six steps, from recording to themes, with a theme table that keeps quotes and timestamps, and where AI helps. - URL: https://tracyinterview.ai/blog/interview-transcript-analysis - Published: 2026-11-13 - Author: Tracy team > **In short** For customer research, Tracy AI Interviewer runs the interviews from your guide, asks follow-ups on what each person says and leaves a recording and timestamped transcript for every result. > > Below: six steps from recording to themes, and a theme table that carries quotes, counts and timestamps. > > In customer research use, each Tracy finding links back to the words and the moment in the recording. > > Next step: see Tracy run one interview from your guide. You have a folder of recordings and a deadline. Interview transcript analysis is the work of turning them into a short list of themes, each with the exact words behind it and a count of how many people said it. Done well, anyone can open a theme, find the quote, hear the moment and judge whether the theme is fair. Below is a method for interview transcript analysis you can run in a spreadsheet, the theme table to copy, and how [Tracy AI Interviewer](/product) produces transcripts that are ready for this work in customer research. ## Interview transcript analysis in six steps The steps are adapted from the six phases Braun and Clarke describe in a 2006 paper in Qualitative Research in Psychology: familiarise yourself with the data, generate codes, search for themes, review them, define and name them, then write up ([source](https://doi.org/10.1191/1478088706qp063oa)). In a spreadsheet it looks like this. | Step | What you do | What you end up with | | --- | --- | --- | | 1. Prepare | Fix obvious transcription errors, remove names, label each speaker and keep the timestamps. | A clean transcript per participant | | 2. Read | Read each transcript once, start to finish, before tagging anything. | A page of first impressions | | 3. Code | Tag each meaningful passage with a short label, in the participant's own words where you can. | A list of codes, each with the line it came from | | 4. Group | Cluster related codes and name each cluster with a sentence, not a topic. | Draft themes | | 5. Test | For each theme, find two or three quotes, count how many participants said it and look for a participant who said the opposite. | Themes with evidence and a count | | 6. Write up | State each finding as a claim, then the quotes, then the count. | A report someone can check | A code is a label for one passage. A theme is a claim about what the passages mean together. "Pricing" is a topic. "Buyers cannot tell what they will pay until the quote arrives" is a theme. ## A theme table with the words behind each theme Copy this layout into a sheet. It is the part of the work other people will read, so keep one row per theme and never a theme without a quote. | Theme | Said by | Quote | Participant and time | | --- | --- | --- | --- | | Reports take most of the week | 5 of 6 | "Most of the week goes on pulling the same export by hand." | P2, 02:20 | | Someone senior waits on the numbers | 4 of 6 | "The finance lead waits on me every Monday, and I am rarely ready." | P2, 06:10 | | Home-made fixes break | 3 of 6 | "I built a spreadsheet template, but it breaks when a column moves." | P2, 09:05 | Every row has a count, a quote and a place to find the moment. The counts and quotes are fictional. ## Where interview transcript analysis goes wrong - **Themes that are topics.** A list of subjects is not a finding. Write each theme as a sentence someone could disagree with. - **One loud participant.** Count participants, not mentions. Five mentions from one person is one voice. - **Quotes without a place.** A quote with no participant and time cannot be checked, and nobody can hear the tone. - **Leading follow-ups.** "Was it frustrating?" gives you the word you hoped for. "What happened next?" gives you the story. - **Guides that drift.** If each interviewer asks different questions, the transcripts cannot be compared. Fix the core questions first; see [how to screen candidates](/blog/how-to-screen-candidates) for the same idea applied to hiring. - **Reading the summary instead of the transcript.** A summary can drop the exceptions, so read the transcript before you rely on it. ## Interview transcript analysis software and AI: what to look for Software and AI can do the slow parts: transcribing, suggesting codes, grouping similar passages and drafting a summary. Treat each as a first draft. The test for any tool is whether a theme leads you to the exact words and the moment in the recording, so you can check it in seconds. A summary that you cannot trace back is a guess you cannot defend. ## Where Tracy fits: run the interviews and keep the evidence Analysis is easier when every transcript follows the same shape. In customer research use, Tracy runs the interviews so that it does. - **A guide from what you want to learn.** Tracy drafts an interview guide from what you want to learn, or uses yours, for discovery, concept tests and win-loss interviews. - **Participants join when it suits them.** They open a link in the browser, with no scheduling. Participants answer the guide you set, and Tracy follows up on what each one says, so the transcripts share a spine you can compare. - **Follow-ups on what was said.** She follows up and waits for the answer, which is where a vague first reply turns into a reason. A Director at a top-tier early-stage VC says: "Tracy sometimes brings up exactly those questions I might have forgotten to ask, which has been a pleasant surprise." - **Evidence per result.** A recording of both sides, a timestamped transcript and a written assessment per question with its reasoning. - **Findings you can check.** For customer research use, each Tracy finding links back to the words and the moment in the recording, so each row of your theme table has its quote and its time. A person still decides what the findings mean. See the [customer research](/use-cases/customer-research) page, [how an AI interview report reads](/blog/ai-interview-reports) and what participants see in the [participant experience](/participant-experience). To try it, describe what you want to learn in one study and run the first interviews. ![Transcript card: a participant says reports took too long, and Tracy asks which numbers and who was waiting.](/blog/images/interview-transcript-analysis-1.png) *Tracy follows up on a vague answer so the transcript holds the reason, not only the complaint. Illustration with sample data* The follow-up turns a vague answer into a reason you can quote. ![Sample research session report: the recording beside a timestamped transcript, and three questions with what the participant said and quoted evidence.](/blog/images/interview-transcript-analysis-2.png) *Customer research use: quotes from one session, each matched to a timestamp in the transcript, beside the recording. Illustration with sample data* Each quote sits beside its timestamp in the transcript, next to the recording. ## Which path fits you - Discovery, concept tests and win-loss interviews: [customer research](/use-cases/customer-research). - Check-ins and exit interviews: [people teams](/use-cases/people-teams). - Cost before you start: [pricing](/pricing). ## Common questions ### How do you analyse interview transcripts? Prepare a clean transcript, read it through, tag passages with short codes, group the codes into themes, test each theme against the data with quotes and a count of participants, then write each finding as a claim with its evidence. ### What is the difference between a code and a theme? A code labels one passage. A theme is a claim about what several passages mean together, written as a sentence. ### Can AI do interview transcript analysis? It can transcribe, suggest codes and draft a summary. Check each theme against the words and the moment in the recording before you rely on it. ### How many interviews do you need? There is no fixed number. Keep going until new interviews stop adding new themes, and report how many participants said each one. ### How do you write a good interview transcript summary? Lead with the findings, give each a count and one or two quotes with the participant and time, and list the exceptions. --- See how Tracy runs one interview from your guide. Tell us what you want to learn and we will set up the interview guide in days, with a recording and report you can check. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=interview-transcript-analysis&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=interview-transcript-analysis&utm_content=demo) ## Sources - [Braun and Clarke (2006), Using thematic analysis in psychology, Qualitative Research in Psychology 3(2), 77-101](https://doi.org/10.1191/1478088706qp063oa), retrieved 9 Nov 2026. - [Tracy AI Interviewer home page, customer quote](https://tracyinterview.ai/), retrieved 9 Nov 2026. - [Tracy AI Interviewer customer research use case](https://tracyinterview.ai/use-cases/customer-research), retrieved 9 Nov 2026. # Speed to Submit: How Staffing Firms Interview Every Candidate the Same Day > Speed to submit in staffing is lost before the submit button. See how firms interview every candidate the same day and send a package clients can check. - URL: https://tracyinterview.ai/blog/staffing-speed-to-submit - Published: 2026-11-13 - Author: Tracy team > **In short** A requisition can be lost while candidates wait for a first call. Tracy AI Interviewer starts each interview on click, at any hour, so every candidate who gets the link can interview the day they apply. > > She asks follow-ups in the candidate's own field and writes an assessment for each question; the recruiter reads a package with resume match, assessment, strengths, gaps and recording, then submits. > > Next step: pilot it on one requisition this week and time each step from application to submit. Picture a client who sends a requisition late on a Thursday. By Friday morning other agencies may have submitted. Your applicants are in the system, and your recruiters can call a few of them before the weekend. Speed to submit in staffing is decided before the submit button, in the time between a candidate applying and a recruiter hearing them speak. This guide covers where that time goes, what a same-day interview has to capture, and how Tracy AI Interviewer runs it. ## Where does speed to submit get lost? The submit step is quick; the time goes before it. - The first call waits for a recruiter. Applications arrive in the evening and at weekends, and the first screen happens when someone has an hour. - Scheduling adds a round trip. A candidate who is at work at nine and free at eight needs a slot that suits both sides. - A no-show restarts the clock. A missed call means a reminder, a new time and another day. - The recruiter screens a few and guesses at the rest. The candidates who get a call are often the ones who applied first, and the best candidate may be further down the pile. - A thin submit costs time too. A client who asks why this candidate and gets a CV summary sends the submittal back. A same-day interview removes the first four, and the package answers the fifth. Our guide to [screening when there are too many applicants](/blog/how-to-screen-candidates) covers the screening side. ## What must a same-day interview capture? A fast interview that leaves a thin record only moves the delay to the client's side. Five things make a submittal hold up. 1. The candidate's own account of the work, in the role's field, not a yes or no. 2. Follow-ups on what they said, so a rehearsed first answer does not carry the interview. 3. The skills the client cares about for that requisition, each with an assessment and the reasoning behind it. 4. A record the client can check: the recording and a timestamped transcript beside the assessment. 5. A recruiter's decision before submit, with the evidence in front of them. [How AI interviews work](/blog/how-do-ai-interviews-work) explains the mechanics behind the first three. ## How does Tracy interview every candidate the same day? The interview starts when the candidate clicks. There is no scheduling and no recruiter present, so an application at ten at night gets the same interview as one at ten in the morning. The candidate needs a browser and a link, with no account, and the join screen says the conversation is recorded and transcribed. The interview is ready in under 10 seconds. A candidate who drops out can rejoin within 10 minutes (configurable per customer), a reminder goes out after a day, and failed interviews are chased automatically, which replaces the no-show restart. The [participant experience](/participant-experience) page shows the candidate's side. You set up from the job description or bring your own questions, and choose the skills to evaluate for each requisition as tags. With the CV check on, she reads the CV first. She asks follow-ups in the candidate's own field and waits for the answer. Create the interview for your next requisition and put the link in your first message to applicants. ![Four steps on a line: candidate applies, interview starts on click, an assessment per question with evidence, package sent.](/blog/images/staffing-speed-to-submit-1.png) *Tracy moves a candidate from application to a client-ready package without waiting for a recruiter's calendar. Illustration. Illustration.* Nothing between apply and the package waits on a recruiter's calendar. Each candidate gets a written assessment for every question, with the reasoning, so the recruiter reads evidence from everyone who took the interview and not only the first applicants. The package for each candidate holds resume match, the assessment per question, strengths and gaps, the recording and the fit rationale in one record, shared by link or PDF. [AI interview reports](/blog/ai-interview-reports) shows how a reviewer reads one. ![Tracy client package in three parts: resume match and assessment, strengths and gaps, recording and transcript.](/blog/images/staffing-speed-to-submit-2.png) *Tracy gives the client a package they can check: assessment per question, strengths and gaps, and a recording with a timestamped transcript behind every result. Illustration. Illustration.* The client gets a resume match, an assessment per question, strengths and gaps, and a recording and timestamped transcript. ## What stays with your recruiters? - The decision to submit. Tracy writes the assessment, and a recruiter reads the package and decides. A recruiter can move a candidate forward or invite them to interview again, and the history stays on the record. - Reading the integrity signals. Tab changes, gaze, live face detection, delay before answers, fullscreen exits and external screens arrive as evidence for a person, never as proof, and nothing is decided in the interview. - The client conversation: why this candidate, availability and pay. - Entering the submission in the client's system. Tracy's pages state no ATS write-back. She connects through API, MCP, Zapier, Slack, WhatsApp and the Workable and Breezy HR chat connectors, so a team on another ATS should plan the hand-off. [Connect](/connect) lists the routes. ## What have customers seen? On the candidate side, Steven Perry, Account Executive at IT Resources, an IT staffing firm in Tampa, US, says: "By the time they finish with this platform, they're even laughing and joking." One Tracy customer cut its average time to hire from 44 days to 7, and Tracy did the work of 53 full-time recruiters in one month. That figure is time to hire, so measure time to submit directly in your own pilot. The [volume hiring page](/use-cases/volume-hiring) shows how that kind of intake is set up. ## How do you pilot same-day interviews in one week? 1. Pick one requisition type you fill often, and one owner, such as a delivery lead or a senior recruiter. 2. Put the interview link in the first message to every applicant, including the ones you would not have called. 3. Time each step: application, interview complete, package read, submitted. 4. Stop if a recruiter cannot decide from the package without calling the candidate. The interview plan needs to change, or the tool is not the fit. The first 3 results are free, then $1 per credit, with one credit per started 20 minutes, so 100 interviews of 20 minutes cost $100. There are no seats and no upfront cost (see [pricing](/pricing)). Our [staffing agencies comparison](/blog/ai-interview-tools-for-staffing-agencies) shows where other tools fit. ## Which numbers show it is working? - Hours from application to interview complete. This is the headline number. - Hours from application to a submit-ready package. - The share of applicants who finish the interview. - The share of submittals the client moves to its own interview. - Recruiter time per submittal. The number of interviews run is the vanity metric. A high count with slow submits means the results are not reaching recruiters. ## Common questions ### What is speed to submit in staffing? It is the time from a candidate applying, or a requisition arriving, to the recruiter submitting that candidate to the client. The clock starts at the application, so it includes every wait before the recruiter speaks to the candidate. ### How can a staffing firm submit faster without lowering quality? Put the interview link in the first message so each applicant can interview the day they apply, read the evidence and send the client a package that answers why this candidate. Recruiters then spend their time on the candidates the evidence supports and on the client conversation. ### Can candidates interview outside business hours? Yes. With Tracy every candidate starts on click, at any hour, from a browser link with no account and with no recruiter present. ### What happens when a candidate drops out? A candidate who drops out can rejoin within 10 minutes. ### Does it connect to our ATS? Tracy works through API and MCP, Slack, WhatsApp, Zapier and the Workable and Breezy HR chat connectors. ATS write-back is not stated, so test the hand-off into your ATS in the pilot. --- Create an interview for one requisition and review the results. Prefer a plan built for you? Send us one role and we build it, customised in days. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=staffing-speed-to-submit&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=staffing-speed-to-submit&utm_content=demo) ## Sources - [Tracy AI Interviewer, Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Volume hiring](https://tracyinterview.ai/use-cases/volume-hiring), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Connect](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. - [Tracy AI Interviewer, customer proof on the home page](https://tracyinterview.ai), retrieved 9 Nov 2026. - [Tracy AI Interviewer, Home](https://tracyinterview.ai), retrieved 9 Nov 2026. # Tracy and Breezy HR: Add AI Interviews Next to Your ATS > Breezy HR AI interviews: keep candidates in Breezy HR and invite shortlisted people to a Tracy AI Interviewer interview from the same chat. - URL: https://tracyinterview.ai/blog/tracy-breezy-hr-integration - Published: 2026-11-13 - Author: Tracy team > **In short** **Tracy AI Interviewer** works next to Breezy HR in one chat: your candidates stay in Breezy HR, and shortlisted people are invited to a Tracy interview from the same conversation. > > Tracy shows who she would invite and waits for your approval before anything is emailed. Results stay in Tracy, where the chat can read them. > > It runs in Claude or ChatGPT with both connectors added. Pay as you go has no seats, and the first 3 results are free. > > Next: add Tracy to your chat app, then ask for your shortlist. A recruiter who tracks candidates in Breezy HR wants an interview step without moving the pipeline. This page describes the Breezy HR AI route Tracy AI Interviewer documents: keep your candidates in Breezy HR, and invite them to an interview from the same chat. Tracy runs the interview. She asks follow-ups in the person's own field, waits for the answer and leaves a recording, a timestamped transcript and a written assessment per question behind each result. The person interviewed needs only a browser link. How the interview works is in [how AI interviews work](/blog/how-do-ai-interviews-work). Breezy HR's own AI features are Breezy's to describe, and this page does not cover them. ## How the two fit together The guide describes the setup as a chat with two connectors: Tracy AI Interviewer and the connector Breezy HR provides for the same chat app. The guide names Claude and ChatGPT. You ask the chat for the people you want and for the interview to send them. What each side holds: - **Breezy HR** keeps the candidates. It shares each candidate's name and email, which is what Tracy needs to invite them. - **Tracy AI Interviewer** holds the interview and its results. The guide says results stay in Tracy, and you ask the chat to read them from there. The guide states no write-back from Tracy to Breezy HR. If you want a result noted on a candidate's Breezy record, a person adds it. ## Set up Breezy HR AI interviews in three steps The public guide gives three steps. 1. **Add Tracy AI Interviewer** to the AI app you chat in. Each app has its own guide, and the [Connect page](/connect) links both. 2. **Add the Breezy HR connector.** Open the same app and add the MCP connector that Breezy HR provides for it. Follow Breezy HR's own instructions, because its connector name, address and plan requirements are Breezy HR's to state. 3. **Ask for both in one chat.** Have an interview set up in Tracy first, or ask the chat to set one up. Tracy shows who she would invite and waits for your approval before anything is emailed. ![Four steps: add Tracy, add Breezy HR, ask in one chat, approve the preview.](/blog/images/tracy-breezy-hr-integration-2.png) *Tracy AI Interviewer next to the Breezy HR connector: you add both, ask in one chat, and approve the invite preview before anything is emailed. Illustration with sample data* Two connectors, one chat, one approval. ## Three prompts to start with The guide offers three prompts. Swap in your own job and interview names. - "List the candidates shortlisted for our Customer Success Manager job in Breezy HR. Then invite them all to my customer success interview in Tracy AI Interviewer with invite_participants, and show me the preview before anything is sent." - "Take the newest candidate in Breezy HR for our Customer Success Manager job and invite them to my customer success interview in Tracy AI Interviewer. Show me the preview first." - "Which of the candidates I invited from Breezy HR have finished my customer success interview? Read their results from Tracy AI Interviewer and summarise each one." The first suits a shortlist. The second suits a single new applicant. The third closes the loop, reading results back from Tracy. ![Flow: shortlist, ask in chat, approve preview, invite emailed.](/blog/images/tracy-breezy-hr-integration-1.png) *A sample chat with Tracy AI Interviewer and Breezy HR: ask for the shortlist, check the preview, approve, and Tracy emails each person their link. Illustration with sample data* Shortlist in Breezy HR, invite from the chat. ## Before you invite a shortlist A shortlist invite is a real email to real people, so a few habits help. - **Read the preview.** The approval step exists so you check who is invited and to which interview. - **Read the interview first.** Every person invited will be asked its questions. Review it in the [web app](/product) before you invite a list. - **Start with one candidate.** Use the second prompt on a test applicant, ideally yourself, before a whole shortlist. - **Keep the decision with a person.** The chat moves names and reads results. A recruiter opens the recording and report and decides. [AI interview reports](/blog/ai-interview-reports) shows what is there. - **Mind credits.** Credits are used by interviews that produce a result: first 3 results free, then $1 per credit, no seats. The [pricing page](/pricing) explains them. ## What this route does not do Be clear on the limits before you build on this route. - **It needs a chat app.** The guide covers Claude and ChatGPT with both connectors added. A recruiter who does not chat with an AI app will not use it. - **Breezy HR sets its own terms.** Its connector name, address and plan requirements are Breezy HR's to state, so confirm them with Breezy before you plan the rollout. - **Names and emails only.** Breezy HR shares each candidate's name and email, which is what an invite needs. The guide describes nothing beyond that. - **Results are read, not synced.** Ask the chat to read them from Tracy. The guide describes no sync back to Breezy HR. Within those limits the route removes a real chore: copying a shortlist into another tool and sending each invite yourself. ## Who this suits This route suits a team that has three things: Breezy HR as the system of record, a chat app such as Claude or ChatGPT already in use, and a first-round step that is slow by hand. It fits [volume hiring](/use-cases/volume-hiring) and [applications](/use-cases/applications) well, because the invite goes to a whole shortlist in one request. Write-back to Breezy HR is not stated in the guide, so if you need results written into Breezy HR automatically, ask us first. The route is also chat-driven; for flows that fire on their own events, the [Connect page](/connect) lists the automation routes. ## Common questions ### Does Breezy HR use AI? Breezy HR works with AI chat apps such as Claude and ChatGPT through the connector it provides, which is how the Tracy guide has you use it. Its own AI features are not covered here. Tracy adds an AI interview next to it: you invite shortlisted candidates from a chat that has both connectors. ### Where do candidates stay? In Breezy HR. It shares each candidate's name and email, which is what Tracy needs to invite them. ### Where do interview results live? In Tracy. You ask the chat to read them from there. The guide states no write-back to Breezy HR. ### Does Tracy email people without my approval? No. Tracy shows who she would invite and waits for your approval before anything is emailed. ### What does it cost? Credits are used by interviews that produce a result: first 3 results free, then $1 per credit, no seats. Breezy HR and your chat app set their own pricing, not covered here. --- Add Tracy to your chat app, connect Breezy HR and draft one interview for a shortlist. Or send us one role and we will build the interview plan in days, with a recording and report you can check. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=tracy-breezy-hr-integration&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=tracy-breezy-hr-integration&utm_content=demo) ## Sources - [Tracy AI Interviewer Connect page](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. - [Use Tracy AI Interviewer with Breezy HR (public guide)](https://tracyinterview.ai/help/breezy), retrieved 9 Nov 2026. - [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 5 Nov 2026. # Tracy in Microsoft Copilot Studio: Give Your Agent an Interview Tool > Copilot Studio HR agent: add Tracy AI Interviewer as an MCP tool at tracyinterview.ai/mcp so your agent can draft interviews and read results. - URL: https://tracyinterview.ai/blog/tracy-copilot-studio-integration - Published: 2026-11-13 - Author: Tracy team > **In short** **Tracy AI Interviewer** runs as an MCP server at tracyinterview.ai/mcp, so a Copilot Studio agent can set up an interview and read its results when you ask in a chat. > > You add it once as a tool, sign in with your Tracy account and approve. Nothing is created until you approve the draft, and a person decides from the evidence. > > A data policy on Power Platform connectors governs this server too. Pay as you go has no seats, and the first 3 results are free. > > Next: add the tool and send the first prompt. A Copilot Studio HR agent can get an interview step from Tracy AI Interviewer. Copilot Studio connects an agent to Tracy as an MCP tool, so the agent can set up an interview and read its results when you ask for them in a chat. Tracy runs the interview itself. She asks follow-ups in the person's own field and waits for the answer. A recording, a timestamped transcript and a written assessment per question stay behind each result. Your agent drafts and reads. A person decides. [How AI interviews work](/blog/how-do-ai-interviews-work) covers the interview itself. ## The server address and what to check first Paste this as the Server URL. When Copilot Studio asks for a name, use Tracy AI Interviewer. `https://tracyinterview.ai/mcp` The public guide notes two things to check first. Copilot Studio supports the Streamable transport for MCP servers, and Tracy uses it. And Copilot Studio reaches an MCP server through Power Platform connectors, so a data policy that governs connectors governs this server too. If adding it is blocked, ask whoever manages your organization's data policies. ## Add Tracy to a Copilot Studio HR agent in six steps The steps below come from the Copilot Studio guide. 1. Open your agent in Copilot Studio and go to its Tools page. 2. Select Add a tool, then New tool, then Model Context Protocol. The MCP onboarding wizard opens. 3. Fill in Server name with Tracy AI Interviewer, a Server description, and the Server URL above. The guide suggests this description: Runs AI interviews, on video or audio, for screening candidates, customer and user research, reference checks and employee check-ins. 4. For authentication, choose OAuth 2.0, then Dynamic discovery as the Type. Tracy registers clients and publishes its sign-in endpoints, so there is no client ID, client secret or URL to enter. 5. Select Create, then Next. In the Add tool dialog, select Create a new connection, finish the Tracy sign-in and approve access. 6. Select Add to agent. Microsoft's own guide to connecting an agent to an existing MCP server covers the Copilot Studio side. ![Four steps: open Tools, enter the address, choose OAuth, connect.](/blog/images/tracy-copilot-studio-integration-1.png) *Tracy AI Interviewer in Copilot Studio: add an MCP tool, enter the server address, choose OAuth with dynamic discovery, then sign in and add it to the agent. Illustration with sample data* Add the tool, point it at one address, sign in once. ## Sign in, approve and send a first prompt You sign in on a Tracy page with your own Tracy account. It asks for one permission, named tracy, and acts only on your own account. Then start a new chat with your agent and send the guide's first prompt: "Use Tracy AI Interviewer to set up a customer discovery interview for people who cancelled in the last month, then give me the public link." Tracy answers with a draft: the title, the questions and the result fields. Nothing is created until you approve the draft, so read it, ask for changes and say yes when it is right. The same pattern works for a screening interview. Describe the role, check the questions, then approve. Teams that screen [applicants](/use-cases/applications) can start the same way. ![Flow: ask the agent, draft shown, you approve, public link.](/blog/images/tracy-copilot-studio-integration-2.png) *A sample chat in a Copilot Studio agent: ask for an interview, read the draft, approve it, and receive a public link. Illustration with sample data* Ask, read the draft, approve, share the link. ## Check it worked and fix common problems Two checks in the guide confirm the connection. - Open your Tracy account and find Connected AI apps. The app is listed, with the time you connected it. - Open the agent's Tools page. Tracy AI Interviewer is listed as a tool, and its settings show the names and descriptions of the MCP tools it offers. If something is off, the guide lists the usual causes. - **Not on the Tools page.** Go through the steps again and finish with Add to agent. - **Blocked by a data policy.** Copilot Studio shows the tool as disabled. Select the information icon to see which policy applies, then ask your administrator. - **Sign-in not finished.** Select Create a new connection and sign in again. - **Agent not using the tool.** Turn the Enabled toggle on and select Save. Under Details, select Allow agent to decide dynamically when to use the tool. - **Access removed.** If you revoked Copilot Studio under Connected AI apps, it stopped at once. Connect it again. ## What the agent can do with Tracy The guide describes what the agent does once the tool is added: it sets up an interview and reads its results when you ask for them in a chat. Tracy shows a draft first, and nothing is created until you approve it. Two settings decide how freely the agent uses her. On the Tools page, the Enabled toggle switches the tool on. Under Details, Allow agent to decide dynamically when to use the tool lets the agent call Tracy on its own, instead of waiting to be told. If you turned off Allow all in the Tools section, each tool there has its own toggle. The person interviewed never touches Copilot Studio. They get a browser link and need no account, as the [participant experience](/participant-experience) page explains. The report, with its recording and transcript, is what your team reads. See [AI interview reports](/blog/ai-interview-reports). ## Who this suits This route suits an IT or HR-tech lead who already builds agents in Copilot Studio and wants interviews inside them. The agent calls Tracy as a tool, and you can remove access under Connected AI apps in your Tracy account. It fits [customer research](/use-cases/customer-research) as readily as screening, since the guide's own prompt is a customer discovery interview. Event-triggered flows are not covered by the guide. The [Connect page](/connect) lists the other routes, including the same server address for other AI apps. ## Common questions ### What is the server address? https://tracyinterview.ai/mcp. Paste it as the Server URL, and use the name Tracy AI Interviewer. ### Do I need a client ID or secret? No. Choose OAuth 2.0 with Dynamic discovery. Tracy registers clients and publishes its sign-in endpoints, so there is nothing to enter. ### Why is the tool disabled in Copilot Studio? A data policy that governs Power Platform connectors also governs this server. Select the information icon next to the tool to see the restriction, then ask your administrator. ### How do I remove access? Open your Tracy account, find Connected AI apps and revoke the connection. It stops at once, and you can connect again later. --- Add the tool to one agent and send the first prompt to draft an interview. Or send us one role and we will build the interview plan in days, with a recording and report you can check. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=tracy-copilot-studio-integration&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=tracy-copilot-studio-integration&utm_content=demo) ## Sources - [Tracy AI Interviewer Connect page](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. - [Add Tracy AI Interviewer to Microsoft Copilot Studio (public guide)](https://tracyinterview.ai/help/copilot-studio), retrieved 9 Nov 2026. - [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 5 Nov 2026. # AI Interviews in Gemini: Create and Review Interviews From a Chat > Run AI interviews in Gemini with Tracy AI Interviewer: add the custom app, ask in a chat, approve the tool and read results with evidence. - URL: https://tracyinterview.ai/blog/tracy-gemini-integration - Published: 2026-11-13 - Author: Tracy team > **In short** **Tracy AI Interviewer** connects to Gemini as a custom MCP server, so you set up an interview, invite people and read results by asking in a chat. > > Tracy still runs the interview on video, with follow-ups on what each person says, and leaves a recording, transcript and assessment per question for you to decide from. > > The Gemini app takes it today only for a personal Google Account in the US; Gemini Enterprise is set up by an administrator. > > Next step: connect Gemini and draft one interview. Most teams already keep a Gemini tab open. Running AI interviews in Gemini means you stop opening a second tool to build an interview: you ask for it in the chat, Gemini calls Tracy AI Interviewer, and the answer comes back where you asked. Think of it as a Gemini recruiting assistant for setup and review, while Tracy does the interviewing. Google offers custom MCP servers in two chat places, the Gemini app and Gemini Enterprise, and each has its own steps. Gemini CLI, the command-line tool, connects the same server with its own commands. This page covers all three, in the order the guide gives them, and says where each one has limits. ## What you can ask Gemini to do MCP, the Model Context Protocol, is a standard way for an AI app to call tools in another service. Tracy publishes one at tracyinterview.ai/mcp, and the [Connect page](/connect) lists Gemini next to the other AI apps that can add it, including [ChatGPT](/blog/tracy-chatgpt-integration). Tracy's guide gives one first prompt: "Use Tracy AI Interviewer to set up a customer discovery interview for people who cancelled in the last month, then give me the public link." Once connected, Gemini lists the actions it found on the server when you save the app. Through them you can set up an interview, invite people and read results by asking. Two rules from the guide keep this safe. Gemini asks for one permission, named tracy, and acts only on your own Tracy account. And to draft an interview, Gemini asks before it uses the tool: you choose Allow once or Always allow. ![Four steps: add Tracy as a custom app, ask in a Gemini chat, approve the tool, read results.](/blog/images/tracy-gemini-integration-1.png) *Gemini sets up and reads, Tracy interviews and keeps the evidence. No second tool to learn. Illustration with sample data. Source: setup steps from the public Tracy AI Interviewer Gemini guide.* Four steps take you from a Gemini chat to a result a person can review. ## Which Gemini you have decides the steps The Gemini app is for a personal Google Account. Gemini Enterprise is set up for a whole team by an administrator. For chat, Tracy's guide covers only those two, because Google's documentation for custom MCP servers names only those two. Gemini CLI is a separate command-line tool with its own steps, covered after Gemini Enterprise. No other Gemini surface or plan is covered here. Whichever you use, Gemini asks for a server address. Paste this wherever it asks for a server URL, and use Tracy AI Interviewer as the name: `https://tracyinterview.ai/mcp` Tracy needs you to sign in and approve it, and Gemini opens that sign-in for you. ## AI interviews in Gemini: set up the Gemini app Google calls these custom apps. For now you can add one only if all of this is true: - You are 18 or over and in the US. - You sign in with a personal Google Account. A work or school Google Account cannot add a custom app yet. - Keep Activity is on. Custom apps are unavailable when it is off. - You use Gemini in English, in the web app at gemini.google.com or in the Gemini mobile app. Once added on the web, the custom app is in both. Then, on your computer: 1. Go to gemini.google.com. Custom apps can only be added in the web app. 2. At the bottom, click **Settings > Connected Apps**. If you do not find Connected Apps, click Personal Intelligence first, then Connected Apps. 3. Click **Custom** at the top right, next to the search box. 4. Paste the server address as the app's MCP server URL. Leave Additional settings closed, because Tracy supports dynamic client registration, then click **Next**. 5. Read Gemini's warning, tick the box to accept the risks, then click **Connect**. 6. When Gemini opens the Tracy AI Interviewer sign-in, finish it and click **Allow** on the consent screen. 7. Gemini lists the actions it found on the server. Click **Connect** to save the app. Ready to try it? Connect Gemini and draft one interview. ## Set up Gemini Enterprise An administrator adds Tracy for the team. Google lists these limits: Gemini Enterprise reaches a server over Streamable HTTP only, and the server must sit at a secure public HTTPS address with a certificate signed by a publicly trusted authority. ### Business edition Google lists this as a Pre-GA offering, which might have limited support. 1. Sign in at business.gemini.google, click Settings & help, select your team, then open Manage team > Connected apps. 2. Click Add MCP Server, paste the server address as the Server URL, enter Tracy AI Interviewer as the Name and click Next. 3. Under Authentication settings choose OAuth 2.0, fill in the fields below, then click Verify Auth and Continue. 4. Choose a location and a connector name, then click Add. 5. The connection is disabled by default, so enable it to make it visible to your team. ### Google's Standard, Plus, Pay-as-you-go and Frontline editions of Gemini Enterprise These use the Google Cloud console. You need the Discovery Engine Editor role, and an organization policy has to allow custom MCP data stores, so an organization policy administrator may need to act first. Allow tracyinterview.ai as an egress domain, since Tracy serves the server, authorization and token hosts from one host. 1. In Gemini Enterprise, open Data stores, click Create data store, search for Custom MCP Server and click Add MCP server. 2. Choose OAuth 2.0, paste the server address as the MCP Server URL, select Enable PKCE Support (Tracy requires PKCE), then click Verify Auth and Continue. 3. Pick a location and a name, create the data store and wait for it to read Active. 4. Every action starts off, so open the data store and choose Actions > Reload custom actions. 5. Select the Tracy actions and click Enable actions. ### The OAuth 2.0 fields Register Gemini Enterprise as an OAuth client, then enter: - Redirect URL: `https://vertexaisearch.cloud.google.com/oauth-redirect` - Authorization URL: `https://tracyinterview.ai/oauth/authorize` - Token URL: `https://tracyinterview.ai/oauth/token` - Scopes: `tracy` Enter the URLs exactly as shown. One caveat from the guide: Google's forms also ask for a Client ID and a Client Secret. Tracy registers each app itself and issues no client secret, and Google's docs do not say what the form takes for a server like that. Setup may stop at this step. If it does, write to Tracy support rather than guess a value. ## Set up Gemini CLI Gemini CLI adds Tracy as a remote MCP server from a terminal. In a terminal, run: `gemini mcp add --transport http tracy https://tracyinterview.ai/mcp` This saves the server in your project's .gemini/settings.json. Add `--scope user` to save it in ~/.gemini/settings.json and have it in every project. Then start Gemini CLI and run `/mcp auth tracy`. Your browser opens on Tracy AI Interviewer. Sign in and approve access once, and the CLI keeps it. Sign-in needs a browser on the same computer as the CLI, so it will not finish over SSH or in a container without one. Check it with `/mcp`: tracy should read CONNECTED. Then send the first prompt from above. To remove it, run `gemini mcp remove tracy`. The [Connect page](/connect) has the full guide. ![Sample Tracy interview: Tracy asks about a production change, hears the answer, then asks a follow-up about it.](/blog/images/tracy-gemini-integration-2.png) *Follow-ups in the person's own words. You read the reasoning, not only the first answer. Illustration with sample data* Tracy follows up on what the person just said, inside the interview. ## A first prompt, and what happens next Once connected, start a new chat and send the prompt from the guide. In the Gemini app, type @ and pick the custom app first, so Gemini uses it. Gemini answers from your Tracy account, and for the draft it asks before it uses the tool. That is where Gemini's job ends. Tracy runs the interview itself. The person you invite opens a link in a browser, needs no account, and is told on the join screen that the conversation is recorded and transcribed. Tracy asks questions in their own field, follows up on what they say and waits for the answer. This works for [customer research](/use-cases/customer-research) calls as much as for hiring. Back in the chat, ask who has finished. Each result opens to a recording, a timestamped transcript and a written assessment for each question. Integrity signals, with their evidence, are there for a human reviewer to interpret. They are a prompt to look, not proof, and the decision stays with you. The [web app](/product) holds the same reports. ## What it costs Cost follows the web app. Pay as you go is $1 per credit, one credit per started 20 minutes of an interview that produced a result. The first 3 results are free with no card, there are no seats, and a session with no answers is never charged. Full API and MCP access is included, so there is no separate fee for the connection. The [pricing page](/pricing) has the current plan. ## Remove access, or fix a stuck setup - **Cut Gemini off.** Open your Tracy account, find Connected AI apps and revoke it. - **In the Gemini app.** Open Settings > Connected Apps, find the app under Custom apps, click More details, then choose Disconnect to revoke its permissions or Remove app to delete it. - **In Gemini Enterprise Business edition.** Select the MCP server in the connectors list and delete it. - **Connected but not used.** In the Gemini app, start the message with @ and the app's name. Without it, Gemini may not call Tracy. - **No Connected Apps entry.** Click Personal Intelligence first, then Connected Apps. Still stuck? Write to Tracy support. Teams on Slack can use the [Slack guide](/blog/tracy-slack-integration) instead, and candidates see what to expect on the [participant experience](/participant-experience) page. ## Common questions ### Can Gemini run the interview itself? No. Gemini sets the interview up and reads results. Tracy interviews each person on video in a browser, with follow-ups on what they say. ### Is this a Gemini AI interview for candidates to practise with? No. The connection is for the person running interviews. People who are interviewed open a link and need no account. ### Can I use a work or school Google Account in the Gemini app? Not yet. Per Tracy's guide, the Gemini app needs a personal Google Account, and a work or school account cannot add a custom app. Gemini Enterprise is the route for a team, set up by an administrator. ### Does Gemini get access to my other data? Gemini asks for one permission, named tracy, and acts only on your own Tracy account. You can revoke it under Connected AI apps in your account. ### Is there an extra charge for Tracy in Gemini? No separate charge. Pay as you go includes full API and MCP access, at $1 per credit. ### What if Gemini Enterprise asks for a client secret? Tracy issues none, and Google's docs do not say what to enter. Setup may stop there. Write to Tracy support rather than guess. --- Connect Gemini and draft one interview, or send us one role and we will build the interview plan, with a recording and report you can check. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=tracy-gemini-integration&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=tracy-gemini-integration&utm_content=demo) ## Sources - [Tracy AI Interviewer Connect page](https://tracyinterview.ai/connect), retrieved 10 Nov 2026. - [Add Tracy AI Interviewer to Gemini (public guide)](https://tracyinterview.ai/help/gemini), retrieved 10 Nov 2026. - [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 5 Nov 2026. - [Add Tracy AI Interviewer to Gemini CLI (Tracy guide)](https://tracyinterview.ai/help/gemini-cli), retrieved 10 Nov 2026. # AI Interviews With Zapier: Invite People and React When They Finish > Zapier interview automation with Tracy AI Interviewer: five triggers and an invite action, set up with one sign-in, plus three sample workflows. - URL: https://tracyinterview.ai/blog/tracy-zapier-integration - Published: 2026-11-13 - Author: Tracy team > **In short** **Tracy AI Interviewer** connects to Zapier with five triggers and an Invite Participant action, so invites go out and results reach your team without copying data by hand. > > The interview and report stay in Tracy: own-field follow-ups, a recording, a transcript, and a person decides. The person interviewed needs only a browser link. > > Zapier checks for events on a schedule, so a Zap starts a few minutes after the event. Pay as you go has no seats, and the first 3 results are free. > > Next: connect Zapier and invite your first person. An interview usually sits between two other tasks. Someone is added to a list, then someone sends an invite by hand. Someone finishes, then someone tells the team. Zapier interview automation removes both hand-offs. With Tracy AI Interviewer connected, a Zap can invite a person the moment they appear in another app, and another Zap can react when that person starts, finishes or has a result. Tracy does the interview itself. She asks follow-ups in the person's own field and waits for the answer. A recording, a timestamped transcript and a written assessment per question stay behind each result, so the people on the receiving end of a Zap still read evidence before they decide. Our guide to [how AI interviews work](/blog/how-do-ai-interviews-work) covers the interview itself. The Zapier route fits the workflow you already have: your other apps change by one step. ## The five triggers and the invite action Each trigger runs a Zap once per event. - **Invite Opened:** the person opened their invite link. - **Interview Started:** the person began the interview. - **Interview Completed:** the person finished. - **Interview Incomplete:** the interview ended without being completed. - **Result Ready:** the result can be read. The one action is **Invite Participant**. You choose the interview, enter an email, and can add a name and your own reference ID, such as a CRM record ID. Tracy emails the person their link. Every event carries the interview title, the person's name and email, and the interview link. Interview Completed and Result Ready also carry a link to the result. Your reference ID comes back on each event for that person, so a later Zap can match the event to the record that started it. A practical use of the reference ID: put your own record number in it when you invite, then use the same number in a later Zap to find the record that started the interview. Without it you are matching on an email address, which works until two records share one. ## Set up Zapier interview automation in four steps Before you start, have a Tracy account, a Zapier account and at least one interview. 1. In Zapier, add Tracy AI Interviewer as the app for a trigger or action step. 2. Click the sign-in button in that step. 3. Approve access in the Tracy window that opens. If it does not open, allow pop-ups. 4. Test the step, then turn the Zap on. Zapier checks for new events on a schedule, so a Zap starts a few minutes after the event, not at the same second. For deeper setup of your account, see the [Connect page](/connect), which lists every route into Tracy. Start with one Zap, such as the invite action, and add the triggers once that works. You can see and revoke the connection under Connected AI apps on your Tracy account at any time. ![Four steps: add the app, sign in, approve access in the Tracy window, test then turn the Zap on.](/blog/images/tracy-zapier-integration-1.png) *Tracy AI Interviewer connects to Zapier with one sign-in, so the owner can automate invites without a developer. Illustration, not a screenshot. Source: Tracy visual library, process template; steps from the public Zapier guide.* Connecting Zapier takes one sign-in, then a test and a switch. ## Three sample workflows These are examples with invented data, to show how the pieces fit. Your apps and fields will differ. **1. A new applicant row invites a person.** A staffing team adds a new row to its applicant sheet: Priya Nair, priya@example.com, reference REC-1042. A Zap runs Invite Participant for your "Sales recruiter screen" interview, passing the name, the email and REC-1042 as the reference ID. Priya gets her link by email. Teams that screen [applicants](/use-cases/applications) in batches save the copy and paste. **2. A finished interview posts to a channel.** Result Ready fires for Priya's interview. The Zap posts a message to a team channel: the interview title, her name, REC-1042 and the link to the result. The team opens a recording and transcript, and a person reads them. See [AI interview reports](/blog/ai-interview-reports) for what a reviewer finds there. **3. An unfinished interview prompts a follow-up.** Interview Incomplete fires for a second person. The Zap creates a task for a recruiter to nudge them. Because Tracy also sends reminders, keep the task light, and see [hiring at volume](/use-cases/volume-hiring) for how teams handle large lists. To try the first one, connect Zapier and invite yourself with your own email. ![Illustration: new row in a sheet, Invite Participant, Result Ready, post to a channel.](/blog/images/tracy-zapier-integration-2.png) *Tracy AI Interviewer in a sample workflow: a new row invites a person, and Result Ready tells the team the report can be read. Illustration with sample data. Source: Tracy visual library, flow template; sample workflow, illustration.* Two Zaps can cover the hand-offs on either side of the interview. ## Check these before you automate invites An invite is a real email to a real person, so test with care. - **Test with your own email.** Zapier's test on Invite Participant sends a real invite. Use your address. - **Check the interview is live.** An invite may be held if the interview is paused or credits are used up. The [pricing page](/pricing) explains credits. - **Read the interview first.** Every person invited will be asked its questions. Review it in the [web app](/product) before a Zap sends to a list. - **Pass a reference ID.** It lets later Zaps tie an event to your own record. - **Decide who reads results.** Zaps move data. A person still decides from the evidence. - **Watch the connection.** Reconnect if it expires, which can happen after revocation or after 60 days without contact from Zapier. You can see and revoke it under Connected AI apps on your Tracy account. ## Who this suits Three kinds of team get the most from it. - **Teams with a list in another app.** If people are already tracked somewhere else, a Zap turns a new entry into an invite, and the reference ID keeps the two records linked. - **Teams that want a nudge, not a dashboard.** A Zap that posts when a result is ready tells people who need to read it, without anyone checking a screen. - **Teams with a mixed process.** [People teams](/use-cases/people-teams) and others can automate the invite and leave every decision to a person. The person interviewed needs only a browser link, as described on the [participant experience](/participant-experience) page. If you would rather work from a chat, the other routes on the Connect page cover that. ## Common questions ### Which Tracy events can start a Zap? Five: Invite Opened, Interview Started, Interview Completed, Interview Incomplete and Result Ready. Each event starts a Zap once. ### How fast does a Zap start? A few minutes after the event. Zapier checks for new events on a schedule rather than at once. ### Does the Zapier connection cost extra? Invites sent through Zapier use your Tracy credits like any other invite: first 3 results free, no seats. Zapier's own plan is billed by Zapier. ### A Zap is not starting. What should I check? Check the Zap is on, that the connection has not expired, and that the interview is not paused or out of credits. --- Connect Zapier and invite your first person: you get a real interview link in your inbox. Or send us one role and we will build the interview plan in days, with a recording and report you can check. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=tracy-zapier-integration&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=tracy-zapier-integration&utm_content=demo) ## Sources - [Tracy AI Interviewer Connect page](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. - [Connecting Tracy AI Interviewer to Zapier (public guide)](https://tracyinterview.ai/help/zapier), retrieved 9 Nov 2026. - [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 5 Nov 2026. - [Tracy AI Interviewer home](https://tracyinterview.ai/), retrieved 9 Nov 2026. # How to Run a Coding Interview With Follow-Ups on Trade-Offs > Coding interview follow-up questions that test trade-offs, not recall: how to ask them, what a strong answer holds, and how to run them at scale. - URL: https://tracyinterview.ai/blog/coding-interview-follow-up-questions - Published: 2026-11-12 - Author: Tracy team > **In short** **Tracy AI Interviewer** runs a coding task inside the interview, then follows up on each candidate's approach and trade-offs, so engineers review reasoning and not only output. > > Candidates start on click at any hour. Your team reads a recording, a timestamped transcript and a written assessment for each question, and a person decides. > > Next step: run one coding task as a Tracy interview and read the follow-ups. Picture a candidate who writes a rate limiter that passes every test. The interviewer says "great" and moves on to the next problem. Nobody asks why the candidate chose a counter over a list of timestamps, or what happens under load. The first answer showed they had seen the problem. Coding interview follow-up questions show whether they can reason about it. This page gives four kinds of follow-up, example questions with what to listen for, and how Tracy AI Interviewer runs them for every candidate: a coding task in the same interview, then follow-ups on approach and trade-offs, with the recording and transcript for your engineers to review. ## How do you follow up on a coding answer? A passing solution shows the candidate has seen the problem. Tests check an output, not why the candidate chose a hash map or a queue, or what changes when the input grows a hundred times. In our view, ten minutes of follow-ups on one problem tell you more than a second problem does. Four kinds of coding interview follow-up questions reach the reasoning behind the code: - **The rejected option.** "You used a list of timestamps. Why not a counter that resets each minute?" It makes the candidate name an alternative and say what it costs. - **The changed constraint.** "Now there are 10,000 requests a minute from one user. What changes?" It shows whether they see where the design bends. - **The failure mode.** "Two servers handle the same user. What goes wrong?" It tests whether they think beyond one machine. - **The cost.** "What does this use in memory and time, and who pays when it is slow?" It tests whether they can put a rough size on it. A generic "what is the time complexity?" gets a recalled answer. A follow-up that quotes the candidate's own choice cannot be recalled, because that choice did not exist before they wrote it. ## Coding interview follow-up questions for four common problems Pick one problem and two follow-ups from the table, and write down what a strong answer contains before the first interview. | Problem | Follow up with | Listen for | | --- | --- | --- | | Rate limiter for an API | Why a list of timestamps and not a counter that resets each minute? Then: what happens at 10,000 requests a minute from one user? | Names the burst at the window edge. Gives a rough memory cost and a cheaper option such as a sliding counter, and says what accuracy it gives up. | | Cache with expiry | What does a reader see if an invalidation arrives late? | Separates stale reads from lost writes. Says which the product can tolerate and why. | | Remove duplicates from a sorted list | What if the list does not fit in memory? | Moves from in-place to streaming and says what order guarantee that needs. | | Slow SQL report | The query took two seconds yesterday and a minute today. What do you check first? | Starts with the plan and the data, not with a rewrite. Names an index and what it costs on writes. | These questions are our own examples. An engineer on your team should adjust each to the stack and the level. ## How many follow-ups should you ask per problem? How you ask matters as much as what you ask: - **Let the candidate finish.** The better part of an answer often comes after the first sentence. - **Quote their words.** "You said you would drop old timestamps. When?" shows you listened. - **Go two levels.** The first follow-up tests the choice and the second tests the cost. - **Ask every candidate the same opening problem.** The follow-ups differ, because they follow the answers. The starting point does not. Doing this well takes an experienced engineer in the room. The limit is time: your senior engineers have a few hours a week, and a hiring round can bring far more candidates than those hours cover. ## What makes a trade-off answer strong? Whatever the problem, a strong answer to a trade-off follow-up usually holds four things: - **It names a cost.** Memory, latency, complexity, a failure mode. "It is faster" is not a trade-off. - **It names the condition that would flip the choice.** "At this size a list is fine. Past about this many requests I would use a counter." - **It puts a rough size on it.** Thousands or millions, milliseconds or seconds. - **It says what the candidate does not know.** "I would measure that" is a better answer than a confident guess. A weak answer restates the first answer in more words, or names a technique with no cost attached. Both are easy to see in a recording and hard to recall in a debrief a week later. ## How Tracy runs the follow-up for every candidate Tracy runs a coding task inside the interview itself: live coding, SQL, debugging or multiple choice. The candidate talks through the approach in a coding room. She asks about what they said and wrote, waits for the end of the thought and follows up on the approach and trade-offs. Your engineers choose the problem, or describe the role and edit the questions she drafts, and choose the skills to evaluate. A Director at a top-tier early-stage VC says of founder interviews that Tracy "sometimes brings up exactly those questions I might have forgotten to ask, which has been a pleasant surprise." Candidates take the interview from a browser link, with no engineer on the call. Afterwards your team reads a recording of both sides, a timestamped transcript and a written assessment for each question with its reasoning. Tracy also shows integrity signals such as tab changes and delay before answers, with their evidence, to the reviewer, as signals for a person to interpret and never as proof. ![Card: Tracy asks what to check on a slow report; the candidate says the query plan; Tracy asks what an index costs on writes.](/blog/images/coding-interview-follow-up-questions-1.png) *Tracy follows up on the cost of a candidate's fix to a slow query, so engineers see whether they reason about trade-offs. Illustration with sample data* Tracy follows up on the cost of a candidate's fix. ![Transcript card: Tracy asks why a list of timestamps, the candidate names the burst, Tracy asks what changes at scale.](/blog/images/coding-interview-follow-up-questions-2.png) *Tracy follows up on a candidate's own design choice, so your engineers see whether they can reason about cost. Illustration with sample data* A follow-up built on the candidate's own answer shows whether they can reason about cost. ## Keep these calls with your engineers A follow-up shows reasoning. It does not replace your engineers' judgement about the person: - **Reading the code.** Style, clarity and how the solution would live in your codebase. Review the code itself, not only the discussion. - **Large system design.** It needs a longer conversation with a person. - **The decision.** A recruiter decides: they can move a candidate forward or invite them to interview again, and the history stays on the record. For a first round that covers every applicant, see [how an AI screening interview works](/blog/ai-screening-interviews). For a bank of problems to use inside it, see [software engineer interview questions with follow-ups](/blog/software-engineer-interview-questions). Our guide to [AI technical interviews](/blog/ai-technical-interviews) covers the formats, the [technical interviews use case](/use-cases/technical-interviews) shows the flow, and [what Tracy does](/product) and [AI interview reports](/blog/ai-interview-reports) cover the product and what your team reads afterwards. ## Pilot it on one role family Start with one role family, one owner and a stop condition. 1. **Pick the role and the owner.** One backend role, one engineering manager. 2. **Choose one problem and two follow-ups from the table.** Write the strong answer for each first. 3. **Run 20 candidates through it.** Have two engineers read the same five results and say who they would advance. 4. **Compare with your current screen.** Note who advanced here that the old screen would have dropped. **Stop condition.** If two engineers agree on who advances no more often than they did from the old screen, rewrite the follow-ups and rerun on the next 20 before widening. At $1 per credit and one credit per started 20 minutes, 20 candidates at 20 minutes cost $20, and the first three results are free (see [pricing](/pricing)). Want it set up for you? Send us one role and we will build the interview plan in days. ## Common questions ### What are good coding interview follow-up questions? Ask about the candidate's own choice: why this structure and not another, what changes if the input grows a hundred times, what breaks with two servers, and what it costs in memory or time. Each should refer to what they said or wrote. ### How many follow-up questions should you ask per problem? Two levels per decision: one on the choice, one on its cost. More turns the interview into an interrogation. ### Can an AI interviewer ask coding follow-up questions? Yes. Tracy AI Interviewer runs a coding task inside the interview and follows up on approach and trade-offs. Your engineers read a recording, a timestamped transcript and an assessment for each question, and a person decides. ### What if the candidate uses an AI assistant? State your rule in the invitation. Tracy shows tab changes and delay before answers to the reviewer as evidence, never as proof, and a follow-up on the candidate's own choice shows whether they can explain the code. --- Run one coding task as a Tracy interview and read the report. The first 3 results are free, with no card. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=coding-interview-follow-up-questions&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=coding-interview-follow-up-questions&utm_content=demo) ## Sources - [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer home page, customer testimonial](https://tracyinterview.ai/), retrieved 9 Nov 2026. # Interview Questions for Finance Experts on AI Training Projects > Finance expert interview questions for AI training projects: 18 questions with follow-ups, strong and weak answers, and how to run them at volume. - URL: https://tracyinterview.ai/blog/finance-expert-interview-questions - Published: 2026-11-12 - Author: Tracy team > **In short** **Tracy AI Interviewer** puts every finance applicant through an interview in their own specialty and follows up on what they said, so reviewers see how they reason, not only what they know. > > Use the 18 finance expert interview questions below, grouped by what they test. > > Each result keeps a recording, a timestamped transcript and an assessment per question, with integrity signals and their evidence for a person to review. > > Next step: paste these questions into a Tracy interview. A finance review project lands on a Monday. By Wednesday you have hundreds of applicants whose CVs say CFA, ex-bank or Big Four, and two senior reviewers who can interview them between other work. The CV says where someone trained. It does not say whether they will catch a model that quotes a margin without saying which one. Finance expert interview questions for AI training projects have to test that gap. These ask for finance reasoning applied to review work: reading a model's answer, writing the criteria for judging it, and knowing where an answer should stop short of advice. Each has a follow-up and a note on what a strong answer contains. They are for heads of expert supply and talent teams approving analysts, accountants and risk officers; job seekers can use them to prepare. They test reasoning only: they are not investment, tax or accounting advice, and the Tracy result is interview evidence, not a credential check, so verify credentials separately. Tracy AI Interviewer runs the part that does not scale. She asks each expert in their own field, follows up on what they just said and waits for the end of the thought. A recording, a timestamped transcript and a written assessment per question go to your reviewers, who decide. For the generalist version, see [AI trainer interview questions](/blog/ai-trainer-interview-questions), and for the wider process, [expert vetting](/use-cases/expert-vetting). ## Eighteen finance expert interview questions, grouped by what they test Pick one or two from each group. Ask the follow-up only after the first answer, write down what a strong answer contains before the first expert joins, and give everyone the same opening questions. Questions that quote a model's answer should keep the quote to a sentence. ### Group 1: reasoning with missing information | Ask | Follow up with | Listen for | | --- | --- | --- | | 1. A user asks a model "Is this company a good investment?" and pastes one line about revenue growth. What do you want to know first? | Which answer would change your view most? | Margins, cash, debt, valuation, time horizon, who is asking. Ranks by impact. | | 2. Revenue grew and operating cash flow fell. What do you test first? | What would you open to tell a slow payer from early revenue booking? | Working capital, receivables, one-offs. Names a document: receivables ageing, revenue policy note. | | 3. Tell me about accounts where your first read was wrong. | What would you have needed to see earlier? | A real case, a named turning point, a check they now run. | ### Group 2: reviewing a model's answer | Ask | Follow up with | Listen for | | --- | --- | --- | | 4. A model says "Margin improved from 12 to 15 percent, so profitability rose." What do you check first? | What would make you reject it if both numbers are right? | Which margin, one definition and period, margin versus absolute profit. | | 5. What separates a wrong answer from a misleading one? | An example from your own reviewing? | A defensible line. A correct figure can mislead through what it leaves out. | | 6. A model gives a discounted cash flow value to two decimals. How do you grade it? | What would you want instead? | Assumptions stated, a range, discount rate and terminal growth named. Calls out false precision. | | 7. A model quotes a figure "from the latest annual report". How do you check? | Which year, entity and accounting basis? | Opens the filing. Checks period, currency, restatements, group against parent, reported against adjusted. | ### Group 3: scope, advice and confidentiality | Ask | Follow up with | Listen for | | --- | --- | --- | | 8. When should an answer decline to say what to buy or sell? | A case where declining is too cautious? | A view on both failures. Explaining factors can help; "see an adviser" can say nothing. | | 9. What do you do when a task is outside your specialty? | How would your lead know the flag was fair? | Flags it with a one-sentence reason. Does not guess to keep the queue moving. | | 10. What would you never put in a task or chat about a client or employer? | How do you build a realistic rubric case? | Invents the numbers, removes names, keeps non-public information out. | ### Group 4: rubrics and agreement | Ask | Follow up with | Listen for | | --- | --- | --- | | 11. Pick a question small investors ask. Write three criteria for judging an answer. | Which could two reviewers read differently? | Criteria scorable yes or no. Spots their own vague wording. | | 12. A rubric line says "mentions relevant risks." What is wrong with it? | Rewrite it so a second reviewer scores it the same. | "Relevant" hides the judgement. Splits into named, checkable risks. | | 13. You and another reviewer score a valuation answer far apart. What now? | When do you hold your score? | Compares reasons, not numbers. Holds for a fact, moves for taste. | ### Group 5: specialty depth and currency | Ask | Follow up with | Listen for | | --- | --- | --- | | 14. What rule of thumb do people apply where it does not hold? | How would you spot a model doing it? | Something specific, such as a ratio that misleads for loss-making or heavily leased businesses. | | 15. What did you last change in your work because a standard or rule changed? | How did you judge that source? | A named source and a concrete change. | | 16. Where does treatment differ between countries or accounting frameworks in your field? | How would you mark an answer that fits only one? | A real difference known first-hand. Marks down answers that do not say where they apply. | ### Group 6: working on the project | Ask | Follow up with | Listen for | | --- | --- | --- | | 17. Tasks arrive in batches against one rubric. How do you score the 200th the way you scored the fifth? | What did you do when you last saw drift? | A routine: reread the rubric, spot-check early scores, report gaps. | | 18. Do you hold a role, position or restriction that could conflict with this work? | What do you do when a task names a company you cover or hold? | Discloses unasked, passes the task back, keeps non-public information out. | Ask question 18 of everyone. Hours and time zone are better collected in the application. ## A strong answer names what it would check, and a weak one stops at the fact The difference shows up in the second answer. These pairs are written by us as examples of how it sounds, not quotes from real experts. | | Weak answer | Strong answer | | --- | --- | --- | | Question 4, the margin claim | "The numbers match the report, so I'd score it high." | "They match, but one is gross and one is operating margin. I'd ask for the same measure in both years." | | Question 6, the single-number valuation | "The method is right, so it's fine." | "Two decimals on a number that moves with the discount rate says more than the inputs support. I'd want a range and the assumptions." | | Question 18, conflicts | "No, I don't think so." | "I cover two listed software companies. If a task names either, I'd tell you and hand it back." | Listen for ownership. "The team signed off" tells you about the team. "I checked this, and this is what I do now" tells you about the person. ## One follow-up chain shows where the real answer starts This is an illustration with sample data, not a real interview. The expert is fictional and nothing they say is financial advice. It shows the shape of what Tracy runs: an opening question from your list, the answer, and a follow-up built on it. It starts from question 2. > **Tracy:** Revenue grew and operating cash flow fell. What do you test first? > > **Expert (fictional):** Working capital. Receivables usually explain it. > > **Tracy:** What would you open to tell a slow payer from early revenue booking? > > **Expert:** The receivables ageing, then the revenue policy note. Ageing shows who is late. The note shows if terms changed. > > **Tracy:** Tell me about the last time that was the answer. > > **Expert:** A distributor whose sales rose after it moved to longer payment terms. Cash fell for two quarters. I asked for the contract terms before I scored anything. The first answer is the textbook one. The third is where you learn whether it ever happened. A Director at a top-tier early-stage VC says Tracy "sometimes brings up exactly those questions I might have forgotten to ask, which has been a pleasant surprise." ![Transcript card: revenue grew, cash flow fell. Expert says working capital. Tracy asks what separates slow payers from early booking.](/blog/images/finance-expert-interview-questions-1.png) *Tracy asks a follow-up built on the expert's own answer. The reviewer sees whether the expert has done the work. Illustration with sample data* A follow-up built on the expert's own answer shows whether they have actually done the work. ## Run the same interview for every applicant without losing the follow-up A live call for every applicant does not survive a project that draws hundreds of them. - **Your questions or ours.** Describe the role and Tracy drafts questions you edit (see the [product page](/product)). - **CV check, as a setting.** With it on, Tracy can read the CV and ask about what it claims, so a stated specialty becomes a question about it. - **Start on click.** Each expert opens a link in a browser, with no account and no calendar, at any hour, and is told on the join screen that the conversation is recorded and transcribed. See the [participant experience](/participant-experience). - **A result you can check.** A recording, a timestamped transcript and a written assessment per question sit behind every result. - **Integrity signals, with the evidence.** Tracy captures tab changes, gaze, face detection, delay before answers, fullscreen exits and external screens, and shows the evidence to a human reviewer. A signal is not proof, and your team decides. See [AI interview cheating detection](/blog/ai-interview-cheating-detection). It is $1 per credit, one credit per started 20 minutes of an interview that produced a result, so a 34-minute interview costs $2. The first 3 results are free with no card, there are no seats, and a session with no answers is never charged. See [pricing](/pricing). Customisation takes days: send us one role and we will build the interview plan. ![Recreated Tracy result report for a fictional finance reviewer: recording with timestamped transcript, and three fields marked Stated with quoted evidence.](/blog/images/finance-expert-interview-questions-2.png) *A recording and timestamped transcript sit behind every Tracy result. Your reviewer checks each stated point against the expert's own words before deciding. Illustration with sample data* A reviewer can read what the expert stated for each field, beside the recording and transcript. ## Keep four checks with people 1. **Credentials.** Verify a qualification, licence or registration separately, at the source in the expert's country. The Tracy result is interview evidence, not a credential check. 2. **A senior reviewer.** Have one of your own score five or six transcripts blind. If they disagree with the written assessments, fix the questions before inviting the next hundred. 3. **Conflicts and confidentiality.** Whether a stated conflict rules someone out is your call. 4. **The decision.** A person decides from the evidence. For ten experts a month, a senior reviewer on a video call is likely cheaper and better than any tool; see [AI interview vs human interview](/blog/ai-interview-vs-human-interview). The case for Tracy starts when applicants outnumber your reviewers' hours. Pilot on one specialty: 15 to 30 applicants, the same six opening questions plus question 18, and one reviewer. If that reviewer disagrees with more than a quarter of the assessments after fifteen, fix the questions first (our rule of thumb, not a researched figure). Measure completion, hours from application to result, reviewer agreement and integrity flags needing a closer look. See [AI interview reports](/blog/ai-interview-reports). ## Common questions ### What should you ask a finance expert in an AI training interview? Ask about reasoning with missing information, how they review a model's answer, when an answer should decline to advise, how they write a rubric criterion and where a rule of thumb stops working. Pick one or two from each group and keep the follow-up beside each. ### How do you test finance reasoning without giving financial advice? Ask about process, not about what to buy. Use invented companies and numbers, and treat nothing in the interview as guidance. ### Can an AI interviewer interview finance experts? Yes. Tracy asks questions in the expert's own field, follows up on what they said and waits for the answer. Your team gets the recording, transcript and an assessment per question, and a person decides. ### Does Tracy check a finance qualification or licence? A Tracy result is interview evidence, not a credential check. Verify qualifications and licences separately at the source that applies in the expert's country. ### Are there finance specialist interview questions for different specialties? Use the same six groups and change the cases to the specialty. A credit officer and an equity analyst should both be asked how they would check a quoted figure, but not on the same company. --- Create a Tracy interview for this exact role from these questions, then read the recording and report; the button opens the Connect page. Or send us one role and we will build the interview plan, with a recording and report you can check. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=finance-expert-interview-questions&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=finance-expert-interview-questions&utm_content=demo) ## Sources - [Tracy AI Interviewer home page, customer proof](https://tracyinterview.ai/), retrieved 9 Nov 2026. - [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer Connect page](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. # AI Interviews for Founders: A First Call for Every Application > A founder interview for every applicant: what it covers, eight questions to start from, a worked example and the report a partner reads. - URL: https://tracyinterview.ai/blog/ai-interviews-for-founders - Published: 2026-11-11 - Author: Tracy team > **In short** **Tracy AI Interviewer** gives every applicant a first conversation from one link in your application form. > > She asks your diligence questions, follows up on the answer and leaves partners a recording, transcript and assessment per question. > > A Director at a top-tier early-stage VC went from around six hours a day of founder screening to about 1.5 hours. > > A partner decides. Next: create an interview for your programme. A programme opens applications on a Monday, and by Friday the inbox holds more founders than the partners have hours to meet. Founders with a polished form can get the calls. A founder with a sharper story and a thinner write-up never reaches one. A form cannot ask a founder to explain a number, and a partner cannot meet everyone. Tracy AI Interviewer gives every applicant that first conversation. She asks about the company, follows up on what the founder says and waits for the end of the thought. Partners read a recording, a timestamped transcript and a written assessment for each question. Our [applications page](/use-cases/applications) shows the flow from form to report, and our guide to the [accelerator interview](/blog/accelerator-interview) covers the interview from the programme's side. **Use Tracy for the first founder interview when** applications arrive in a rush, when partners meet only the first slice of them, or when you want every founder asked the same opening questions. **Skip it when** your cohort is small enough for partners to meet every founder. A form and a call are enough then. ## What does a founder interview need to find out? Techstars says its staff personally review every application, and a Managing Director there says the first interview aims to understand a company well enough to write an investment memo. He asks founders to be ready on four things: the customer and the pain, speed and execution, market size, and what is different. Y Combinator describes 10-minute conversations with 2 to 3 partners who have read the application, and says the first question is usually what the company is working on. A form can hold those answers. It cannot ask a founder what they learned from the customers who left, and that follow-up is the job of the first interview. See [one-way versus two-way video interviews](/blog/one-way-vs-two-way-video-interviews) for the format difference. ## Eight opening questions for a first founder call Use six to eight, in any order, and ask the follow-up only after you have heard the first answer. The table is ours, adapted to the four areas above. Tracy takes your own questions, so you can load these as they stand. For a longer bank, see [founder interview questions](/blog/founder-interview-questions). | Ask | Follow up with | Listen for | | --- | --- | --- | | What is the company working on, and who is it for? | Who paid you last, and what were they using before? | A named buyer and the thing it replaced. | | What did the last customer who left tell you? | How many left for that same reason? | A count and a cause, not a mood. | | What did you build first, and what did you leave out? | What did leaving it out teach you? | Choices made on evidence. | | A number moved last quarter. What changed? | What would it have been if you had changed nothing? | Whether they know what caused it. | | How big is the market, built from the bottom up? | Which assumption would you most like to test? | Arithmetic they can say aloud. | | What is different about this, and why now? | What would a well-funded competitor do first? | A real change in the world and a real answer to it. | | Who does what on the team, and who is missing? | What have you not hired for yet, and why? | Honesty about gaps. | | What would you do with the first six months of the programme? | What would tell you it was going wrong? | A plan with a stop condition. | Give every founder the same opening questions so partners can compare them. ## How does Tracy run a founder interview from your form? You put one link in your application form. A founder opens it in a browser, with no account, at any hour, and is told on the join screen that the conversation is recorded and transcribed. Tracy asks your questions, or questions she drafted from your description of what you want to learn, which you edit freely. She follows up on what the founder said and gives them room to describe what they are building and where they want to take it. Your form stays as it is, and you add the link, as the [product page](/product) shows. ![Recreated Tracy application page: a founder recording beside Tracy's question on trial sign-ups and a short form.](/blog/images/ai-interviews-for-founders-1.png) *Tracy runs the first conversation from the application page, so every founder is asked about their company before a partner reads the form. Illustration with sample data* A founder pitches to Tracy from the application page, and she asks about what they said. ## A worked example: one churn question and its follow-up This is an illustration with sample data. The founder and company are fictional. > **Tracy:** What did the last customer who left tell you? > > **Founder (fictional):** Setup took them a week. We cut it to a day. > > **Tracy:** How many of the customers who left had that same week of setup? > > **Founder:** Four of the seven who left. The other three left for price. The sample assessment reads: "Gave a cause and a count: four of seven left over a week of setup, three over price. Had cut setup to a day." ![Tracy asks what the last customer who left said, then asks how many customers left for that reason.](/blog/images/ai-interviews-for-founders-2.png) *Tracy follows up on a churn answer by asking for a count. Illustration with sample data* The follow-up turns a story about churn into a count the partner can weigh. ## What did the VC change? Before Tracy, the VC held a 30-minute Zoom interview with every founder, sometimes eight back to back in a day. The director says Tracy "asks meaningful, investment-relevant questions and follows up on important points", and advises teams to "spend time designing the depth and structure of your interview questions during the initial setup." The question table above is a place to start. ## What does the partner read afterwards? For each founder, the partner gets a recording of both sides and a timestamped transcript and a written assessment for each question with its reasoning. One record can go to an investment committee by link or PDF. A person decides who moves on: they can move a founder forward or invite a founder to interview again. See [AI interview reports](/blog/ai-interview-reports). ## Will founders find an AI interview impersonal? Some partners say founders should be met by a person, and they are right about the second round. Tracy is the first conversation for everyone, so partners spend their hours on the best founders. Candidates rate Tracy 92 on NPS. Test the feel with your own cohort before you roll it out. ## How to pilot a founder interview on one cohort Pilot it on one cohort. Run one application window with one partner as owner and set the stop condition first. 1. Put the link in your form for one track and load six to eight questions. 2. Have the owner read the reports for a sample of founders and compare them with who the partners would have picked from the form alone. 3. Stop if the owner would not change the shortlist after five reports. 4. Measure founders who finished, partner minutes per applicant and founders no partner would have met. Tracy costs $1 per credit, one credit per started 20 minutes, so 100 interviews of 20 minutes cost $100. The first 3 results are free with no card, there are no seats and a session with no answers is never charged. Try one on the [Connect page](/connect), and see [pricing](/pricing). Which path fits you: - Programme, fellowship or scholarship applications: [applications](/use-cases/applications). - Experts to vet for a project: [expert vetting](/use-cases/expert-vetting). - Hundreds of applicants in a rush: [volume hiring](/use-cases/volume-hiring). ## Common questions ### What is a founder interview? A conversation in which a programme or investor learns what a company does, who pays, how the founders think and why now. Programmes such as Y Combinator hold a 10-minute one after the application. Tracy AI Interviewer runs that first conversation for every applicant. ### What are good founder interview questions? Ask about the customer, what customers who left said, what changed in the numbers, how the market is sized and why now. Follow each answer with a question that asks for a count or an example. ### Can an AI interviewer run a first call with founders? Yes. Tracy asks your questions, follows up on what the founder says and leaves partners a recording, a timestamped transcript and a written assessment per question. A partner decides. --- Create an interview for your programme. Send us your application form and we will build the interview plan in days, with one link for the form and a recording and report a partner can check. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=ai-interviews-for-founders&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=ai-interviews-for-founders&utm_content=demo) ## Sources - [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 5 Nov 2026. - [Tracy AI Interviewer Connect page](https://tracyinterview.ai/connect), retrieved 5 Nov 2026. - [Tracy AI Interviewer product page](https://tracyinterview.ai/product), retrieved 5 Nov 2026. - [Techstars, How to prepare for your first Techstars interview (22 January 2026)](https://www.techstars.com/blog/founder-advice/how-to-prepare-for-your-first-techstars-interview), retrieved 3 Nov 2026. - [Y Combinator, interview guide](https://www.ycombinator.com/interviews), retrieved 3 Nov 2026. - [Tracy AI Interviewer home page, customer testimonials](https://tracyinterview.ai/), retrieved 9 Nov 2026. - [Techstars, Inside a Techstars Accelerator: The Application & Interview Process (12 January 2026)](https://www.techstars.com/blog/advice/inside-a-techstars-accelerator-the-application-and-interview-process), retrieved 3 Nov 2026. # AI Interviews for Software Engineers > An AI interview for software engineers: what it covers, the coding task and follow-ups, what the report shows and where engineers still decide. - URL: https://tracyinterview.ai/blog/ai-interviews-for-software-engineers - Published: 2026-11-11 - Author: Tracy team > **In short** Engineering teams have more applicants than interview hours. **Tracy AI Interviewer** runs the first technical conversation for every software engineer: questions on their own stack, a coding task and follow-ups on approach and trade-offs. > > A recording, timestamped transcript and assessment per question let an engineer decide from the candidate's own words. > > A Tracy customer cut its average time to hire from 44 days to 7. > > Engineers keep system design, team fit and the offer. Next: create an interview for one role. Engineering teams get more applicants than they have engineers to interview them. The usual compromise is a take-home or a timed test, which shows an output and hides the thinking, followed by a phone screen that takes engineer time. An AI interview for software engineers moves the conversation to the start. Every applicant is interviewed on their own stack, asked to solve a problem and asked why. Tracy AI Interviewer does that. She asks questions in the candidate's own field, follows up on their approach and trade-offs and waits for the end of the thought. Your engineers read a recording, a timestamped transcript and a written assessment per question, and spend their hours on the people worth meeting. Our guide to [AI technical interviews](/blog/ai-technical-interviews) covers the format in general. **Use Tracy for the first technical round when:** - applicants outnumber the engineers who can interview them; - you hire for the same role again and again; - you want each candidate asked about their own stack, not a shared quiz. **Keep a person for** the system-design round, the team-fit conversation and the offer. ## What does an AI interview for software engineers cover? Describe the role and Tracy drafts the questions from the job description, or bring your own. Each item below is a setting on the role, as the [product page](/product) shows. - **Your stack and your skills.** You pick the skills the role must show as tags, such as backend design, databases or debugging, per role or for a family of roles. - **Style.** Interview style runs from quick to deep and from technical to behavioural. - **A coding task inside the interview.** Live coding, SQL, debugging or multiple choice, followed by questions on approach and trade-offs. - **CV check, if you turn it on.** Tracy can read the CV and ask about what it claims, so a listed system becomes a question about that system. - **A spoken-language check,** when spoken communication matters for the role. It evaluates fluency, coherence, grammar, pronunciation and vocabulary on a 0 to 100 scale beside the transcript. For a bank of questions to start from, see [software engineer interview questions](/blog/software-engineer-interview-questions). ## What do follow-ups find that a test does not? A test scores what a candidate produced. A follow-up asks about the part they did not write down. | Opening question | Follow-up | What it shows | | --- | --- | --- | | Your endpoint is slow after a release. Where do you look first? | The query plan looks the same as before. What next? | Whether they have a method or a list of guesses. | | Your cache returns stale data after a write. How do you fix it? | Two services write at once and an invalidation arrives late. What does a reader see? | Whether they understand the trade-off behind the fix. | | Walk me through how you would split this service in two. | Which half would you be sorry to own at 3 a.m., and why? | Operating experience as well as design. | | You chose a hash map here. What would make you pick a sorted array? | How does that change if the input does not fit in memory? | Judgement about approach, tested in the coding task. | | Tell me about a bug you shipped. | What did you change in how you review after that? | Whether they learned something and can say what. | The first answer is the one a candidate prepared. The follow-up depends on what the candidate said, which is why a fixed form cannot ask it. ## A worked example: a coding task and the follow-up This is an illustration with sample data, not a real interview. The candidate is fictional. In this illustration, Tracy sets two tasks and asks the candidate to talk through their approach, then turns to a design question. > **Tracy:** Your cache returns stale data after a write. How do you fix it? > > **Candidate (fictional):** Invalidate the key on write, with a short expiry as a backstop. > > **Tracy:** Two services write at once and an invalidation arrives late. What does a reader see? > > **Candidate:** Stale data until the expiry. I would version the keys so a late invalidation cannot remove a newer value. A sample excerpt of the written assessment, written by us: | Question | Assessment (sample) | | --- | --- | | Task 1, merge overlapping intervals | Sorted by start, then merged in one pass, and said why sorting first makes one pass enough. Named the case of touching intervals and chose to merge them. | | Follow-up, input too large for memory | Proposed sorting chunks and merging them. Missed duplicates across chunks until asked. | | Cache invalidation | Gave write-time invalidation with an expiry. On the late-invalidation follow-up, described stale reads and proposed versioned keys. | ![Recreated Tracy coding room: a merge intervals task, a Python editor and Tracy's chat beside two video tiles.](/blog/images/ai-interviews-for-software-engineers-1.png) *Tracy runs the coding task and the conversation in one interview, so approach is asked about while the code is on screen. Illustration with sample data* The coding task and Tracy's conversation sit in one interview. ![Tracy asks how to fix stale cache data, then asks what a reader sees when two writes race.](/blog/images/ai-interviews-for-software-engineers-2.png) *Tracy asks about the trade-off behind a fix, so an engineer sees whether the candidate can explain a fix as well as produce it. Illustration with sample data* The follow-up asks about the trade-off behind the fix, using the candidate's own words. ## What evidence does an engineering lead get? An engineer reading the result can check it. - A recording of both sides and a timestamped transcript behind every result. - A written assessment for each question, with its reasoning. - Fields marked Stated, each with the candidate's own words as evidence. - One record to send on: the summary, the assessments and the recording, shared by link or PDF. - A person's last word: Tracy rejects no one, and a reviewer decides and can invite a candidate to interview again. See [AI interview reports](/blog/ai-interview-reports) for how to read one. For integrity, Tracy captures tab changes, abnormal eye tracking, live face detection, delay before answers, fullscreen exits and external screens, and shows the evidence to a reviewer. A signal is a prompt to look, never proof. Candidates are told about the recording on the join screen, and nothing is decided in the interview. See [AI interview cheating detection](/blog/ai-interview-cheating-detection) for how to handle a flag. ## How do you run it for hundreds of applicants from your own stack? Each applicant starts on click, at any hour, with no scheduling and no recruiter present. The interview is ready in under 10 seconds. A dropped candidate can rejoin within 10 minutes (customers can configure this), a reminder goes out after a day and failed interviews are chased automatically. Three ways to start an interview fit an engineering pipeline: an invite link, an apply page, or your own system through the API. Tracy also works beside Workable and Breezy HR through their chat connectors, so candidates stay in the ATS and shortlisted people are invited from chat. The [Connect page](/connect) has each guide, and [technical interviews](/use-cases/technical-interviews) shows how teams set this up. ## What stays with your engineers? Tracy runs the first technical conversation. Your engineers keep three things. 1. **System design and team fit,** in a live round with someone senior. 2. **A blind read.** Before you trust the assessments, have one engineer read five or six transcripts and judge them without seeing Tracy's assessments. If they disagree, change the questions first. 3. **The decision.** A person decides from the evidence. Tracy publishes no bias audit, so treat the report as evidence to weigh and not a verdict. For a few candidates a month, an engineer on a video call is likely cheaper; see [AI interview versus human interview](/blog/ai-interview-vs-human-interview). ## What does it cost, and how do you pilot it? Tracy costs $1 per credit, and one credit covers each started 20 minutes of an interview that produced a result. A 34-minute interview takes 2 credits, which is $2, and 100 interviews of 20 minutes cost $100. The first 3 results are free with no card, there are no seats and no upfront cost, and a session with no answers is never charged. See the [pricing page](/pricing) for the current plan. Pilot it on one role: name an engineer as owner, send the link to 20 applicants, and compare the reports with who a phone screen would have passed. Stop if the owner would change fewer than a quarter of the decisions (our rule of thumb). Try one role on the [Connect page](/connect); the first 3 results are free. Which path fits you: - Engineers and AI labs hiring technical roles: [technical interviews](/use-cases/technical-interviews). - Hundreds of applicants for one role: [volume hiring](/use-cases/volume-hiring). ## Common questions ### What is an AI interview for a software engineer? A two-way video interview run by an AI interviewer. It asks questions on the candidate's own stack, can include a coding task, and follows up on what the candidate says. The hiring team reads a recording, a transcript and a written assessment. ### Can an AI interview for software engineers include coding? Yes. Tracy can run a coding task inside the same interview, in live code, SQL, debugging or multiple choice, then ask follow-ups about approach and trade-offs. ### How do you stop candidates reading answers from another screen? Tracy captures tab changes, gaze, face detection, delay before answers, fullscreen exits and external screens, and shows the evidence to a reviewer. Treat a signal as a reason to look, and decide as a person. ### What does an AI interview for software engineers cost? $1 per credit, one credit per started 20 minutes. The first 3 results are free, and there are no seats. --- Create an interview for this exact role. Send us one role and we will build the interview plan in days, with a recording and report you can check. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=ai-interviews-for-software-engineers&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=ai-interviews-for-software-engineers&utm_content=demo) ## Sources - [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 5 Nov 2026. - [Tracy AI Interviewer Connect page](https://tracyinterview.ai/connect), retrieved 5 Nov 2026. - [Tracy AI Interviewer product page](https://tracyinterview.ai/product), retrieved 5 Nov 2026. - [Tracy AI Interviewer home page, customer testimonials](https://tracyinterview.ai/), retrieved 9 Nov 2026. # Backend Developer Interview Questions for Hiring Teams > Backend developer interview questions for hiring teams: 15 questions with follow-ups, strong and weak answers, and how to run them at volume. - URL: https://tracyinterview.ai/blog/backend-developer-interview-questions - Published: 2026-11-11 - Author: Tracy team > **In short** **Tracy AI Interviewer** interviews every backend candidate in their own field and follows up on design trade-offs, so you see how they reason, not only which tools they have used. > > Use the 15 backend developer interview questions below, each with a follow-up and what to listen for. > > Every result keeps a recording, a timestamped transcript and a written assessment per question; your engineers decide. > > Next step: paste three of these questions into a Tracy interview. You open a backend role and 200 applications arrive. Most list the same stack, and the CV cannot tell you who has run a service under load and who has read about it. Backend developer interview questions have to find that gap, so ask what they did when their queue backed up. These 15 backend developer interview questions are written for engineering managers and recruiters hiring mid-level to senior backend developers. Each has a follow-up and a note on what a strong answer contains. They suit a first technical conversation of 20 to 30 minutes; deep architecture and coding belong to a later round with your own engineers. Tracy AI Interviewer runs that first conversation at volume. She asks each candidate in their own field, follows up on what they just said and waits for the end of the thought. For the wider process, see [technical interviews](/use-cases/technical-interviews), and for general engineering questions, [software engineer interview questions](/blog/software-engineer-interview-questions). ## Fifteen backend developer interview questions, grouped by what they test Pick one or two from each group. Ask the follow-up only after the first answer. Choose your opening questions for the role and write down what a strong answer contains before the first candidate joins. ### Group 1: APIs and failure *Table 1: APIs and failure, with follow-ups and what a strong answer contains.* | Ask | Follow up with | Listen for | | --- | --- | --- | | 1. A payment request times out. How do you decide whether to retry it? | What makes the call safe to repeat, and where do you record that? | Backoff, a request key, idempotency. Names what a duplicate would cost. | | 2. How do you design an API change that existing clients must survive? | What do you do when one client cannot upgrade for a quarter? | Versioning, deprecation notice, additive changes. A real client they had to protect. | | 3. A downstream service is slow, not down. What happens to yours? | How would you find out it is happening before users do? | Timeouts, circuit breakers, queues filling. Metrics they would watch, not a general "monitoring". | ### Group 2: data and storage *Table 2: Data and storage, with follow-ups and what a strong answer contains.* | Ask | Follow up with | Listen for | | --- | --- | --- | | 4. Walk me through how you chose a database for your last service. | What did the choice make harder? | A trade-off named without prompting. Consistency, query shape, operations, cost. | | 5. A query that took 50 ms now takes 4 seconds. What do you do first? | What if the plan has not changed? | Reads the plan, checks indexes and data growth, then locks and load. Measures before changing. | | 6. How do you change a column on a table with millions of rows without downtime? | What do you do if the migration fails halfway? | Expand and contract, backfill in batches, a way back. Has done it. | ### Group 3: design trade-offs *Table 3: Design trade-offs, with follow-ups and what a strong answer contains.* | Ask | Follow up with | Listen for | | --- | --- | --- | | 7. When would you add a cache, and when would you refuse to? | What goes wrong when it serves stale data? | Invalidation as the real cost. Says when correctness matters more than speed. | | 8. Design a rate limiter for a public API. What do you ask before drawing anything? | How does it behave across several servers? | Asks who is limited, burst rules, failure mode. Counter versus window, and the cost of each. | | 9. Where would you use a queue instead of a direct call? | How do you handle a message that always fails? | Decoupling, retries, dead-letter handling, ordering. Knows delivery is not exactly-once by default. | ### Group 4: production and debugging *Table 4: Production and debugging, with follow-ups and what a strong answer contains.* | Ask | Follow up with | Listen for | | --- | --- | --- | | 10. Tell me about the last incident you helped fix. | What would have caught it earlier? | A specific timeline, their own role, a change they made afterwards. | | 11. Memory use climbs for a week, then the service restarts. How do you investigate? | How do you tell a leak from normal growth? | Heap profiles, comparing builds, load patterns. Does not guess. | | 12. What do you log, and what do you never log? | How would you trace one request across three services? | Request ids, structured logs, no secrets or personal data. Has used tracing. | ### Group 5: working with others *Table 5: Working with others, with follow-ups and what a strong answer contains.* | Ask | Follow up with | Listen for | | --- | --- | --- | | 13. Describe a review where you pushed back on a design. | Were you right? | Gives their reasoning and how it ended, including being wrong. | | 14. How do you decide something is ready to ship? | What do you do when the deadline says ship and your tests say no? | Checks, rollout plan, a way to turn it off. Says it to the team early. | | 15. What have you stopped doing in the last two years? | Why? | Something specific and a reason they changed their mind. | ## A strong answer names what it would check, and a weak one names the tool Listen to the second answer, where the two diverge. We wrote these pairs to show how it sounds; none is a quote from a real candidate. *Table 6: Weak and strong answers to the same questions, written as examples.* | | Weak answer | Strong answer | | --- | --- | --- | | Question 1, the timeout | "I'd add a retry with backoff." | "Backoff, but only if the call is safe to repeat. For a payment I'd send a request key so a second attempt returns the first result." | | Question 5, the slow query | "I'd add an index." | "I'd read the plan first. If the data grew, an index may help. If it did not change, I'd look at locks and concurrent load." | | Question 10, the incident | "The team fixed it quickly." | "I found it was a missing timeout. I added it, then added an alert on queue depth so we'd see it before users." | Listen for ownership. "We" tells you about the team. "I checked this, and this is what I do now" tells you about the person. ## One follow-up chain shows where the real answer starts This is an illustration with sample data, not a real interview. The candidate is fictional. It shows the shape of what Tracy runs: an opening question from your list, the answer, and a follow-up built on it. It starts from question 1. > **Tracy:** A payment request times out. How do you decide whether to retry it? > > **Candidate (fictional):** I retry with backoff, but only if the call is safe to repeat. > > **Tracy:** What makes this call safe to repeat, and where would you store that? > > **Candidate:** The client sends a request key. The server stores it with the result and returns the same result on a repeat. > > **Tracy:** What happens if two requests with the same key arrive together? > > **Candidate:** The second should wait or fail on a unique constraint, not run the payment twice. The first answer is the one in every tutorial. By the third, you can hear whether the candidate has built it. ![Transcript card: a payment times out, candidate says retry with backoff, Tracy asks what makes the call safe to repeat.](/blog/images/backend-developer-interview-questions-1.png) *Tracy asks a follow-up built on the candidate's own answer. Your engineers hear whether the candidate has built it. Illustration with sample data* A follow-up on the candidate's own answer separates having read about retries from having built them. ## Run the technical interview for every applicant A live call for every applicant does not survive 200 applications. - **Your questions or ours.** Describe the role and Tracy drafts questions from the job description, which you edit; see the [product page](/product). - **A task inside the interview.** Live coding, SQL, debugging or multiple choice can sit in the same interview, with follow-ups on approach and trade-offs. See [AI technical interviews](/blog/ai-technical-interviews). - **CV check, as a setting.** With it on, Tracy can read the CV and ask about what it claims. - **Start on click.** Each candidate opens a link in a browser, with no account and no calendar, at any hour. They are told on the join screen that the conversation is recorded and transcribed. See the [participant experience](/participant-experience). - **A result you can check.** A recording and a timestamped transcript sit behind every result, with a written assessment per question. See [AI interview reports](/blog/ai-interview-reports). - **Integrity signals, with the evidence.** Tracy captures tab changes, gaze, face detection, delay before answers, fullscreen exits and external screens and shows the evidence to a person. A signal is not proof. It is $1 per credit, one credit per started 20 minutes of an interview that produced a result. The first 3 results are free with no card, and there are no seats. See [pricing](/pricing). ![Recreated Tracy result report for a fictional backend developer: recording, transcript with the 11:49 answer on repeated requests, and quoted evidence per question.](/blog/images/backend-developer-interview-questions-2.png) *A recording and a timestamped transcript sit behind every Tracy result. The engineer who reads it checks the exact answer before moving a candidate forward. Illustration with sample data* Your engineers can open the recording and the transcript behind each answer before they decide. ## Keep three decisions with your engineers 1. **Who moves forward.** A senior backend engineer reads the report and decides. Tracy writes an assessment for each question, with quoted evidence; a person has the last word. 2. **The calibration.** Have one of your engineers read five transcripts blind. If they disagree with the written assessments, change the questions before inviting the next hundred. 3. **The deep round.** Architecture whiteboarding and pairing stay with your team. For five hires a quarter, a call with your own engineer is simpler and probably better; see [AI interview vs human interview](/blog/ai-interview-vs-human-interview). The case for Tracy starts when applicants outnumber your engineers' hours, as in [volume hiring](/use-cases/volume-hiring). To try it, put three of these questions into one open backend role, send the link to your next batch of applicants and have an engineer read what comes back. ## Common questions ### What are good backend developer interview questions? Ask about failure handling, data and storage, design trade-offs, production debugging and working with others. Pick one or two from each group and keep the follow-up beside each, because the follow-up is where the real answer starts. ### What backend developer interview questions suit experienced engineers? Use the design trade-off and production groups. Experienced candidates should name what a choice made harder and describe an incident with a timeline and their own part in it. ### How do you interview a backend developer without a live call? Run a structured first-round interview: Tracy asks in the candidate's field, follows up on what they said and returns a recording, transcript and assessment per question for your engineers to review. ### Can Tracy include a coding or SQL task? Yes. A coding task, SQL, debugging or multiple choice can sit inside the same interview, with follow-ups on approach and trade-offs. ### Does Tracy decide who to hire? No. A person reads the evidence and decides. --- Create an interview for this exact role from these questions, then read the recording and report; the button opens the Connect page. Or send us one role and we will build the interview plan, with a recording and report you can check. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=backend-developer-interview-questions&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=backend-developer-interview-questions&utm_content=demo) ## Sources - [Tracy AI Interviewer home page](https://tracyinterview.ai/), retrieved 9 Nov 2026. - [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer Connect page](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. # micro1 Zara Interview: What It Is and Who It Is For > The micro1 Zara interview explained: what micro1's AI recruiter asks, what the candidate gets, who it is for, and how to interview your own applicants. - URL: https://tracyinterview.ai/blog/micro1-zara-interview - Published: 2026-11-11 - Author: Tracy team > **In short** Tracy AI Interviewer runs a comparable interview on your own applicants: it asks in their field, follows up, waits, and leaves a recording, a timestamped transcript and an assessment per question. > > The micro1 Zara interview is micro1's own screen for experts who apply to work for it; micro1 says it asks open-ended technical and scenario questions, and a coding challenge for technical roles. > > Whether companies can use Zara on their own applicants: not stated on the pages read. > > Try three real Tracy interviews free. The micro1 Zara interview is a conversational interview with an AI recruiter that micro1 gives every expert who applies to work on its projects. Candidates take it on demand, from anywhere, and get a result straight away. It is written for those candidates. If you run expert supply for an AI lab or a talent platform, the useful question is different: what would that format look like on your own applicants? This page explains the interview first, then how [Tracy AI Interviewer](/product) runs an interview of its own for you. ## What is the micro1 Zara interview? micro1's interview guide says: "At micro1, every candidate goes through an AI-powered interview designed to assess real-world skills in a fair, fast, and structured way." The interviewer is its AI recruiter, Zara. The guide describes it as "A conversational interview with our AI recruiter, Zara" that is "Available on demand" and can be completed "anytime, from anywhere." One expert's account on micro1's reviews page places it first in the funnel: an AI-based interview, then a chat with HR, then a discussion with a specialist. micro1's pages read do not state that sequence as a rule, so treat it as one candidate's experience. ## What does Zara ask? The guide lists four kinds of task: open-ended technical questions based on the chosen role, scenario-based questions "to evaluate how you think, problem-solve, and communicate", a coding challenge for technical roles, and a human data exercise for annotator roles. micro1's questions page adds that "No two candidates receive the same questions" and that Zara "adapts to your background." Its examples run from full-stack architecture trade-offs to physics derivations and enterprise sales scenarios. The design matches what a good first-round interview does: ask in the person's field, and shape the questions to the person. ## What does the candidate get at the end? According to the guide, the candidate gets immediate access to the interview, "A real-time result indicating whether you meet micro1's global certification criteria" and the option to request feedback if certification is not achieved. The criteria are micro1's own, so the result tells a candidate where they stand with micro1, not how they would do on another lab's work. ## What do candidates say about it? Reviews of micro1 on Trustpilot (4.7 from 11,219, read 10 Nov) are broadly positive; one reviewer calls the AI-led screening "clear, structured, and straightforward to navigate" (Trustpilot, Nov 2026). ## Who is the micro1 Zara interview for? Candidates are the reader of micro1's guide. It tells them to choose one core skill or job type, try its Interview Prep tool, optionally write their own questions and watch a short demo video. For preparation, micro1's own guide is the place to look. For a company, the position is plain. micro1 sells expert human data and evaluations, and its home page sends companies to a contact form. The pages read do not say whether the Zara interview can be used on your applicants, so asking micro1 is the only way to find out. ## How do you interview your own applicants in their field? If your experts come from your own funnel, referrals or an existing pool, you need an interviewer you control. Tracy generates questions from a job description or uses your own, interviews each applicant in their own field, follows up on what they said and waits for the answer. Applicants open a browser link, with no account, at any hour. You set the skills and their weights per role, and a person keeps the last word. a Director at a top-tier early-stage VC says of Tracy: "Tracy sometimes brings up exactly those questions I might have forgotten to ask, which has been a pleasant surprise." The table sets the two formats side by side. ![Tracy asks a fictional applicant how they found a pipeline fault, then what they changed so a retry stays safe.](/blog/images/micro1-zara-interview-1.png) *Tracy follows up on a fictional expert's own words, so the reviewer sees how they reason. Illustration with sample data* Tracy builds the next question from what the applicant just said. micro1 column: its own pages, read 9 Nov 2026. Tracy column: Tracy's published capabilities. | | Zara at micro1, as it describes itself | Tracy on your applicants | | --- | --- | --- | | Who is interviewed | Experts applying to micro1 | Your own applicants, referrals or contractors | | Questions | Role-based technical and scenario questions; a coding challenge for technical roles | From your job description or your own; coding task in the same interview | | Result | A real-time result against micro1's certification criteria | On the skills you choose, with a recording, timestamped transcript and assessment per question | | Who decides | Not stated on the pages read | A person; the history stays on the record | | Cost to the company | Not stated on the pages read | $1 per credit, first 3 results free, no seats | ![Four steps: paste the role, send a link, Tracy interviews, you review the recording, transcript and assessment.](/blog/images/micro1-zara-interview-2.png) *Tracy turns a job description into an interview and leaves the reviewer a recording, a transcript and an assessment per question. Illustration. Illustration with sample data* Four steps take you from a job description to evidence a reviewer can check. ## Which should you use? micro1 is the right place to ask if you want its Zara interview applied to experts it supplies, or if you want expert data and evaluations from it. Tracy is the right choice if you already have the applicants and need an interviewer you control. Create one interview from a real role and read the result: [try Tracy free](/connect). For an example of what a result holds, see [what an interview report contains](/blog/ai-interview-reports), and for the wider picture, [AI interview tools for expert vetting](/compare/ai-interview-tools-for-expert-vetting). The [pricing page](/pricing) has the rule: $1 per credit, no seats. ## Common questions ### What questions does the micro1 Zara interview ask? Open-ended technical questions for the chosen role, scenario-based questions, a coding challenge for technical roles and a human data exercise for annotator roles. micro1 says no two candidates get the same questions. ### How do you prepare for the micro1 Zara interview? micro1's guide suggests choosing one core skill or job type, using its Interview Prep tool, optionally writing custom questions and watching the demo video. ### Is Zara a real person? No. micro1 calls Zara its AI recruiter, and describes the interview as a conversation with her. ### Can a company use the Zara interview on its own applicants? Not stated on the pages read. micro1's site is written for expert applicants, and companies are sent to a contact form. Tracy is built for that job. ### What does Tracy give a reviewer after an interview? A recording, a timestamped transcript and a written assessment per question. A person decides. --- See the method applied in one real interview: send us a role and we build the interview plan in days, with a recording and report you can check. The first 3 results are free. [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=micro1-zara-interview&utm_content=demo) · [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=micro1-zara-interview&utm_content=try-free) ## Sources - [micro1: AI interview guide](https://www.micro1.ai/ai-interview-guide), retrieved 9 Nov 2026. - [Tracy AI Interviewer: Pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. # Proxy Interviews: How to Spot and Prevent Them > Proxy interviews put someone else in the interview or feed them answers. How the types differ, what the FBI says, how to spot one and how to prevent it. - URL: https://tracyinterview.ai/blog/proxy-interviews - Published: 2026-11-11 - Author: Tracy team > **In short** **Tracy AI Interviewer** runs and records the first round for every applicant, asks follow-ups on the person's own answers and shows signals with their times, so a reviewer has evidence to compare when a later round does not match. > > Proxy interviews are interviews where someone other than the applicant answers, or feeds answers from off camera. > > Identity is checked as a separate step, and a person decides from the evidence. > > Next step: send us one role and see the evidence on one interview. ## What are proxy interviews? A proxy interview is an interview in which someone other than the applicant takes part. Either they answer in the applicant's place, or they supply answers from off camera while the applicant appears on screen. Proxy interviews are described in several ways, so it helps to separate them before deciding what to check. - **Stand-in.** A different person takes the interview, and the applicant, or someone else, starts the job. - **Live helper.** The applicant is on camera and another person feeds the answers. The search term "proxy interview support" is used for this. - **Relay across rounds.** One person does the screening interview and another does the technical round. - **Impersonation.** The applicant uses someone else's identity, name and personal data. Candidate impersonation is the wider term. The four overlap, and one case can combine them. ## What the FBI says about someone else attending an interview The FBI's Internet Crime Complaint Center has described the pattern in a specific scheme. Its 16 May 2024 alert says that US-based facilitators of North Korean IT workers provided services that included attending "virtual interviews and meetings on behalf of North Korean IT workers", alongside receiving company laptops, setting up remote desktop connections and creating accounts on job sites. The same alert says companies that outsource IT work to third-party vendors "can face additional vulnerabilities since these companies are removed from the direct hiring process". The 2022 alert describes stolen personal data used to apply for remote jobs, with background checks finding data that belonged to another person. These alerts cover one set of cases. They give no rate, so this page does not either. They do show that someone else attending an interview is a documented tactic and not a thought experiment. ## How to spot a proxy interview No sign settles the question, so the table pairs each one with the kind of proxy it points to and where it comes from. | Sign | Which kind it points to | Where it comes from | |---|---|---| | Cannot discuss their own CV one level deeper | Stand-in or live helper | Our practice. The FBI lists trouble with key details of the person's past | | Extra noise, or sounds as if others are doing similar work nearby | Live helper or a group working together | FBI IC3, 16 May 2024 | | A pause before most answers, gaze that looks abnormal | Live helper feeding answers | Tracy signals: delay before answers, gaze | | The person in round two sounds or knows differently from round one | Relay across rounds | Our practice. A recording of round one makes the comparison possible | | The same phone number or email on applications under different names | Facilitated or organised applications | FBI IC3, 23 January 2025 | | A new address or payment platform appears during onboarding | Facilitated employment | FBI IC3, 23 January 2025 | The table lists six signs and what each one points to. The FBI's May 2024 tips also include noting "inconsistencies in interviews, especially applicants being unable to field questions about where they are located or key details about their past". Treat each item as a reason to look closer and not a verdict. Slow connections, shared flats and nervous answers produce several of these signs in honest interviews. Our guide to [AI interview cheating detection](/blog/ai-interview-cheating-detection) covers the live-helper case with AI in more depth. ## How to prevent proxy interviews Prevention is mostly process, and it costs little. Steps 4 and 5 are the FBI's advice. 1. **State the rule in the invitation.** Say who may attend and whether anyone may help, and say that the interview is recorded. A rule nobody was told is not one a person broke on purpose. 2. **Keep a recording of round one** and let the next interviewer watch a few minutes of it, so a different voice or a different depth of knowledge is noticed. 3. **Use questions that need the person's own experience.** A follow-up on the answer just given is harder for a helper to cover than the opening question. 4. **Verify identity as its own step.** The FBI's January 2025 advice is to "implement identity-verification processes during interviewing, onboarding, and throughout the employment of any remote worker". 5. **Do more in person where you can.** The same alert advises completing "as much of the hiring and onboarding process as possible in person". 6. **Apply every step to everyone.** A check that only some candidates get is a check on those candidates. ![Four steps: round one interviewed by Tracy and recorded, follow-ups, signals with times, round two compared by a person.](/blog/images/proxy-interviews-1.png) *Tracy runs and records round one, shows signals with their times, and a person compares it with the next round. Illustration with sample data* Tracy interviews and records round one, and a person compares it with later rounds. ## How Tracy gives a reviewer something to compare Tracy AI Interviewer runs the first round for every applicant, which matters most in [hiring at volume](/use-cases/volume-hiring). She asks in the person's own field, follows up on what they said and waits for the answer. A recording of both sides and a timestamped transcript sit behind each result, with a written assessment per question. When round two is run by a person, they have the first interview to compare against. Identity is checked in your own process, and a signal is a reason to look, not a finding that a proxy took part. The person interviewed is told on the [join screen](/participant-experience) that it is recorded and transcribed, and the decision stays with your team. ![Sample chain: Tracy asks about a data pipeline rebuild on the CV, the candidate answers, Tracy asks which retries were capped.](/blog/images/proxy-interviews-2.png) *Tracy follows up on one CV claim, so a reviewer can see whether the person can explain their own work. Illustration with sample data* Tracy follows up on the person's own work, which a stand-in or helper has to answer on the spot. ## What to do when you suspect a proxy Read the transcript around the signal, compare how the person explained their work in each round, and run your identity check. Then decide, with the evidence in one place so a second person can review it. ## Common questions ### What is a proxy interview? An interview where someone other than the applicant takes part, either answering in the applicant's place or feeding answers from off camera. ### What is proxy interview support? The phrase is used for both a person who sits the interview for someone else and a helper who supplies answers live. For an employer, either one means the interview does not assess the person you are hiring. ### How can you check for a proxy interview? Use several checks together: follow-ups on the person's own CV, a recording of round one to compare with later rounds, signals read by a person, and a separate identity check. No single check proves it. ### What is candidate impersonation? Using another person's identity to apply or interview. The FBI's 2022 alert describes background checks that found applicant data belonging to another person. ### Can an AI interviewer stop proxy interviews? No tool settles identity on its own. Tracy gives a reviewer follow-ups, signals with times and a recording to compare, and a person decides. --- See the method applied in one real interview: send us one role and we will build the interview plan in days, with a recording and report you can check. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=proxy-interviews&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=proxy-interviews&utm_content=demo) ## Sources - [FBI IC3 public service announcement I-051624-PSA, 16 May 2024: North Korea leverages US-based individuals to defraud US businesses](https://www.ic3.gov/PSA/2024/PSA240516), retrieved 9 Nov 2026. - [FBI IC3 public service announcement I-012325-PSA, 23 January 2025: North Korean IT workers conducting data extortion](https://www.ic3.gov/PSA/2025/PSA250123), retrieved 9 Nov 2026. - [FBI IC3 public service announcement I-062822-PSA, 28 June 2022: deepfakes and stolen PII used to apply for remote work](https://www.ic3.gov/PSA/2022/PSA220628), retrieved 9 Nov 2026. # SQL Interview Questions for Hiring Teams > SQL interview questions for hiring teams: 15 questions with follow-ups, strong and weak answers, and how to run them at volume. - URL: https://tracyinterview.ai/blog/sql-interview-questions - Published: 2026-11-11 - Author: Tracy team > **In short** **Tracy AI Interviewer** can run a SQL task inside the same interview and follow up on approach and trade-offs, so you hear how an analyst or engineer reasons, not only whether the query runs. > > Use the 15 SQL interview questions below, each with a follow-up and what to listen for. > > Every result keeps a recording, a timestamped transcript and a written assessment per question; your data lead decides. > > Next step: paste three of these questions into a Tracy interview. A data lead posts an analyst role and gets 150 applications. Every CV says SQL. The CV cannot tell you who writes a join that double counts revenue and who notices. SQL interview questions have to test that, so ask why a total is too high. These 15 SQL interview questions are written for data leads and recruiters hiring analysts and data engineers. Each has a follow-up and a note on what a strong answer contains. They suit a first conversation of 20 to 30 minutes; a take-home or pairing round belongs later with your own team. Tracy AI Interviewer runs that first conversation at volume. She asks each candidate questions in their own field, follows up on what they said and waits for the end of the thought. For the wider process, see [technical interviews](/use-cases/technical-interviews) and [AI technical interviews](/blog/ai-technical-interviews). ## Fifteen SQL interview questions, grouped by what they test Pick one or two from each group. Ask the follow-up only after the first answer. Choose your opening questions for the role and write down what a strong answer contains before the first candidate joins. ### Group 1: reading the problem *Table 1: Reading the problem, with follow-ups and what a strong answer contains.* | Ask | Follow up with | Listen for | | --- | --- | --- | | 1. You get a table of orders. What do you ask before writing a query? | What would change your query most? | What one row means, the grain, the time zone, whether deleted or test rows exist. | | 2. A colleague asks for "active users last month". How do you respond? | Who decides the definition? | Pins down active, month and users. Writes the definition down and gets it agreed. | | 3. When do you use a subquery, a CTE or a temporary table? | When would you avoid a CTE? | Readability first, then reuse and cost. Knows it depends on the engine. | ### Group 2: joins and duplicates *Table 2: Joins and duplicates, with follow-ups and what a strong answer contains.* | Ask | Follow up with | Listen for | | --- | --- | --- | | 4. Orders join to payments, and revenue looks too high. Where do you look first? | How would you change the query so each order counts once? | The join. Counts per key, aggregates before joining, or a distinct on the right grain. | | 5. When does a LEFT JOIN turn into an INNER JOIN by accident? | How would you spot it in a result? | A filter on the right table in the WHERE clause. Compares row counts before and after. | | 6. How do you find rows in one table with no match in another? | Which approach would you pick on a large table, and why? | Anti-join, NOT EXISTS, NULL handling. Knows NOT IN with NULLs misleads. | ### Group 3: aggregation and windows *Table 3: Aggregation and windows, with follow-ups and what a strong answer contains.* | Ask | Follow up with | Listen for | | --- | --- | --- | | 7. How do you get the latest order for each customer? | What if two orders share a timestamp? | A window function with a tie-break. Says what to do on ties, not only how. | | 8. What is the difference between WHERE and HAVING? | Give me a query where using the wrong one gives a wrong answer. | Row filter versus group filter, with a real example. | | 9. Calculate a seven-day moving average. What can go wrong? | What if some days have no rows? | Gaps in dates, partial windows at the start, a calendar table. | ### Group 4: performance and scale *Table 4: Performance and scale, with follow-ups and what a strong answer contains.* | Ask | Follow up with | Listen for | | --- | --- | --- | | 10. A report query slows down as the table grows. What do you check? | What if the plan looks fine? | Reads the plan, indexes, scans, data skew, then load. Measures before changing. | | 11. When does an index not help? | Which columns would you index for a query you choose? | Low selectivity, functions on columns, write cost. Picks columns with reasons. | | 12. How would you move a heavy daily report off the live database? | How do you know the copy is correct? | Replica, summary table, schedule. A reconciliation check. | ### Group 5: trust in the numbers *Table 5: Trust in the numbers, with follow-ups and what a strong answer contains.* | Ask | Follow up with | Listen for | | --- | --- | --- | | 13. Tell me about a number you reported that was wrong. | What do you check now? | A real case, how they found it, and a routine they added. | | 14. How do you test a query before someone relies on it? | What would you compare against? | Row counts, a known total, edge dates, a second method. | | 15. How do you explain a result to someone who does not know SQL? | Show me with the last query you wrote. | Plain words, a caveat stated, no jargon. | ## A strong answer names what it would check, and a weak one recites syntax The gap opens in the second answer. We wrote these pairs to show how it sounds; none is a quote from a real candidate. *Table 6: Weak and strong answers to the same questions, written as examples.* | | Weak answer | Strong answer | | --- | --- | --- | | Question 4, the inflated revenue | "I'd use DISTINCT." | "If an order has two payments, its amount counts twice. I'd sum payments per order first, then join, and check the total against the orders table." | | Question 7, the latest order | "ORDER BY date DESC, LIMIT 1." | "That gives one row total. I'd rank within each customer and add an order id as a tie-break." | | Question 13, the wrong number | "I've never reported a wrong figure." | "I joined on the wrong key once and overcounted by 8 percent. I now compare every total with a source count before I send it." | Listen for ownership. A candidate who describes a mistake and the check they added has done the work. ## One follow-up chain shows where the real answer starts This is an illustration with sample data, not a real interview. The candidate is fictional. It shows the shape of what Tracy runs: an opening question from your list, the answer, and a follow-up built on it. It starts from question 4. > **Tracy:** Orders join to payments, and revenue looks too high. Where do you look first? > > **Candidate (fictional):** At the join. If an order has two payments, its amount counts twice. > > **Tracy:** How would you change the query so each order counts once? > > **Candidate:** I'd sum the payments per order in a subquery, then join that to orders. > > **Tracy:** How would you check the new total is right? > > **Candidate:** Compare it with the total in the orders table and spot-check two orders by hand. The first answer is the textbook one. The third tells you whether the candidate verifies their own work. ![Transcript card: revenue looks too high, candidate says the join counts twice, Tracy asks how to count each order once.](/blog/images/sql-interview-questions-1.png) *Tracy asks a follow-up built on the candidate's own answer. Your data lead hears whether the candidate checks their own work. Illustration with sample data* A follow-up on the candidate's own answer shows whether they verify the numbers they report. ## Run the SQL interview for every applicant A live call for every applicant does not survive 150 applications. - **A SQL task inside the interview.** A SQL task can sit in the same interview as the spoken questions, with follow-ups on approach and trade-offs. Your team reads the answers and the transcript. - **Your questions or ours.** Describe the role and Tracy drafts questions from the job description, which you edit; see the [product page](/product). - **Start on click.** Each candidate opens a link in a browser, with no account and no calendar, at any hour, and is told on the join screen that the conversation is recorded and transcribed. See the [participant experience](/participant-experience). - **A result you can check.** A recording and a timestamped transcript sit behind every result, with a written assessment per question. See [AI interview reports](/blog/ai-interview-reports). - **Integrity signals, with the evidence.** Tracy captures tab changes, gaze, face detection, delay before answers, fullscreen exits and external screens and shows the evidence to a person. A signal is not proof; see [AI interview cheating detection](/blog/ai-interview-cheating-detection). It is $1 per credit, one credit per started 20 minutes of an interview that produced a result. The first 3 results are free with no card, and there are no seats. See [pricing](/pricing). ![Tracy result report for a fictional data analyst: a recording, a transcript with the 7:38 answer about duplicate payments, and three questions with what they said and quoted evidence.](/blog/images/sql-interview-questions-2.png) *A recording and timestamped transcript sit behind every Tracy result. The reviewer reads the candidate's own words before moving them forward. Illustration with sample data* Your data lead can read what the candidate said to each question and open the recording and transcript behind it before deciding. ## Keep three decisions with your team 1. **Who moves forward.** Your data lead reads the report and decides. A person has the last word. 2. **The calibration.** Have one analyst read five transcripts blind. If they disagree with the written assessments, change the questions before inviting the next hundred. 3. **The hands-on round.** A real dataset, a take-home or pairing stays with your team. For a handful of hires, a call with a senior analyst is simpler; see [AI interview vs human interview](/blog/ai-interview-vs-human-interview). The case for Tracy starts when applicants outnumber your team's hours, as in [volume hiring](/use-cases/volume-hiring). Start small: take one analyst opening, use three of these questions and have your data lead read the first batch of reports before you widen it. ## Common questions ### What are good SQL interview questions? Ask about reading the problem, joins and duplicates, aggregation and windows, performance, and how candidates check their numbers. Keep the follow-up beside each question, because that is where the real answer starts. ### What SQL interview questions suit a data analyst? Lean on joins, windows and trust in the numbers. For a data engineer, add the performance and scale group and ask about moving heavy work off the live database. ### What SQL interview questions suit a data engineer? Use the performance group and the reconciliation follow-ups: plans, indexes, replicas and how the candidate knows a copy is correct. ### Can Tracy run a SQL task in the interview? Yes. A SQL task can sit in the same interview, with follow-ups on approach and trade-offs. The report gives a recording, a timestamped transcript and a written assessment per question. ### Does Tracy decide who to hire? No. A person reads the evidence and decides. --- Create an interview for this exact role from these questions, then read the recording and report; the button opens the Connect page. Or send us one role and we will build the interview plan, with a recording and report you can check. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=sql-interview-questions&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=sql-interview-questions&utm_content=demo) ## Sources - [Tracy AI Interviewer home page](https://tracyinterview.ai/), retrieved 9 Nov 2026. - [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer Connect page](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. # Structured Interview Scoring: A Guide With Examples > Structured interview scoring made practical: write anchors, score each answer alone, record evidence and compare, with a worked example. - URL: https://tracyinterview.ai/blog/structured-interview-scoring - Published: 2026-11-11 - Author: Tracy team > **In short** **Tracy AI Interviewer** interviews every candidate on the skills you set for the role. > > Each result has a recording, a timestamped transcript and a written assessment per question, so a reviewer can check each point. > > Below: how structured interview scoring works, from anchors to disagreement and common errors. > > Next: load your criteria into one role and read three reports. Two interviewers meet the same candidate. One writes "strong, 4". The other writes "fine, 3". In the debrief they argue about the candidate when they should be arguing about the scale, because neither wrote down what a 4 means. Structured interview scoring fixes that by deciding, before anyone is interviewed, what each score looks like and what evidence earns it. This guide covers the scale, the anchors, when to score, how to combine scores and where scoring goes wrong. Tracy AI Interviewer interviews each candidate on your skills and writes an assessment for each question, and the method below works on any sheet. Our guide to [screening candidates](/blog/how-to-screen-candidates) covers the step before it. ## Why structured interview scoring beats a gut call The reason to bother is predictive. A 2022 re-analysis in the Journal of Applied Psychology by Sackett and colleagues, covering studies of selection procedures, estimated a mean validity of .42 for structured interviews and .19 for unstructured ones. That is a correlation with job performance ratings, about twice as strong a link, not a hiring success rate. The structured figure rests on 105 studies and 7,864 people. Structure helps because it narrows how much one person's mood decides the outcome. It does not make the criteria correct. Scoring only works if the criteria came from the job. ## Write the anchors before the first interview Use a 1 to 5 scale with three anchors written out: what a 1, a 3 and a 5 sound like for each criterion. Interviewers fill in the 2 and the 4. Five points give enough room to separate strong from good without pretending to be exact. Some teams use four points to remove the safe middle. Pick one and keep it. Write anchors as behaviours in an answer, not adjectives. "Confident" is an adjective. "Names the steps in order and says what each ruled out" is a behaviour. | Score | Troubleshooting (support engineer) | | --- | --- | | 1 | Cannot describe a fault they worked on, or describes only the outcome | | 3 | Describes steps but not why that order, or what they ruled out | | 5 | Names the steps in order, says what each ruled out and what they would check next | The table shows three written anchors for one criterion. The anchors are an example. Write yours with the people who do the job, and test them on two past hires before you use them. ## Score each answer, alone, with a line of evidence Score each answer straight after you hear it, not at the end. By the end, the last answer and the first impression have taken over. Score alone before the debrief, so the loudest voice cannot set everyone's number. Then compare. Write one line of evidence for every score: a quote, a number or a moment. "Strong" is not evidence. "Ruled out DNS first because the error was intermittent" is. A score with no evidence line is an opinion. A fictional worked example, one answer, scored against the anchors above: > **Question:** Tell me about the last hard fault you diagnosed. > > **Candidate (fictional):** A customer's sync failed twice a week. I checked logs first, then ruled out the network because it only failed at night, then found a scheduled job holding a lock. > > **Score: 4.** Steps in order, one thing ruled out with a reason. Did not say what they would check next, which a 5 needs. ![Sample scoresheet for a fictional support engineer: four weighted criteria scored 1 to 5, total 3.8.](/blog/images/structured-interview-scoring-1.png) *A sample scoresheet, not a Tracy report screen. Illustration with sample data* Each criterion gets a score against written anchors, and the weights produce one comparable number. ## Combine scores with weights and floors, not averages Weights set what matters. If troubleshooting counts for 35 percent and written clarity for 20, a candidate strong at the first and weak at the second can still lead. Set weights before the interviews and do not change them after you have met the candidates. Do not average away a must-have. If a skill is a hard requirement, set a floor, for example no score under 3 on troubleshooting, and apply it before the weighted total. When two interviewers differ by two points or more on a criterion, do not split the difference. Read the evidence lines and ask which one the anchor supports. If the anchor cannot settle it, the anchor was vague. Rewrite it. ## Know the five errors that bend a score These errors recur, and each has a cheap fix. - **Halo.** One strong answer lifts every score. Score each criterion on its own evidence. - **Leniency.** Everyone gets 4. Compare your average with a colleague's on the same candidates. - **Recency.** The last answer counts most. Score as you go. - **Similarity.** You rate people like you higher. Keep the evidence line and read it again. - **Drift.** By candidate 40 your 3 is a different 3. Re-read the anchors every ten interviews. The last one is why scoring a few interviews by hand is easy and scoring hundreds is not. ## Interview on your skills with Tracy You pick the skills for a role. Tracy interviews each candidate on them and follows up on what each person says, and every result carries a written assessment per question with its reasoning, a recording and a timestamped transcript. A reviewer reads the evidence for each person, and results can be shared with the hiring team. A person keeps the last word. They can move a candidate forward or invite them to interview again, and the history stays on the record. Tracy does not publish a bias audit, so use the evidence to check a reviewer's score, not to excuse it. See [AI interview reports](/blog/ai-interview-reports) for what the reviewer reads. ![Four steps: set the anchors, ask questions, review the evidence, a person scores.](/blog/images/structured-interview-scoring-2.png) *Tracy asks about your skills and leaves evidence a person can check. Illustration with sample data* The routine is the same at any volume: anchors first, evidence per score, a person decides. ## Pilot your scale on one role Pilot your scale on one role. 1. Write four criteria, weights and three anchors each. 2. Set the skills in Tracy and interview everyone who applies for a fortnight. 3. Have the hiring manager read ten reports and score each by hand against the same anchors. 4. Stop if the manager's scores and the assessments disagree on more than a few candidates and the anchors cannot explain why. 5. Measure the share of scores with an evidence line and the hours the manager spent. Tracy costs $1 per credit, one credit per started 20 minutes, and the first 3 results are free with no seats. Try one on the [Connect page](/connect), and see [pricing](/pricing). Which path fits you: - Many applicants for the same role: [volume hiring](/use-cases/volume-hiring). - Technical roles: [technical interviews](/use-cases/technical-interviews). Good looks like this: two reviewers can read the same evidence and land within a point of each other. ## Common questions ### What is structured interview scoring? Scoring every answer against anchors written in advance, on one scale, with a line of evidence for each score and weights that reflect the job. ### What is the best structured interview rating scale? A 1 to 5 scale with the 1, 3 and 5 described as behaviours is a common choice. Some teams prefer four points to remove the middle. What matters is that anchors are written and used by everyone. ### How do you evaluate a structured interview? Score each answer straight after it, alone, with evidence. Combine the scores with weights, apply any floors, then compare with other interviewers' scores and discuss gaps of two points or more. ### Can an AI interviewer run a structured interview? Yes. You pick the skills for the role, and Tracy AI Interviewer interviews each candidate on them and leaves a recording, a transcript and an assessment per question. A person reviews the evidence and decides. --- See the method applied in one real interview. Send us one role and we will build the interview plan from your skills in days, with a recording and report you can check. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=structured-interview-scoring&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=structured-interview-scoring&utm_content=demo) ## Sources - [Sackett, Zhang, Berry and Lievens (2022), Revisiting meta-analytic estimates of validity in personnel selection, Journal of Applied Psychology 107(11)](https://doi.org/10.1037/apl0000994), retrieved 9 Nov 2026. - [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 5 Nov 2026. # Time to Hire Benchmark: What the Numbers Say > A time to hire benchmark from SHRM's 2025 survey of 2,371 members: 44 days median, how it is defined, and what a benchmark must state to be usable. - URL: https://tracyinterview.ai/blog/time-to-hire-benchmark - Published: 2026-11-11 - Author: Tracy team > **In short** **Tracy AI Interviewer** takes out the scheduling days: each applicant starts an interview on click, at any hour, so screening runs while you sleep. > > One customer cut its average time to hire from 44 days to 7, and Tracy did the work of 53 full-time recruiters in one month. A customer's own figure uses its own definition and is not a benchmark. > > The time to hire benchmark to cite: SHRM's 2025 median of 44 days to fill a nonexecutive role, from 2,371 US members. > > Next: measure your own stages, then run one role in Tracy. Picture a board asking whether 38 days to hire is good. Someone pastes a number from a vendor blog, and nobody asks what it measured. Is the clock running from the day the requisition opens or the day the candidate applies? Is it a median or a mean? Whose roles? A time to hire benchmark is only useful if it answers those questions. The one we can cite in full comes from SHRM's 2025 benchmarking survey, and it is below. Tracy AI Interviewer works on the part of the clock a team controls most directly, the days between a person applying and a first interview happening. ## The time to hire benchmark we can cite: SHRM, 2025 SHRM, the Society for Human Resource Management, fielded the survey to a random sample of its active US members from 9 January to 3 March 2025, and 2,371 members responded (members at VP level or above were deliberately oversampled). Respondents span industries, sectors and organisation sizes. The data is not weighted, not every respondent answered every question, and SHRM reports medians. SHRM calls the metric time to fill: the days from the requisition being opened to the offer being accepted, counted in calendar days including weekends and holidays. | Median days to fill | 2017 | 2022 | 2025 | | --- | --- | --- | --- | | Nonexecutive roles | 30 | 44 | 44 | | Executive roles | 47 | 60 | 45 | Median days to fill, by role type, from SHRM's 2017, 2022 and 2025 data. Source: SHRM, 2025 Recruiting Executives Benchmarking data brief, US, medians. Two more points from the same brief. Extra-large organisations, those with 5,000 or more employees, reported a median of 61 days for nonexecutive roles and 60 for executive roles, and spent nearly twice as long on approval, posting and interviewing. Executive roles got faster between 2022 and 2025, but executive cost per hire rose 21 percent over the same period and is up 113 percent from 2017. SHRM reads that as organisations possibly paying more to go faster. ![Bar chart of SHRM median days to fill: nonexecutive 30, 44 and 44; executive 47, 60 and 45, for 2017, 2022, 2025.](/blog/images/time-to-hire-benchmark-1.png) *Median days to fill from requisition opened to offer accepted, as SHRM reported them in 2017, 2022 and 2025. Source: SHRM, 2025 Recruiting Executives Benchmarking data brief.* Nonexecutive time to fill has held at a median of 44 days since 2022; executive time fell from 60 to 45. ## Time to hire, time to fill and time to offer are different clocks The three words get used as if they were one. They are not. - **Time to fill** (SHRM's measure): requisition opened to offer accepted. It includes the days a role sits waiting for approval, before a single applicant exists. - **Time to hire:** most often counted from when a candidate enters the process, whether by applying or being found, to offer accepted. It leaves out the approval days. - **Time to offer:** the same start, stopped when the offer goes out. It leaves out the candidate's decision. SHRM's own brief splits time to fill into seven stages, from requisition approval through screening, interviews, the final decision and offer acceptance. That is why a single number hides so much. Two teams with the same 44 days can have different problems. We would not set a target on the total before looking at the stages. ## What a time to hire benchmark must state Before you put a figure in a board paper, check that it states each of these. If a number misses one, treat it as colour, not a benchmark. 1. **The start and stop events.** Requisition opened, application received, first contact. Offer sent, offer accepted. 2. **Calendar or business days.** SHRM counts calendar days. 3. **Median or mean.** A few slow executive searches drag a mean up. SHRM reports medians. 4. **Who answered.** SHRM's sample is US members of a professional body, answering for themselves. 5. **When.** SHRM fielded in early 2025. Labour markets move, and a figure from 2021 describes another market. 6. **How many.** 2,371 respondents overall, with fewer per metric. 7. **Which roles.** Executive and nonexecutive differ, and so do organisation sizes. If a vendor has data from its own customers, it is a fair guide to those customers and no more. ## Measure your own stages before you chase a benchmark Pick one role family and pull the last 20 hires. For each, write down five dates: requisition opened, job posted, first applicant screened, last interview done and offer accepted. Take the median of each gap, not the average. That gives you your own stage table to set against SHRM's definitions. One measure to set aside is the total on its own. A fast hire who leaves in four months is not a good result. SHRM's brief notes that few organisations measure quality of hire in 2025, so pair the days with something you can check, such as whether the hire passed probation or the hiring manager's score a quarter later. ## Where Tracy takes scheduling out of the clock In SHRM's stages, the days a recruiter can change fastest are screening and interviews. Both wait on calendars: a person has to read the CV, book a slot and be free to talk. Tracy removes the booking. Each applicant starts on click, at any hour, with no scheduling and no recruiter present, and is ready in under 10 seconds. A reminder goes out after a day, and failed interviews are chased automatically. Tracy follows up on what each applicant says, and a person reads the recording, transcript and assessment per question, then decides. ![Four steps: job posted, screening starts with a link, interview done, report ready for a person to decide.](/blog/images/time-to-hire-benchmark-2.png) *Tracy covers the screening and first-interview stages of the clock without scheduling. Illustration with sample data* Screening and first interviews stop waiting on calendars: every applicant gets a link, and a person reads the report. ## Pilot it on one role and measure the stage Pilot Tracy on one role and compare the same role's stages before and after. 1. Record the baseline: median days from first applicant screened to last interview done, for the last 20 hires. 2. Run the next requisition with Tracy as the first round. Load your questions or let her draft them from the job description. 3. Stop if the median does not fall, or if reviewers say the reports do not help them decide. 4. Read the results against quality, not only days. Tracy costs $1 per credit, one credit per started 20 minutes, and the first 3 results are free with no seats. Try a role on the [Connect page](/connect), and see [pricing](/pricing). Which path fits you: - Many applicants per role: [volume hiring](/use-cases/volume-hiring). - Staffing teams racing to submit: [AI interview tools for staffing agencies](/blog/ai-interview-tools-for-staffing-agencies). Good looks like this: you can state your own median per stage, how it is defined, and which stage moved. ## Common questions ### What is a good time to hire benchmark? SHRM's 2025 survey of 2,371 US members found a median of 44 days to fill nonexecutive roles and 45 days for executive roles, measured from requisition opened to offer accepted. Treat it as a reference for your own definition, not a target. ### What is the average time to hire? It depends on the definition. SHRM reports medians, not averages, and measures time to fill. A figure that does not say where the clock starts and stops cannot be compared with yours. ### What is the difference between time to hire and time to fill? Time to fill starts when the requisition opens and ends when the offer is accepted. Time to hire usually starts when the candidate enters the process, so it leaves out approval days. ### Is there a cost per hire benchmark? SHRM's 2025 brief reports a median cost per hire of $1,200 for nonexecutive roles and $10,625 for executive roles. Executive cost per hire is up 113 percent from 2017 and 21 percent from 2022. --- See the method applied in one real interview. Send us one role and we will build the interview plan in days, with a recording and report you can check. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=time-to-hire-benchmark&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=time-to-hire-benchmark&utm_content=demo) ## Sources - [SHRM (2025), 2025 Recruiting Executives Benchmarking: Insights to Maximize Recruitment, data brief](https://www.shrm.org/content/dam/en/shrm/research/2025-recruiting-benchmarking-report.pdf), retrieved 9 Nov 2026. - [Tracy AI Interviewer home page, customer quotes](https://tracyinterview.ai/), retrieved 9 Nov 2026. - [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 5 Nov 2026. # AI Interviews in Claude: Create and Review Interviews From a Chat > Run a Claude interview with Tracy AI Interviewer: add the connector, draft an interview in chat, approve it, invite people and read results. - URL: https://tracyinterview.ai/blog/tracy-claude-integration - Published: 2026-11-11 - Author: Tracy team > **In short** **Tracy AI Interviewer** lets you draft, invite and review interviews from a Claude chat. Nothing is created until you approve the draft. > > Tracy still runs the interview: video in the person's browser, with follow-ups in their own field, and a recording, transcript and assessment per question for a person to decide from. > > Setup is one custom connector at tracyinterview.ai/mcp. Claude asks before each use, and the connector acts only on your own account. > > Next step: add the connector. You already describe an interview in words before you build it: who you are talking to, what you need to learn, how long it should take. Then you copy those words into a form. A Claude interview setup skips the copying. With Tracy AI Interviewer added to Claude as a connector, you ask for the interview in plain language and Claude comes back with a draft to read. A vague prompt costs you a re-read, not a live link, because nothing exists until you approve the draft. The same chat can invite people and read results, so you do not switch tools to see who has finished. The interview itself stays Tracy's. She interviews each person on video, asks follow-ups in their own field and waits for the end of the thought. Claude is where you set the interview up and look at results. ## What the connection is MCP, the Model Context Protocol, is a standard way for an AI app to call tools in another service. Tracy AI Interviewer publishes one at tracyinterview.ai/mcp. Claude, ChatGPT, Mistral, Gemini and Copilot Studio can all add it. This page covers Claude. Our [ChatGPT guide](/blog/tracy-chatgpt-integration) covers the other large chat app, and the [Connect page](/connect) lists every route, including [Slack](/blog/tracy-slack-integration). ## What Claude can read and change in your Tracy account Once the connector is on, you can ask Claude to: - **Draft an interview** from a short brief. The draft shows the title, the questions and the result fields. You ask for changes in the same chat and say yes when it is right. - **Create it** once you approve, and give you the public link. - **Invite people** to an interview. - **Read results** by asking for them. - **Check your balance** of free results and credits. The connector asks for one permission, named tracy, and acts only on your own account. Claude also asks before it uses Tracy: you choose Allow once, or Always allow to stop being asked. In Claude's connector settings the tools are grouped as 7 read-only and 6 write tools, each group set to Needs approval, and you choose when Claude may use them. You can revoke access at any time under Connected AI apps in your Tracy account, and it stops at once. Tracy, not Claude, conducts the interview. The person being interviewed opens a link in a browser, with no account, and is told on the join screen that the conversation is recorded and transcribed. What they meet there is the same video interview the [web app](/product) runs. ![Claude connector page for Tracy AI Interviewer: tool permissions with 7 read-only and 6 write tools, each group set to Needs approval.](/blog/images/tracy-claude-integration-4.png) *Screenshot from the Tracy AI Interviewer Claude guide, 9 Nov 2026. The connector page lists 7 read-only tools and 6 write tools, each group set to Needs approval. Source: Tracy AI Interviewer Claude guide, 9 Nov 2026.* You choose when Claude may use each Tracy tool, and every group is set to Needs approval. ## Five jobs to hand to the chat 1. **Draft from one sentence.** The guide's own first prompt is: "Use Tracy AI Interviewer to set up a customer discovery interview for people who cancelled in the last month, then give me the public link." Claude replies with a draft, not a live interview. 2. **Rework the draft.** Ask for fewer questions, a different order or a deeper style before you approve. Tracy can also start from your own questions. 3. **Invite people.** Ask Claude to invite them, and read the step Claude shows before you allow it. Tracy's Connect page shows invitations from chat in ChatGPT as a preview before anything is sent. 4. **Read where results stand.** Ask who has finished. Each result opens to a recording and a timestamped transcript, with a written assessment for each question. 5. **Check credits before a batch.** Ask how many free results and credits you have left. Claude asks for approval first, as in the screenshot below. ![Claude prompt: Claude wants to use Get the balance from Tracy AI Interviewer. Buttons: Deny, Always allow, Allow once.](/blog/images/tracy-claude-integration-1.png) *Screenshot from the Tracy AI Interviewer Claude guide, 9 Nov 2026. Claude asks to use Get the balance from Tracy AI Interviewer, with Deny, Always allow and Allow once. Source: Tracy AI Interviewer Claude guide, 9 Nov 2026.* Claude asks for your approval before the connector reads your Tracy balance. ## Set up a Claude interview in five steps The steps are the same in Claude Desktop and in claude.ai. Custom connectors work on the Free, Pro, Max, Team and Enterprise plans. The Free plan allows one custom connector. 1. Open **Customize > Connectors** in Claude. 2. Click **+ Add**, then **Add custom connector**. 3. Name it `Tracy AI Interviewer`, paste the address `https://tracyinterview.ai/mcp`, then click **Continue** and **Add**. Leave the detected sign-in options as they are. 4. Click **Connect**, sign in to Tracy if asked, then click **Allow**. 5. Start a new chat. Select the **+** button at the lower left, then **Connectors**, and make sure Tracy AI Interviewer is switched on for that conversation. Send the first prompt. On Team and Enterprise, an owner adds the connector first under **Organization settings > Connectors**: Add, Custom, then Web, with the same name and address. Each member then opens Customize > Connectors and clicks Connect. Ready to try it? Connect Claude and draft one interview. To check it worked, open your Tracy account and find Connected AI apps. Claude is listed there with the time you connected it. Tracy answers the first prompt with a draft of the title, the questions and the result fields. Developers who work in Claude Code add it from a terminal with `claude mcp add --transport http tracy https://tracyinterview.ai/mcp`, then run `/mcp` and sign in. Run `claude mcp list` to confirm: tracy should read Connected. ![Claude Add custom connector dialog filled with the name Tracy AI Interviewer and the address https://tracyinterview.ai/mcp.](/blog/images/tracy-claude-integration-3.png) *Screenshot from the Tracy AI Interviewer Claude guide, 9 Nov 2026. The Add custom connector dialog with the name Tracy AI Interviewer and the address https://tracyinterview.ai/mcp. Source: Tracy AI Interviewer Claude guide, 9 Nov 2026.* The whole setup is a name and one address in Claude's custom connector dialog. ![Four steps: ask in Claude, approve the draft, Tracy interviews in a browser, review recording, transcript, assessments.](/blog/images/tracy-claude-integration-2.png) *The owner asks and approves in Claude while Tracy interviews and produces the report. Nothing goes live without your yes. Illustration, not a screenshot. Source: Tracy visual library, process template; workflow from the public Claude guide.* You approve the draft in Claude, and Tracy runs the interview and keeps the evidence. ## If it does not work - **Tracy is not in your connectors.** On Team and Enterprise an owner has to add it first. Ask them, then connect it yourself. - **The sign-in did not finish.** Open Customize > Connectors, find Tracy AI Interviewer and click Connect again. If Claude is missing from Connected AI apps in your account, the sign-in was not completed. - **The connector is off in the chat.** Use the + button, then Connectors, and switch it on for that conversation. - **Access was removed.** If you revoked Claude under Connected AI apps, it stopped at once. Connect it again with the steps above. ## What stays with people Claude drafts and reports, and you decide. Read each draft before you approve it, because the questions are what every person will be asked. Read the results yourself: a recording, a timestamped transcript and a written assessment per question sit behind each one, with integrity signals and their evidence for a human reviewer to interpret. A signal is a prompt to look, not proof of anything, and the decision stays with you. Our guide to [AI interview reports](/blog/ai-interview-reports) shows what a reviewer reads. Cost follows the same rule as the web app. It is $1 per credit, one credit per started 20 minutes of an interview that produced a result. The first 3 results are free with no card, there are no seats, and a session with no answers is never charged. The [pricing page](/pricing) has the current plan. ## Which path fits you - Research leads running discovery calls: start from [customer research](/use-cases/customer-research). - Expert supply and talent platforms vetting specialists: see [expert vetting](/use-cases/expert-vetting). - Teams with many applicants: see [hiring at volume](/use-cases/volume-hiring). - Slack or ChatGPT instead of Claude: the [Slack guide](/blog/tracy-slack-integration) and the [ChatGPT guide](/blog/tracy-chatgpt-integration) cover them. - Prefer a browser: the web app does the same job without a chat. What the interviewed person sees is on the [participant experience](/participant-experience) page. ## Common questions ### What is a good Claude interview prompt for Tracy? Start with the one in Tracy's guide: "Use Tracy AI Interviewer to set up a customer discovery interview for people who cancelled in the last month, then give me the public link." Swap in your own role or topic. Claude replies with a draft you read before anything is created. ### Can Claude run the interview itself? No. Claude sets it up, invites people and reads results. Tracy interviews each person on video in a browser, with follow-ups in their own field. ### Is this Claude interview prep for candidates? No. The connector is for the person running interviews. If you are preparing to be interviewed, ask Claude for practice questions directly, and see what to expect in a Tracy interview on the participant experience page. ### Which Claude plans work? Custom connectors work on Free, Pro, Max, Team and Enterprise. The Free plan allows one custom connector. On Team and Enterprise an owner adds it first. ### Does Claude get access to my other data? The connector asks for one permission, named tracy, and acts only on your own Tracy account. Claude asks before it uses it, and you can revoke access under Connected AI apps. ### Is there an extra charge for the connector? No separate charge. Pay as you go includes full API and MCP access. --- Add the connector in five steps, or send us one role and we will build the interview plan, with a recording and report you can check. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=tracy-claude-integration&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=tracy-claude-integration&utm_content=demo) ## Sources - [Tracy AI Interviewer Connect page](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. - [Add Tracy AI Interviewer to Claude (public guide)](https://tracyinterview.ai/help/claude), retrieved 9 Nov 2026. - [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 5 Nov 2026. # AI Interviews Over WhatsApp: Invite Candidates Where They Already Are > A WhatsApp interview with Tracy AI Interviewer: message Tracy, draft an interview, approve it and invite candidates by email, all from the chat. - URL: https://tracyinterview.ai/blog/tracy-whatsapp-integration - Published: 2026-11-11 - Author: Tracy team > **In short** **Tracy AI Interviewer** works inside WhatsApp for the account owner: message Tracy's number, then draft interviews, invite people by email and check results from the chat. > > Candidates do not interview in WhatsApp. They open the emailed link in a browser, with no account, for a video interview with follow-ups in their own field. > > Every create or invite shows a preview, and nothing is sent until you tap Approve. > > Next step: connect your phone and draft one interview. Your shortlist reaches you in a WhatsApp group at 8 a.m. The interview tool is in a browser tab on a laptop you will not open until noon. By then the people you wanted to move are answering other agencies. A WhatsApp interview setup puts the tool where you already are. With Tracy AI Interviewer in WhatsApp, you message her number and ask in plain words. She drafts the interview, shows you what she is about to do and waits for you to tap Approve. She can invite people by email, tell you who is on an interview and where their results stand, and tell you how many free results and credits you have left. The interview itself is Tracy's: video, with follow-ups in the person's own field and a recording and timestamped transcript behind every result. ## What you can ask Tracy to read and change You talk to Tracy in a WhatsApp chat with her number. The first time, she sends a link that connects your phone to your Tracy account. After that she knows who you are. **Read:** - Your interviews, and who is on each one. - Where their results stand. - How many free results and credits you have left. **Change, always with a preview and an Approve button:** - Draft and create an interview. - Invite people by email. The preview lists who will be emailed. - Turn an interview's open link on or off. The permission floor is your own account. Tracy reads the messages you send to her number and answers in the same chat. She cannot see your other chats or your contacts. She reads typed messages only, so send text and not a voice note or a picture. ## Set up a WhatsApp interview in five steps Before you start you need a Tracy AI Interviewer account, and WhatsApp on the phone whose number you want to use. You do not add Tracy as a contact first. A message to her number is all it takes. The number and a QR code are in the WhatsApp guide on the [Connect page](/connect). 1. Open a WhatsApp chat with Tracy and send any message, such as `hi`. 2. Tracy replies with a short hello, what she can do and a link to connect your account. Open the link yourself. It is good for 10 minutes and is for your number alone, so do not pass it on. If it has run out, send her another message for a new one. 3. If you are signed out, sign in to your Tracy account. The page shows your number with all but the last digits hidden, and the email of the account it will link to. 4. Check that both are yours, then click **Link my account**. Go back to WhatsApp, where Tracy confirms the connection and suggests what to ask. 5. Ask her something. Three to start with: "Which interviews do I have?", "Who is on my customer success interview, and where do their results stand?" and "How many free results and credits do I have left?" To check it worked, ask the first question. She answers from your account, or says you have no interviews yet. Your account page shows the number as connected in the WhatsApp bot panel. The Disconnect button there ends the link, and the next message from that number gets a new connect link. ![Four steps: message Tracy, link your phone within 10 minutes, ask in plain words, tap Approve before anything is sent.](/blog/images/tracy-whatsapp-integration-1.png) *Tracy takes your requests in WhatsApp and waits for Approve. An owner can run interviews from a phone without a stray send. Illustration of the steps in the public guide, not a screenshot. Source: Tracy visual library, process template; steps from the public WhatsApp guide.* You link your phone once, and after that every create or invite waits for your Approve. ## Four jobs for a phone and a minute 1. **Morning check.** Ask which interviews you have and who has finished. You see where results stand before you reach a desk. 2. **Draft and create.** Describe the role or the conversation. Tracy drafts the questions and shows a preview that ends with Approve and Cancel. After Approve she creates the interview and sends its link. Nobody is invited yet. You can also bring your own questions. 3. **Invite a shortlist.** Ask her to invite people by email. The preview lists who will be emailed, and nothing goes out until you tap Approve. Each person gets a link and a reminder goes out after a day. 4. **Close a link.** Turn an interview's open link off when you have enough people, and back on later if you need it. Each of these changes gets a preview too. A decision made on a phone can start a full interview with follow-ups. The [hiring at volume](/use-cases/volume-hiring) page covers the rest of that workflow. Ready to try it? Connect your phone and draft one interview. ## The candidate opens a browser, not a chat You work in WhatsApp. The person you invite gets an email with a link, opens it in a browser, with no account and no calendar, and starts when ready, at any hour. They are told on the join screen that the conversation is recorded and transcribed. A dropped person can rejoin within 10 minutes, which is configurable per customer. What they go through is on the [participant experience](/participant-experience) page. Steven Perry, Account Executive at IT Resources, an IT staffing firm in Tampa, US, says of the candidate side: "By the time they finish with this platform, they're even laughing and joking." You get the rest from the interview: a recording of both sides, a timestamped transcript, a written assessment per question with its reasoning, and integrity signals with their evidence for a person to review. A signal is a prompt to look, not proof, and your team decides. ![Myth: candidates take the interview inside WhatsApp. Fact: you work in WhatsApp and they open a browser link.](/blog/images/tracy-whatsapp-integration-2.png) *The owner drafts and invites from WhatsApp while the candidate interviews with Tracy in a browser. No confusion about where the interview happens. Illustration, not a screenshot. Source: Tracy visual library, myth-fact template.* The chat is for the person running the interviews, and the interview opens from a browser link. ## WhatsApp interview scams: what a real invitation looks like The search "whatsapp interview scams" is common, and recruiting scams do run on chat apps. The general advice applies to any invitation, and it is advice, not a Tracy rule: check who is asking, check that the employer exists, and stop if anyone asks for money or banking details before a job. Here is what a genuine Tracy invitation looks like on our side. The recruiter's Tracy account sends an email with a personal link. The interview opens in a browser, and the join screen says the conversation is recorded and transcribed. If a message only says "join this interview in WhatsApp", ask the sender for the email invitation. ## When it does not work - **No answer from Tracy.** Check that you wrote to her number, then send a typed message. A recording or a picture gets a reply that she reads text only. - **Her link did not work.** A link lasts 10 minutes. Send her any message for a new one. - **You linked the wrong account.** Click Disconnect in the WhatsApp bot panel of your account, message Tracy again and open the new link while signed in to the right account. Prefer a laptop chat? The [Slack guide](/blog/tracy-slack-integration) covers that route, and the [product page](/product) shows the web app. ## Which path fits you - Staffing firms and RPO teams placing many people: start with [AI interview tools for staffing agencies](/blog/ai-interview-tools-for-staffing-agencies) and [hiring at volume](/use-cases/volume-hiring). - Teams who live in Slack: the [Slack guide](/blog/tracy-slack-integration). - Cost: $1 per credit, one credit per started 20 minutes of an interview that produced a result. The first 3 results are free, with no card and no seats. See [pricing](/pricing). ## Common questions ### Are WhatsApp interviews scams? Not by nature, but fake recruiters do use chat apps, so check the sender and the employer and never pay to be interviewed. A Tracy invitation arrives by email with a personal link and opens in a browser. ### Where do candidates take the interview? In a browser. The recruiter works in WhatsApp, and the person invited opens the emailed link and needs no account. ### What can Tracy see in my WhatsApp? Only the messages you send to her number. She cannot see your other chats or your contacts, and she reads typed messages only. ### Do I need a Tracy account first? Yes. If you are signed out when you open Tracy's link, you are asked to sign in or sign up and then brought straight back. ### Does the WhatsApp route cost extra? It uses the same free results and credits as your account. --- Connect your phone and draft one interview from a WhatsApp chat; the button opens the Connect page with the WhatsApp guide. Or send us one role and we will build the interview plan, with a recording and report you can check. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=tracy-whatsapp-integration&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=tracy-whatsapp-integration&utm_content=demo) ## Sources - [Tracy AI Interviewer home page, customer proof](https://tracyinterview.ai/), retrieved 9 Nov 2026. - [Tracy AI Interviewer Connect page](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. - [Chat with Tracy AI Interviewer on WhatsApp (public guide)](https://tracyinterview.ai/help/whatsapp), retrieved 9 Nov 2026. - [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 5 Nov 2026. # AI Engineer Interview Questions for Hiring Teams > AI engineer interview questions for hiring teams: 15 questions on evals, trade-offs and failure cases with follow-ups, and how to run them at volume. - URL: https://tracyinterview.ai/blog/ai-engineer-interview-questions - Published: 2026-11-10 - Author: Tracy team > **In short** **Tracy AI Interviewer** interviews every AI engineer applicant, follows up on trade-offs and failure cases, and can run a coding task inside the same interview, so you read how they reason about systems that fail. > > Use the 15 AI engineer interview questions below, each with a follow-up and what to listen for. > > A recording, a timestamped transcript and a written assessment per question sit behind every result, and your team decides. > > Next step: paste five of these questions into a Tracy interview. Applications for an AI engineer role arrive with the same words on every CV: LLMs, RAG, fine-tuning, evals. Your head of engineering, or the person running lab operations, cannot interview all of them. The CV shows what someone has built with. It does not show how they behave when the system gives a confident wrong answer. AI engineer interview questions should test that. These are written for the interviewer. They ask about evaluation, retrieval, cost, safety and shipping, and each has a follow-up and a note on what a strong answer contains. They avoid trivia about one framework, which goes out of date in a quarter. Candidates can use them to prepare. Tracy AI Interviewer runs the part that does not scale. She asks each applicant in their own field, follows up on what they just said and waits for the end of the thought, with a coding task inside the same interview when you want one. For the wider engineering process, see [AI technical interviews](/blog/ai-technical-interviews), and for specialist vetting, [expert vetting](/use-cases/expert-vetting). ## Fifteen AI engineer interview questions, grouped by what they test Pick one or two questions from each group. Ask the follow-up only after the first answer. Choose your opening questions for the role and write down what a strong answer contains before the first candidate joins. Ask for a system the candidate shipped, because a rehearsed answer about a famous model is easy to find and a failure story is not. ### Group 1: evaluation *Table 1: Evaluation, with follow-ups and what a strong answer contains.* | Ask | Follow up with | Listen for | | --- | --- | --- | | 1. You changed a prompt and the demo looks better. How do you know it is better? | What would make you revert it? | A fixed set of cases, a baseline, a comparison on the same inputs. Names a regression they would accept or not. | | 2. How do you build an evaluation set for a task with no single right answer? | How do you keep it from going stale? | Real examples, written criteria, a second reviewer, refreshed from failures. Avoids one model grading its own output without checks. | | 3. Tell me about a metric that looked good and was wrong. | What did you use instead? | A real case, the cause, and a change of metric or sample. | ### Group 2: retrieval and data *Table 2: Retrieval and data, with follow-ups and what a strong answer contains.* | Ask | Follow up with | Listen for | | --- | --- | --- | | 4. A retrieval system returns plausible but wrong passages. Where do you look? | How would you tell chunking from embedding trouble? | Reads failing queries, separates retrieval from generation, checks chunk boundaries and filters before swapping models. | | 5. When would you fine-tune instead of improving the prompt or retrieval? | What did it cost you the last time? | Names the trigger, such as style or latency at volume, and the cost: data, drift, maintenance. | | 6. How do you treat training or evaluation data that may contain personal information? | Who do you ask first? | Removes or avoids it, asks the data owner and the legal team, keeps a record. | ### Group 3: trade-offs in production *Table 3: Trade-offs in production, with follow-ups and what a strong answer contains.* | Ask | Follow up with | Listen for | | --- | --- | --- | | 7. A feature costs too much per request. What are your levers? | Which would you pull first? | Shorter context, caching, a smaller model on easy cases, batching. Measures quality after each. | | 8. Latency doubled after a release. How do you find the cause? | What would you add so you see it sooner? | Traces by stage, compares versions, finds tokens or retries. Adds a dashboard or alert. | | 9. How do you decide between an API model and one you host? | What would change your answer? | Volume, privacy, latency and maintenance. Gives a number or a threshold, not a preference. | ### Group 4: failure cases and safety *Table 4: Failure cases and safety, with follow-ups and what a strong answer contains.* | Ask | Follow up with | Listen for | | --- | --- | --- | | 10. Tell me about a failure your system had in production. | What did you change afterwards? | A specific incident, the signal that caught it, a fix and a test added. | | 11. How would you stop a model from acting on instructions hidden in a document it reads? | What would you not rely on? | Separates instructions from data, limits tool permissions, asks for confirmation. Says a prompt alone is not enough. | | 12. A model refuses a harmless request. How do you handle it? | How do you know the fix did not open something else? | Collects examples, changes one thing, re-runs the safety cases as well as the helpful ones. | ### Group 5: shipping with a team *Table 5: Shipping with a team, with follow-ups and what a strong answer contains.* | Ask | Follow up with | Listen for | | --- | --- | --- | | 13. How do you explain to a product manager why a feature cannot be exact? | What do you offer instead? | Plain language about error rates, a fallback or a review step, an agreed threshold. | | 14. You are asked to ship in a week and the evals are not ready. What do you do? | What would you refuse to skip? | Proposes a smaller launch, keeps the key checks, says what is accepted and who agreed. | | 15. What did you last learn from reading someone else's code or paper, and what did you do with it? | What did you decide not to copy? | A specific source and a specific change. Can say why it did not apply to their case. | ## A strong answer names what would change its mind, and a weak one names a model The difference shows up in the second answer. These pairs are written by us as examples of how it sounds, not quotes from real candidates. *Table 6: Weak and strong answers to the same questions, written as examples.* | | Weak answer | Strong answer | | --- | --- | --- | | Question 1, the better prompt | "The outputs read more naturally now." | "I ran it on the same 80 cases as before. It improved tone and dropped accuracy on dates, so I would not ship it yet." | | Question 4, wrong passages | "I would switch to a bigger embedding model." | "I would read twenty failing queries first. If the right passage is not retrieved at all it is retrieval, not generation." | | Question 10, a production failure | "We had no major failures." | "The bot quoted a refund policy from an old document. We added a freshness filter and a test for it." | Listen for ownership. "The model hallucinated" tells you about the model. "I read the failing cases and found the filter" tells you about the person. ## One follow-up chain shows where the real answer starts This is an illustration with sample data, not a real interview. The candidate is fictional. It shows the shape of what Tracy runs: an opening question from your list, the answer, and a follow-up built on it. It starts from question 1. > **Tracy:** You changed a prompt and the demo looks better. How do you know it is better? > > **Candidate (fictional):** I run it on a fixed set of cases and compare with the old prompt. > > **Tracy:** What was in that set the last time you did this? > > **Candidate:** Eighty real support questions. The new prompt was friendlier but got dates wrong in eleven of them. > > **Tracy:** What did you decide? > > **Candidate:** I did not ship it. I added date cases to the set and tried again. The first answer is the textbook one. The second shows whether it ever happened. A Director at a top-tier early-stage VC says Tracy "sometimes brings up exactly those questions I might have forgotten to ask, which has been a pleasant surprise." ![Transcript card: better prompt, candidate says fixed test cases, Tracy asks what was in the set last time.](/blog/images/ai-engineer-interview-questions-1.png) *Tracy asks a follow-up built on the candidate's own answer. You see whether the evaluation story came from real work. Illustration with sample data* A follow-up built on the candidate's own answer shows whether they have done the work. ## Run the interview for every applicant without losing the follow-up A live call for every applicant does not survive a hundred of them, and a specialist pool is often the scarcest resource in the building. - **Your questions or ours.** Describe the role and Tracy drafts questions from the job description, which you edit (see the [product page](/product)). - **A coding task in the same interview.** Live coding, debugging or multiple choice, with follow-ups on approach, trade-offs and failure cases. - **CV check, as a setting.** With it on, Tracy can read the CV and ask about what it claims, so "built a RAG system" becomes a question about that system. See [how the CV check works](/blog/ai-interviewer-cv-check). - **Start on click.** Each candidate opens a link in a browser, with no account and no calendar, at any hour, and is told on the join screen that the conversation is recorded and transcribed. See the [participant experience](/participant-experience). - **A result you can check.** A recording, a timestamped transcript and a written assessment per question sit behind every result. See [AI interview reports](/blog/ai-interview-reports). It is $1 per credit, one credit per started 20 minutes of an interview that produced a result, so a 34-minute interview costs $2. The first 3 results are free with no card, there are no seats, and a session with no answers is never charged. See [pricing](/pricing). Customisation takes days: send us one role and we will build the interview plan. ![Recreated Tracy result report for a fictional AI engineer: recording, a transcript with the 7:30 answer on the eighty-question test set, and quoted evidence per question.](/blog/images/ai-engineer-interview-questions-2.png) *A recording and a timestamped transcript sit behind every Tracy result. Your lead reads the exact words before deciding. Illustration with sample data* A reviewer can open the recording and the transcript behind each answer. ## Keep four checks with people 1. **Your own engineer.** Have one score five or six transcripts blind. If they disagree with the written assessments, fix the questions before inviting the next hundred. 2. **The work.** Tracy asks about approach and trade-offs. Whether a design or a piece of code meets your bar is a reviewer's call. 3. **Integrity signals.** Tracy captures tab changes, gaze, face detection, delay before answers, fullscreen exits and external screens, and shows the evidence to a person. A signal is not proof, and with tools like chat assistants everywhere, it is the follow-up that tests the candidate. See [AI interview cheating detection](/blog/ai-interview-cheating-detection). 4. **The decision.** A person decides from the evidence. For ten applicants, a video call with one of your engineers is simple and works well; see [AI interview vs human interview](/blog/ai-interview-vs-human-interview). The case for Tracy starts when applicants outnumber your engineers' hours. To try it, create a Tracy interview for this exact role from these questions and read the recording and report on three real applicants. ## Common questions ### What are good AI engineer interview questions? Ask how they know a change is better, where a retrieval system fails, what they would cut to lower cost, and for a production failure they owned. Keep the follow-up beside each. ### Where can candidates find AI engineer interview questions on GitHub? Public lists, including GitHub repositories, give candidates the questions and sample answers, so expect rehearsed first answers. The follow-up is what separates them: ask what was in their evaluation set and what they decided. ### How do you test an AI engineer without a trivia quiz? Ask about decisions, not definitions: what they measured, what they traded and what failed. A short coding task with follow-ups on approach can sit in the same interview. ### Can an AI interviewer interview AI engineers? Yes. Tracy asks questions in the candidate's field, follows up on what they said and waits for the answer. Your team gets the recording, transcript and an assessment per question, and a person decides. ### Should candidates be allowed to use AI tools in the interview? That is your policy. Tell candidates the rule before they start. Tracy shows integrity signals with their evidence to a person, and the follow-up lets a reviewer hear whether the candidate can explain the answer. --- Create a Tracy interview for this exact role from these questions, then read the recording and report; the button opens the Connect page. Or send us one role and we will build the interview plan, with a recording and report you can check. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=ai-engineer-interview-questions&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=ai-engineer-interview-questions&utm_content=demo) ## Sources - [Tracy AI Interviewer home page, customer proof](https://tracyinterview.ai/), retrieved 9 Nov 2026. - [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer Connect page](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. # AI Interviews for AI Trainers > AI trainer screening that tests how applicants judge a model answer: job description, test versus interview, a worked example and the report you read. - URL: https://tracyinterview.ai/blog/ai-interviews-for-ai-trainers - Published: 2026-11-10 - Author: Tracy team > **In short** AI trainer screening breaks when applicants outnumber the reviewers who can talk to them. **Tracy AI Interviewer** interviews every applicant in their own field, follows up and waits for the answer. > > She tests what a CV hides: judging a model's answer, writing criteria, explaining a score. A reviewer decides from a recording, transcript and assessment per question. > > A Director at a top-tier early-stage VC says Tracy "sometimes brings up exactly those questions I might have forgotten to ask." > > Next: create an interview for one role. AI trainer screening breaks at the same point on every project. A project opens, applicants arrive faster than your reviewers can speak to them, and every CV says the right thing: a degree, a specialty, years in the field. None of that shows whether a person can say why a model's answer is wrong, write the criteria for scoring it, or explain a score so that a second grader reaches the same one. Tracy AI Interviewer runs that conversation for every applicant. She asks questions in the applicant's own field, follows up on what they just said and waits for the end of the thought. Your reviewers then read a recording, a timestamped transcript and a written assessment for each question, and decide. For a question bank to copy, see our guide to [AI trainer interview questions](/blog/ai-trainer-interview-questions). For the wider process, see how we handle [expert vetting](/use-cases/expert-vetting). ## What does AI trainer screening have to find out? AI trainers read a model's answer and judge it. Depending on the project they write a hard prompt with the correct answer, write the criteria for grading, rate and rank several answers, or flag one that is wrong or unsafe. Four abilities carry most of that work, and a CV shows none of them. | Ability | Ask | Follow up with | A strong answer | | --- | --- | --- | --- | | Judging a model's answer | A model's answer is correct but skips a step. How do you score it? | Two graders disagree on that mark. What in your rationale settles it? | Writes the missing step into the rationale and gives a rule another grader can apply. | | Writing criteria | Write one grading criterion for an answer that explains a concept in your field. | Give me an answer that meets it and one that only looks as if it does. | The criterion can be checked with a yes or no, and the two examples differ on one point. | | Explaining a score | You gave this answer 2 out of 5. Tell me why in two sentences. | What would move it to a 4? | Names the gap and what would change the score. A feeling is not a reason. | | Knowing when to stop | When do you send a task back instead of scoring it? | Tell me about the last time that happened. | Names a real case, says what was unclear and what they did next. | Give every applicant the same opening questions so you can compare them. ## How does an AI trainer job description set up the interview? An interview built from a vague AI trainer job description gets vague answers. The description should say what the person will do (write prompts, grade, rank, review), the field, the weekly hours, the quality bar and what experience counts. Tracy drafts the questions from that brief, and you edit them or bring your own. Interview style runs from quick to deep and from technical to behavioural, and you choose the skills the role must show. With CV check turned on, Tracy can also ask about what the CV claims. The [product page](/product) lists each setting. ## AI trainer test or interview: which should you use? An AI trainer test scores the output of a task. It shows what someone produced on the day, not how they got there or whether they can explain it. An interview shows the reasoning. A common setup is to interview everyone who applies, then give a short calibration task to the people who pass and score it against your own answer key. For why a conversation finds more than a recorded form, read [one-way versus two-way video interviews](/blog/one-way-vs-two-way-video-interviews). ## A worked example: one scoring question and its follow-up This is an illustration with sample data, not a real interview. The candidate is fictional. > **Tracy:** A model's answer is correct but skips a step. How do you score it? > > **Candidate (fictional):** I mark the missing step in my rationale and score it down one level. > > **Tracy:** Two graders disagree on that mark. What in your rationale settles it? > > **Candidate:** A rule. A skipped step that changes the reader's conclusion costs a level, and one that does not costs nothing. I would add an example to the rubric. > > **Tracy:** Where would that rule fail? ![Tracy asks how to score a correct answer that skips a step, then asks what settles a grader dispute.](/blog/images/ai-interviews-for-ai-trainers-1.png) *Tracy asks a follow-up built on the applicant's own answer. Illustration with sample data* A follow-up on the applicant's own answer shows how they reason about a score. ## What do your reviewers read after the interview? Each finished interview leaves a record a person can check. - **A recording and a timestamped transcript** behind every result. - **A written assessment for each question,** with its reasoning, assessed on the skills you choose for the role. - **A ranking** against the others who took the same interview, with the strongest flagged to you automatically. - **Integrity signals with their evidence,** covered below. - **One record you can send on,** with the resume match, interview score, strengths and gaps and the recording, shared by link or PDF with the lab team that needs it. A reviewer decides: they can move an applicant forward or invite them to interview again. See [AI interview reports](/blog/ai-interview-reports) for how to read one. ![Recreated Tracy result report for a fictional AI trainer: recording, timestamped transcript and each question with what was said and the quote.](/blog/images/ai-interviews-for-ai-trainers-2.png) *Tracy gives the reviewer what the trainer said to each question, with the time and the quoted words, next to the recording. Illustration with sample data* The report ties each answer to its question, the time it was said and the words behind it. ## Can you screen hundreds of AI trainer applicants without a queue? Each applicant opens a link in a browser, with no account and no scheduling, and starts when they click, at any hour. The interview is ready in under 10 seconds. A dropped applicant can rejoin within 10 minutes (customers can configure that), a reminder goes out after a day, and interviews that failed are chased automatically. Applicants are told on the join screen that the conversation is recorded and transcribed. See what applicants see on the [participant experience page](/participant-experience). Tracy also captures tab changes, abnormal eye tracking, live face detection, delay before answers, fullscreen exits and external screens, and shows the evidence to a person. A signal is a prompt to look, never proof, and nothing is decided in the interview. Our guide to [AI interview cheating detection](/blog/ai-interview-cheating-detection) covers what a reviewer should do with one. ## What stays with your team Tracy runs the first conversation. Three things stay with people. 1. **Credentials.** Not stated for Tracy, so check a degree, licence or work history separately where the project requires it. 2. **A reviewer of your own.** Have one of your graders read five or six transcripts and score them blind. If they disagree with the written assessments, change the questions before you invite the next hundred. 3. **The decision.** Nothing is decided in the interview. A person decides from the evidence, and Tracy publishes no bias audit. If you onboard a handful of trainers a month, a call with your own reviewer is cheaper. The case for Tracy starts when applicants outnumber reviewer hours; see [AI interview versus human interview](/blog/ai-interview-vs-human-interview). ## How to start: one role, one link Start with one role. 1. Write the job description for one role, with the task type, the field and the quality bar. 2. Open the [Connect page](/connect), describe the role and let Tracy draft the questions, or paste your own. 3. Share the link with the first applicants. The first 3 results are free, with no card. 4. Compare the reports with the people you would have approved from the CV alone. After the free results, Tracy costs $1 per credit, and one credit covers each started 20 minutes of an interview that produced a result. There are no seats and no upfront cost, a session with no answers is never charged, and 100 interviews of 20 minutes cost $100. The [pricing page](/pricing) has the current plan, and our guide to [AI interview cost](/blog/ai-interview-cost) sets prices side by side. Which path fits you: - Subject experts or AI trainers to vet: [expert vetting](/use-cases/expert-vetting). - Hundreds of applicants for one project: [volume hiring](/use-cases/volume-hiring). - Engineers for the same projects: [technical interviews](/use-cases/technical-interviews). - Programme or fellowship applications: [applications](/use-cases/applications). ## Common questions ### How do you screen AI trainers? Interview every applicant on the work itself: judging a model's answer, writing grading criteria, explaining a score and knowing when to send a task back. Follow up on each answer, then have a reviewer read the recording and transcript. Credentials are checked separately. ### What should an AI trainer job description include? The task type, the field, the weekly hours, the quality bar and how it is checked, and what experience counts. Tracy drafts interview questions from that description, and you edit them. ### What is an AI trainer test? A task, such as grading three answers against a rubric, that is scored on the output. It shows what someone produced. An interview adds how they reasoned and whether they can explain a score, so most teams use both. --- Create an interview for this exact role. Send us one role and we will build the interview plan in days, with a recording and report you can check. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=ai-interviews-for-ai-trainers&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=ai-interviews-for-ai-trainers&utm_content=demo) ## Sources - [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 5 Nov 2026. - [Tracy AI Interviewer Connect page](https://tracyinterview.ai/connect), retrieved 5 Nov 2026. - [Tracy AI Interviewer product page](https://tracyinterview.ai/product), retrieved 5 Nov 2026. - [Tracy AI Interviewer home page, customer testimonials](https://tracyinterview.ai/), retrieved 9 Nov 2026. # Candidate Screening Questions That Show Depth > Candidate screening questions that separate depth from polish: ten to ask, the follow-up for each, and what to listen for when you screen at volume. - URL: https://tracyinterview.ai/blog/candidate-screening-questions - Published: 2026-11-10 - Author: Tracy team > **In short** **Tracy AI Interviewer** asks your candidate screening questions of every applicant, then follows up on the answer, which is where depth shows. > > Questions come from your job description, edited or your own, and each result carries a recording and an assessment per question. > > Below: ten questions with the follow-up and what to listen for. > > Next: load them into one role and compare the follow-ups. Picture a recruiter who asks 120 applicants "tell me about a time you handled a conflict". She gets 120 smooth answers. Each one has a situation, an action and a result, and none can be told from the next. The question was fine. The screening stopped one step too early. Candidate screening questions only help if they separate people who have done the work from people who can describe it well. That separation rarely sits in the first answer. It sits in the follow-up. Tracy AI Interviewer asks your questions of every applicant and then asks the follow-up a good recruiter would, so the depth shows up before anyone books a call. ## Listen for specifics, trade-offs and their own part Three habits show up in answers from people who did the work, and are missing from answers that were rehearsed: - **Specifics.** A named tool, a number, a date, a person. Polish stays general. - **Trade-offs.** They say what they gave up and why. Rehearsed answers have no cost. - **Their own part.** "I" appears where the work was theirs, and "we" appears where it was not. They can say which is which. Write each of your screening questions so that one of these three can be heard, and write the follow-up before you need it. ## Split the knockout questions from the depth questions Keep a form for facts you can check without a conversation: work authorisation, location, notice period, salary range. Those are knockout questions, and a form does them well. Do not spend interview time on them. Keep the interview for what a form cannot test: how someone reasons, what they have actually done, and what they do when the first answer is not enough. Avoid questions about age, family, health or anything else unrelated to the job. They add risk and no signal. ## Ten candidate screening questions with follow-ups Ten depth questions for technical and high-volume roles. Ask the follow-up only after you have heard the first answer. Swap the nouns for your own stack. Our guide to [how to screen candidates](/blog/how-to-screen-candidates) covers the process around them. | Ask | Follow up with | Listen for | | --- | --- | --- | | What was the last thing you built or fixed that you are proud of? | What would you do differently now? | A real decision, not a title | | Walk me through how you diagnosed your last hard fault. | What did you rule out first, and why? | Order and reasons | | What did you leave out of your last project? | What did leaving it out cost? | A trade-off they own | | Which tool or method did you pick over another? | When would you pick the other? | Conditions, not loyalty | | Tell me about a number you moved. | What would it have been if you had done nothing? | Whether they know the cause | | Describe a time you were wrong at work. | Who found out, and what changed? | Honesty and a fix | | What part of your CV would you rather we asked about? | And the part you would rather we did not? | Candour about gaps | | How do you explain this work to someone outside your field? | What do they usually get wrong? | Plain words | | What did your team do that you disagreed with? | What did you do about it? | Action, not complaint | | What would you learn first in this role? | How would you know it was working? | A plan with a test | The table lists ten questions, the follow-up for each and what to listen for. Ask every applicant the questions you set for the role. The follow-ups differ with each answer, and that is the point. ## Follow up with an example, a number or a decision Weak screening stops at the first answer. A useful follow-up does one of three things: asks for an example ("which project?"), asks for a number ("how many?"), or asks for a decision ("why that one?"). If an applicant cannot give any of the three, you have learned something real. If they can, you have the evidence to write down. > **Tracy:** How do you handle a tight deadline? > > **Candidate (fictional):** I prioritise and communicate early. > > **Tracy:** What did you drop the last time that happened? This is an illustration with sample content. "I prioritise" fits anyone. What they dropped is a fact. ![Tracy asks about a tight deadline, hears a general answer, then asks what the candidate dropped last time.](/blog/images/candidate-screening-questions-1.png) *Tracy follows a general answer with a question about what the candidate dropped. Illustration with sample data* One short follow-up separates a general answer from a specific one. ## Run the questions at volume without cutting the follow-ups Asking these ten questions of two hundred applicants takes a team, and the follow-ups are the part that gets cut first. Tracy takes the job. Give her a job description and she drafts the questions, in a style from quick to deep and from technical to behavioural. Edit them or write your own. Each applicant starts from a link, at any hour, and Tracy follows up on what they said and waits for the end of the thought. For each result you get a recording, a timestamped transcript and a written assessment for each question. A person reads and decides. For technical roles, a coding task can sit in the same interview. See [AI screening interviews](/blog/ai-screening-interviews) and [screening hundreds of applicants](/blog/how-to-screen-hundreds-of-applicants). ![Four steps: ask one question, hear the answer, probe the detail, note the evidence.](/blog/images/candidate-screening-questions-2.png) *Tracy asks a question, hears the full answer, probes the detail and leaves evidence to read. Illustration with sample data* One question, the full answer, a probe, then evidence. ## Pilot the questions on one role A Director at a top-tier early-stage VC gives one piece of advice for teams starting out: "spend time designing the depth and structure of your interview questions during the initial setup." The ten above are a first draft for that work. Tracy costs $1 per credit, one credit per started 20 minutes, and the first 3 results are free with no card and no seats. Run your own questions on one role, then read three reports. Start on the [Connect page](/connect). Which path fits you: - Many applicants for the same role: [volume hiring](/use-cases/volume-hiring). - Technical roles: [technical interviews](/use-cases/technical-interviews). Good looks like this: you can read an applicant's answer and see, in their own words, a decision they made. ## Common questions ### What are good candidate screening questions? Questions that ask for something done, not something believed: the last project, the last fault, a number moved. Each needs a follow-up that asks for an example, a number or a decision. ### What should an applicant screening questionnaire include? Facts you can check on a form, such as work authorisation, location and notice period. Leave reasoning and experience to an interview, where you can follow up. ### How many screening questions should you ask? Ten is a workable first-screen list when each has a follow-up. Fewer works if every question has one. That is our suggestion, not a studied number. ### Can an AI interviewer ask candidate screening questions? Yes. Tracy AI Interviewer asks your questions or ones drafted from your job description, follows up on each answer and leaves a recording and an assessment per question. A person decides. --- See the method applied in one real interview. Send us one role and we will build the interview plan from your questions in days, with a recording and report you can check. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=candidate-screening-questions&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=candidate-screening-questions&utm_content=demo) ## Sources - [Tracy AI Interviewer home page, customer quotes](https://tracyinterview.ai/), retrieved 9 Nov 2026. - [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 5 Nov 2026. # Cybersecurity Interview Questions for Hiring Teams > Cybersecurity interview questions for hiring teams: 15 scenario questions with follow-ups, strong and weak answers, and how to run them at volume. - URL: https://tracyinterview.ai/blog/cybersecurity-interview-questions - Published: 2026-11-10 - Author: Tracy team > **In short** **Tracy AI Interviewer** interviews every security applicant on scenarios and follows up on how they would act, so your security lead reads judgement under pressure, not memorised acronyms. > > Use the 15 cybersecurity interview questions below, each with a follow-up and what to listen for. > > Each result keeps a recording, a timestamped transcript and an assessment per question, with integrity signals and their evidence for a person to review. > > Next step: paste five of these questions into a Tracy interview. A security opening draws applicants who all list CISSP, Splunk and "incident response". Your security lead has a queue of alerts and a handful of hours. The CV shows certifications and tools. It does not show what someone does in the first ten minutes of an incident. Cybersecurity interview questions should test that. These are for the security lead or hiring manager. They ask for scenarios and decisions across detection, response, access, cloud and communication, and each has a follow-up and a note on what a strong answer contains. For freshers, use groups 1 and 2 and accept lab or study examples. Candidates can use them to prepare. Tracy AI Interviewer runs the part that does not scale. She asks each applicant in their own field, follows up on what they just said and waits for the end of the thought. For the wider process, see [technical interviews](/use-cases/technical-interviews) and [software engineer interview questions](/blog/software-engineer-interview-questions) for the engineering side. ## Fifteen cybersecurity interview questions, grouped by what they test Pick one or two questions from each group. Ask the follow-up only after the first answer. Choose your opening questions for the role and write down what a strong answer contains before the first candidate joins. Keep every scenario defensive and invented: nobody should be asked to describe how to attack a real system. ### Group 1: foundations *Table 1: Foundations, with follow-ups and what a strong answer contains.* | Ask | Follow up with | Listen for | | --- | --- | --- | | 1. Explain the difference between authentication and authorisation with an example. | Where do systems confuse them? | Clear definitions, a concrete example, a real failure such as a user who can open another user's record. | | 2. What happens when you visit an HTTPS site, in terms of trust? | What would a certificate warning tell you? | Certificate chain, identity, key exchange at a sensible depth. Does not claim encryption proves a site is safe. | | 3. How would you explain phishing risk to a colleague who does not work in IT? | What would you ask them to do? | Plain words and one action: report it, do not click. Avoids blame. | ### Group 2: detection and triage *Table 2: Detection and triage, with follow-ups and what a strong answer contains.* | Ask | Follow up with | Listen for | | --- | --- | --- | | 4. An alert says a user logged in from two countries within an hour. What do you do? | What would make you escalate? | Checks travel, VPN, device and recent changes before assuming compromise. States the escalation trigger. | | 5. You have fifty alerts and time for ten. How do you choose? | What do you do about the forty? | Ranks by asset value and confidence, documents the rest, tunes noisy rules. | | 6. Tell me about a false positive you tuned. | What did you risk by tuning it? | A real rule, the change and the risk accepted. Reviews the result later. | ### Group 3: response *Table 3: Response, with follow-ups and what a strong answer contains.* | Ask | Follow up with | Listen for | | --- | --- | --- | | 7. You suspect a laptop is infected. What are your first three actions? | Who do you tell, and when? | Contain before wiping, preserve evidence, notify the right owner. Does not improvise alone. | | 8. A colleague says they clicked a link and entered their password. How do you respond? | What do you say to them? | Reset credentials, check sessions and mail rules, thank them for reporting. Keeps the tone blame-free. | | 9. Describe the last incident you worked on. | What would you do differently? | A specific timeline, their own role, a lesson and a change afterwards. | ### Group 4: access and cloud *Table 4: Access and cloud, with follow-ups and what a strong answer contains.* | Ask | Follow up with | Listen for | | --- | --- | --- | | 10. A storage bucket is public by mistake. What now? | How do you stop it happening again? | Closes access, checks logs for reads, then adds a guardrail or policy in the pipeline. | | 11. How do you handle secrets in code and pipelines? | What would you do after a key leaks? | Secret stores, rotation, scanning. After a leak: revoke first, then investigate use. | | 12. A team asks for admin rights for speed. What do you ask? | When would you say yes? | Asks what job needs it, offers narrower access, time-limited approval and a review. | ### Group 5: communication and judgement *Table 5: Communication and judgement, with follow-ups and what a strong answer contains.* | Ask | Follow up with | Listen for | | --- | --- | --- | | 13. Explain a vulnerability to a manager who must decide whether to delay a release. | What would you recommend? | Impact, likelihood and the fix cost in plain language, then one recommendation with a reason. | | 14. You find a flaw outside your scope. What do you do? | What would you never do? | Reports it through the proper channel and never tests it further without permission. | | 15. How do you stay current without chasing every headline? | What did you last change because of it? | A small routine, named sources and a concrete change. | ## A strong answer says what it would check first, and a weak one names a tool The difference shows up in the second answer. These pairs are written by us as examples of how it sounds, not quotes from real candidates. *Table 6: Weak and strong answers to the same questions, written as examples.* | | Weak answer | Strong answer | | --- | --- | --- | | Question 4, two countries in an hour | "I would lock the account." | "I would check for a VPN or a known device first. If the device is new and a mail rule changed, I escalate and lock it." | | Question 7, the suspect laptop | "I would wipe it and reinstall." | "I would isolate it from the network and keep it powered so we can collect evidence, then tell the owner and the incident lead." | | Question 10, the public bucket | "I would make it private." | "I would make it private, then read the access logs to see who read it, then add a rule that blocks public buckets in the pipeline." | Listen for ownership. "The SIEM flagged it" tells you about the tool. "I checked the device and the mail rules" tells you about the person. ## One follow-up chain shows where the real answer starts This is an illustration with sample data, not a real interview. The candidate is fictional and the scenario is invented. It shows the shape of what Tracy runs: an opening question from your list, the answer, and a follow-up built on it. It starts from question 4. > **Tracy:** An alert says a user logged in from two countries within an hour. What do you do? > > **Candidate (fictional):** I check whether it is a VPN or a known device before I assume a compromise. > > **Tracy:** The device is new and the VPN logs show nothing. What next? > > **Candidate:** Look for mail rules, new app consents and password changes. If any of those exist I escalate. > > **Tracy:** What do you tell the user while you wait? > > **Candidate:** That I am checking the sign-in and may reset their password. I ask them to avoid logging in until I confirm. The first answer is the textbook one. The second shows the order they would work in. The scenario changes with every answer, which a fixed question list cannot do. ![Transcript card: two-country login alert, candidate checks VPN first, Tracy asks what to do when the device is new.](/blog/images/cybersecurity-interview-questions-1.png) *Tracy asks a follow-up built on the candidate's own answer. Your security lead sees the order of actions, not a definition. Illustration with sample data* A follow-up built on the candidate's own answer shows how they would act, not what they memorised. ## Run the interview for every applicant without losing the follow-up A live call for every applicant does not survive a queue of them, and your security lead is also your incident responder. - **Your questions or ours.** Describe the role and Tracy drafts questions from the job description, which you edit (see the [product page](/product)). - **CV check, as a setting.** With it on, Tracy can read the CV and ask about what it claims, so "incident response" becomes a question about one incident. See [how the CV check works](/blog/ai-interviewer-cv-check). - **Start on click.** Each candidate opens a link in a browser, with no account and no calendar, at any hour, and is told on the join screen that the conversation is recorded and transcribed. See the [participant experience](/participant-experience). - **A result you can check.** A recording, a timestamped transcript and a written assessment per question sit behind every result. See [AI interview reports](/blog/ai-interview-reports). - **Integrity signals, with the evidence.** Tracy captures tab changes, gaze, face detection, delay before answers, fullscreen exits and external screens, and shows the evidence to a human reviewer. A signal is not proof, and your team decides. See [AI interview cheating detection](/blog/ai-interview-cheating-detection). It is $1 per credit, one credit per started 20 minutes of an interview that produced a result, so a 34-minute interview costs $2. The first 3 results are free with no card, there are no seats, and a session with no answers is never charged. See [pricing](/pricing). Because your reviewers will ask about the vendor: Tracy is SOC 2 Type II attested, and the report is available on request at hello@tracyinterview.ai. Customisation takes days. ![Tracy result report for a fictional security analyst: recording with timestamped transcript, and three questions with what they said and quoted evidence.](/blog/images/cybersecurity-interview-questions-2.png) *A recording and timestamped transcript sit behind every Tracy result. Your security lead reads each answer and its quote before deciding. Illustration with sample data* A reviewer can read what the candidate said to each question, beside the recording and transcript. ## Keep four checks with people 1. **Your own security engineer.** Have one score five or six transcripts blind. If they disagree with the written assessments, fix the questions before inviting the next hundred. 2. **Credentials.** Verify certifications at the issuer. A Tracy result is the recording, transcript and assessment per question. 3. **Background and clearance.** Whether a role needs screening beyond the interview is your call. 4. **The decision.** A person decides from the evidence. For ten applicants, a video call with one of your engineers is simple and works well; see [AI interview vs human interview](/blog/ai-interview-vs-human-interview). The case for Tracy starts when applicants outnumber your team's hours. To try it, create a Tracy interview for this exact role from these questions and read the recording and report on three real applicants. ## Common questions ### What are good cybersecurity interview questions? Ask scenario questions: what they do first with a suspicious login, how they triage fifty alerts, how they contain an infected laptop and how they explain a risk to a manager. Keep the follow-up beside each. ### What cybersecurity interview questions suit freshers? Use groups 1 and 2, and accept lab, course or home-network examples. Listen for the order they would work in and whether they say what they would check, rather than recalled terms. ### How do you test a candidate's cybersecurity judgement without an exploit exercise? Use defensive scenarios with invented details and change one fact after each answer. The follow-up shows whether they adapt. ### Can an AI interviewer interview cybersecurity candidates? Yes. Tracy asks questions in the candidate's field, follows up on what they said and waits for the answer. Your team gets the recording, transcript and an assessment per question, and a person decides. ### Does Tracy verify a security certification? A Tracy result is the recording, transcript and assessment per question. Verify certifications at the issuer. --- Create a Tracy interview for this exact role from these questions, then read the recording and report; the button opens the Connect page. Or send us one role and we will build the interview plan, with a recording and report you can check. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=cybersecurity-interview-questions&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=cybersecurity-interview-questions&utm_content=demo) ## Sources - [Tracy AI Interviewer home page, customer proof](https://tracyinterview.ai/), retrieved 9 Nov 2026. - [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer Connect page](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. # How to Spot a Fake Candidate in a Video Interview > How to spot a fake candidate: signs in the application, the interview and onboarding, from FBI alerts, and the evidence a person should review. - URL: https://tracyinterview.ai/blog/how-to-spot-a-fake-candidate - Published: 2026-11-10 - Author: Tracy team > **In short** **Tracy AI Interviewer** asks follow-ups on each person's own CV and work, then shows a reviewer signals with their times, such as delay before answers and tab changes, beside a recording and transcript. > > A fake candidate is someone who is not who they say they are, or whose claimed experience is invented. No single sign proves it. > > FBI alerts list signs such as lip movement out of step with audio and trouble answering simple questions about location or past. > > Next step: send us one role and review the evidence on one interview. ## How to spot a fake candidate: three places to look How to spot a fake candidate comes down to collecting evidence in three places and letting a person weigh it: the application, the interview and the first weeks of onboarding. A fake candidate here means an applicant who is not who they claim to be, or whose claimed work is invented. It is not a candidate who is nervous, has a poor connection or prepared well. The signs below come from two sources, and our guide to [AI interview cheating detection](/blog/ai-interview-cheating-detection) covers the AI-fed-answers case separately. Where the FBI's Internet Crime Complaint Center has described a sign, the table says so and gives the date. Where a sign comes from Tracy's signals or from our own practice, it says that. Nothing on the list is proof by itself. ## Signs in the application Some checks work before anyone joins a call. The FBI's January 2025 advice to employers includes cross-checking HR systems "for other applicants with the same resume content and/or contact information", and reviewing applicants' communication accounts, because North Korean IT workers "have reused phone numbers (particularly voice-over-IP numbers) and email addresses, on multiple resumes purportedly belonging to different applicants". The alert is about one scheme, and the same check is cheap for any pool of applicants. If your applicant tracking system can search by phone and email, run that search before the interview and not after the offer. ## Signs in the interview These are the signs a reviewer can notice during a video interview, with where each one comes from. | Sign | Where it comes from | What a reviewer does | |---|---|---| | Lip movement and actions that do not match the audio, or a cough that does not line up with the picture | FBI IC3, 28 June 2022 | Read the transcript around that minute, then ask the person to answer again in a short live call | | Cannot answer simple questions about where they are located or key details of their past | FBI IC3, 16 May 2024 | Ask the same everyday questions of every candidate and compare the answers with the CV | | Extra noise, or sounds as if the person is surrounded by others doing similar work | FBI IC3, 16 May 2024 | Note the time and ask in a follow-up whether anyone else is helping | | The same resume content, phone number or email on applications under different names | FBI IC3, 23 January 2025 | Cross-check your applicant records before the interview | | A pause before almost every answer | Tracy signal: delay before answers | Listen to a few of them. A slow connection looks the same | | Leaving fullscreen, switching tabs or a second screen | Tracy signals | Ask what the person was doing, then follow up on the next answer | | Knows the headline of a project but not one detail below it | Our practice | Follow up on the CV claim: what broke, who decided, what changed | The table lists seven signs, each with the action that follows it. The FBI alerts describe remote-work fraud and one North Korean scheme, and they give no rate. Two cautions apply. First, ask any check of everyone. A location question put only to some candidates is a screen on those candidates, and accent, name, camera quality and connection speed are not signs of anything. Second, one sign is a reason to look, not a reason to reject. ## Signs after the interview Checking does not stop at the interview. The FBI's January 2025 alert advises verifying identity "during interviewing, onboarding, and throughout the employment of any remote worker" and watching for changes in address or payment platform during onboarding. If the person who starts is not the person who interviewed, a recording of the interview is the first thing to compare. In the US, the FBI's 2022 alert asks companies that identify this type of activity to report it to the IC3. ## How Tracy puts the evidence in front of a reviewer Tracy AI Interviewer runs the first round for every applicant, so the evidence is the same kind for each person. When the CV check is on, she reads the CV and asks about it (see [the CV check](/blog/ai-interviewer-cv-check)). She asks questions in the person's own field, follows up on what they said and waits for the answer. For engineers she can set a coding task and probe the approach and trade-offs, so a copied solution still has to be explained. See [AI technical interviews](/blog/ai-technical-interviews) and our [technical interviews use case](/use-cases/technical-interviews). She captures tab changes, abnormal eye tracking (gaze), live face detection, delay before answers, leaving fullscreen and external screens, and shows the evidence to a person in the report, with times. A recording of both sides and a timestamped transcript sit behind each result, with a written assessment per question. Nothing is decided in the interview. The person interviewed is told on the [join screen](/participant-experience) that the conversation is recorded and transcribed, and the decision stays with your team. Identity verification is a separate step in your process. Tracy's recording and signals are evidence for the person who runs that check. ![Sample chain: Tracy asks about a payments migration on the CV, the candidate answers, Tracy asks about writes during the cutover.](/blog/images/how-to-spot-a-fake-candidate-1.png) *Tracy follows up on a CV claim, so a reviewer can see whether the person can explain their own work. Illustration with sample data* Tracy follows up on a claim in the CV, so a reviewer sees whether the person can go one level deeper. ## A review routine for when a sign appears When a sign appears, this routine keeps the decision fair and the evidence in one place. 1. **Read around the signal.** Open the transcript at the minute it appeared. 2. **Compare with the CV.** Does the person explain the work they claim, one level deeper? 3. **Ask in a short live call.** Put a different follow-up on the same topic, with the same questions you would ask anyone. 4. **Run the separate identity check** your process already has. 5. **Record the decision** with the evidence behind it, so a second person can review it. Tracy's signals sit inside the [interview report](/blog/ai-interview-reports) with their times, which saves the reviewer from rewatching every interview. The illustration below shows what that looks like with sample data. ![Recreated signals panel, sample data: tab changes three times and delay before answers five times, with times; three signals not detected.](/blog/images/how-to-spot-a-fake-candidate-2.png) *Tracy lists each signal with its times in the report, so a reviewer knows where to look. Illustration with sample data* Signals appear in the report with their times, and your team decides what they mean. ## Common questions ### How do you spot a fake candidate in a video interview? Look for several signs together: video that does not match the audio, trouble with simple questions about location or past, extra noise from others, and answers that fall apart one question deeper. Each is a reason to look closer, and a person decides. ### How do you identify fake candidates in recruitment? Start before the interview by cross-checking applicants for the same resume content, phone number or email. Then add follow-ups on the CV, a separate identity check and a review of the evidence. ### Can software tell you if a candidate is fake? No tool should be read as a verdict. Tracy shows signals with their evidence to a reviewer and asks follow-ups. Identity is checked in your own process, and a person makes the decision. ### What should I do if I suspect a fake candidate? Read the transcript around the signal, ask the person to explain in a short live call and run your identity check. In the US, the FBI asks companies that identify this type of activity to report it to the IC3. ### Is a nervous or slow candidate a fake candidate? No. Pauses, a poor connection and nerves are common in honest interviews. That is why one signal is never enough to reject someone. --- See the method applied in one real interview: send us one role and we will build the interview plan in days, with a recording and report you can check. The first 3 results are free, with no card. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=how-to-spot-a-fake-candidate&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=how-to-spot-a-fake-candidate&utm_content=demo) ## Sources - [FBI IC3 public service announcement I-062822-PSA, 28 June 2022: deepfakes and stolen PII used to apply for remote work](https://www.ic3.gov/PSA/2022/PSA220628), retrieved 9 Nov 2026. - [FBI IC3 public service announcement I-051624-PSA, 16 May 2024: North Korea leverages US-based individuals to defraud US businesses](https://www.ic3.gov/PSA/2024/PSA240516), retrieved 9 Nov 2026. - [FBI IC3 public service announcement I-012325-PSA, 23 January 2025: North Korean IT workers conducting data extortion](https://www.ic3.gov/PSA/2025/PSA250123), retrieved 9 Nov 2026. # How Accelerators Can Meet Every Founder Who Applies > How accelerators select startups, where partner hours cap who gets met, and how to give every founder who applies a first interview. Pilot plan included. - URL: https://tracyinterview.ai/blog/meet-every-founder - Published: 2026-11-10 - Author: Tracy team > **In short** **Tracy AI Interviewer** gives every founder who applies a first interview from one link on your apply page, so partner hours go to the strongest applications and not to whoever wrote the best form. > > Techstars and Y Combinator both run an application, then an interview. Techstars says staff read every application, so the first look decides who is met. > > Tracy starts on click at any hour and follows up on what each founder says. A person decides from a recording, a timestamped transcript and an assessment for each question. > > Next step: send us one programme. A founder submits an application at 11 p.m. on the last day of the call. The form is complete, the deck is attached and the one-line description is sharp. Three weeks later it sits in a long list, and the partner who could have asked the one question that shows what this founder knows has no hours left to ask it. That is how accelerators select startups in practice. They decide who gets a conversation, and weigh the rest from the form. This page shows where that decision happens and what a first look has to capture. It also shows how Tracy AI Interviewer gives every applicant that first conversation: one link in your form, any hour, no calendar, follow-ups on what each founder says. Your partners then read the result and choose. If you run an accelerator, a fellowship or a VC programme, the sections below end in a pilot you can start with one cohort. Our guide to [running the accelerator interview](/blog/accelerator-interview) covers what to ask, and the [applications use case](/use-cases/applications) shows the flow. ## The cut happens at the first look, before anyone is met Techstars and Y Combinator both describe an application followed by an interview. Techstars says its staff review every application, and Y Combinator's interviews are 10-minute conversations with partners who have read the application. In our view the biggest cut comes earlier, when a form and a deck decide who is invited to talk. A form cannot ask a follow-up, so it rewards the founder who writes well and can miss the one who builds well. A first conversation for everyone stops the form being the filter. ## Programmes publish how accelerators select startups, and every stage uses partner hours Programmes describe their own process in public. Three facts show where the hours go: - **Reading.** Techstars says that "thousands of companies apply to Techstars, and our staff personally review every application that comes in" (its post on the application and interview process, dated 12 January 2026). - **Founder effort.** The same post says most teams need one to four hours to complete a first application. A form that no conversation follows costs founders that time too. - **The conversation.** Y Combinator's interview guide describes 10-minute conversations over Zoom with two to three partners who have read the application. Techstars managing director Andres Barreto says the first interview tries to understand a company "well enough to write a strong investment memo" (post dated 22 January 2026). Put together, the first conversation is short, scarce and built on what the founder already wrote. That suits a programme with a small cohort. It becomes a bottleneck for one that gets thousands of applications, because reading an application does not meet the founder. ## A first look has to capture five things a form cannot A first look that is fair to a founder captures five things: - **The founder in their own words**, describing the customer and the problem without a field to fill. - **Customer and pain, with a source.** Techstars lists it first of the four things it asks founders to be ready on. - **What they built first and what they skipped**, which shows judgement about speed. - **How they handle a question they cannot answer yet.** Techstars tells founders that saying "I don't know" is as relevant an answer as any. - **A record a second partner can read** without having been there. None of these fit a form field. ## Give every founder a first interview from the form You add a link in your application form to an apply page where each founder takes the interview, as a step after the short form or instead of a long answer. Each founder starts when they click, at any hour, in a browser, with no calendar and no partner present, and is told on the join screen that the conversation is recorded and transcribed. The interview is ready in under 10 seconds, and a founder whose connection drops can rejoin within 10 minutes (a setting each customer configures). She asks your questions, or questions drafted from a description of your programme that you edit. Then she does the part that does not scale: she listens, waits for the end of the thought and asks about what the founder said. A founder who says two of six pilots renewed gets asked what those two had in common. Our guide to [how AI interviews work](/blog/how-do-ai-interviews-work) walks through a session. Your partners get a recording of both sides, a timestamped transcript and a written assessment for each question with its reasoning. Results can be shared with the rest of your team. A partner decides: they can move a founder forward or invite them to interview again, and the history stays on the record. See [what a good AI interview report shows](/blog/ai-interview-reports). See [what Tracy does](/product) and [what a founder sees](/participant-experience). ![Transcript card: Tracy asks what two renewing pilots shared, the founder names one clinic manager, Tracy asks what that manager did first.](/blog/images/meet-every-founder-1.png) *Tracy follows up on what a founder says about their pilots, so partners meet every applicant on an apply page reached from your form. Illustration with sample data* Tracy asks the follow-up a partner would ask about a founder's own pitch. ## Cost is small enough to cover the whole list You pay for results, not seats: $1 per credit, one credit per started 20 minutes of an interview that produced a result. One hundred founders at 20 minutes each cost $100, and the first three results are free with no card. See the [pricing page](/pricing). ![Four icons: a link in the form, a founder pitching, Tracy following up, and a partner reading the result.](/blog/images/meet-every-founder-2.png) *Tracy gives every founder a first interview from the form, so partner time goes to reading results. Illustration. Illustration with sample data* Every founder gets a first interview from the form, and a partner reads the result. ## What a top-tier early-stage VC saw when it moved first interviews to Tracy A Director at a top-tier early-stage VC describes the before. "Initially, we conducted 30-minute, one-on-one Zoom interviews with every founder." Some days held "eight 30-minute interviews scheduled back-to-back". With preparation, note-taking and review, "I was effectively spending around six hours a day on screening." An earlier AI interviewing tool still took "about an hour to go through their answers." The after, in his words: "With Tracy, I can now review and assess those candidates in approximately 1.5 hours. That's a time saving of more than 50%, and in practice, closer to 75%." He also notes: "For borderline candidates who are difficult to evaluate based on quantitative scores alone, Tracy makes it much easier to share their responses and assessment results with other team members for further discussion." The advice: "spend time designing the depth and structure of your interview questions during the initial setup." ## Keep these calls with your partners A first interview does not make the selection. These stay with people: - **Who moves to the next round.** Tracy gives the evidence. A partner reviews the evidence and decides. - **Fit with your thesis.** That is a judgement about your portfolio. - **Checking the claims.** A founder says two pilots renewed. A partner, or diligence, confirms it. - **Anything about the founder's circumstances.** Visa, funding and availability are conversations, not scores. A form is enough when your cohort is small enough for partners to meet every applicant. Once it is not, the first look is where you lose people. ## Pilot on one intake before you change the programme Run one intake and decide first what would make you stop. Have two partners read the same ten results and say who they would advance, then compare with the form-only shortlist. **Stop condition.** If the partners agree no more often than they did from forms alone, or many founders abandon the interview step, rewrite the questions and run the next intake before widening the pilot. Want it set up for you? Send us one programme and we will build the interview plan in days. ## Measure whether every founder was met, not how many applied Applications received is the number everyone reports and the one that tells you least, so set it aside. Five measures show whether the process works: - **Share of completed forms with a first interview.** The target is all of them. - **Partner hours per founder advanced.** - **Founders advanced who were not on the form-only shortlist.** This is the real test of whether the interview adds information. - **Interview completion rate.** A low rate points at the questions or the join step, not at the founders. - **Agreement between two partners reading the same result.** Read them after the first intake. If the third is zero, the interview adds nothing and you can stop. ## Common questions ### How do accelerators select startups? Techstars and Y Combinator both run an application followed by an interview. Techstars says its staff review every application, then invites a share of founders to interviews that test customer and pain, speed, market and differentiation. The first look decides who is met at all. ### What do accelerators look for in a founder? Techstars asks founders to be ready on four things: customer and pain, speed and execution, market size and differentiation. It also looks for honesty about what a founder does not know. A first interview with follow-ups shows these better than a form does. ### Can an AI interviewer interview founders? Yes. Tracy AI Interviewer asks your questions or ones drafted from your programme description, follows up on what each founder says and waits for the answer. Your partners read a recording, a timestamped transcript and an assessment for each question, and a person decides. ### How long should a first founder interview be? Pick the length to match your questions. A 20-minute interview costs $1, one credit per started 20 minutes. --- Create an interview for your own programme: send us one programme and we will build the interview plan in days, with a recording and report you can check. The first 3 results are free, with no card. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=meet-every-founder&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=meet-every-founder&utm_content=demo) ## Sources - [Techstars, Inside a Techstars Accelerator: The Application & Interview Process, 12 Jan 2026](https://www.techstars.com/blog/advice/inside-a-techstars-accelerator-the-application-and-interview-process), retrieved 3 Nov 2026. - [Techstars, How To Prepare for Your First Techstars Interview, 22 Jan 2026](https://www.techstars.com/blog/founder-advice/how-to-prepare-for-your-first-techstars-interview), retrieved 3 Nov 2026. - [Y Combinator, interview guide (undated)](https://www.ycombinator.com/interviews), retrieved 3 Nov 2026. - [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 5 Nov 2026. - [Tracy AI Interviewer home page, customer testimonial](https://tracyinterview.ai/), retrieved 9 Nov 2026. # Skills-Based Hiring: How to Run It With Interviews > Skills based hiring means checking skills for every applicant, not only a shortlist. How to set the skills, interview everyone and score the evidence. - URL: https://tracyinterview.ai/blog/skills-based-hiring - Published: 2026-11-10 - Author: Tracy team > **In short** **Tracy AI Interviewer** makes skills based hiring work for every applicant: you choose the skills, she interviews each person and follows up on what they say. > > Each result comes with a recording, a transcript and a written assessment per question, so a person can check the score. > > A person decides who moves forward. > > Next: run one role in Tracy and read the reports. Picture a recruiter who removes "degree required" from a job ad on Monday. By Friday the shortlist looks the same as last year, because the same person still skims the same CVs for the same names. The ad changed. The filter did not. Skills based hiring is the practice of deciding who moves forward on evidence of what a person can do, not on where they studied or which titles they held. It only works if you can collect that evidence for every applicant. That is the hard part, and it is what this guide covers: which skills to name, how to check them with an interview, and how to score what you hear. Tracy AI Interviewer does the checking for every applicant, and our guide to [screening candidates](/blog/how-to-screen-candidates) covers the wider screening process. ## Skills based hiring fails when only the job ad changes Dropping a requirement from the ad is not the same as hiring differently. A February 2024 study by the Burning Glass Institute and Harvard Business School's Managing the Future of Work Project followed 11,300 roles at large firms, drawn from the career histories of about 65 million US workers, between 2014 and 2023. On average, firms that removed a degree requirement raised the share of non-degreed hires in those roles by about 3.5 percentage points. Only 3.6 percent of roles dropped a requirement, so the authors estimate the real effect at not even 1 in 700 US hires. The gap was not even. Nearly all of the change came from the 37 percent of firms that backed the policy with new hiring practice. About 45 percent changed in name only. Where it worked, non-degreed hires had a two-year retention rate of 58 percent, against 48 percent for their college-educated colleagues. The study measures degree requirements, which is one part of skills based hiring, but the lesson carries: the practice has to change, not the wording. ## Name the skills and their evidence before the job ad Pick three to five skills the work really needs, and write what evidence of each would sound like in an interview answer. Weight them before you meet anyone. This is the step most teams skip, and every later step depends on it. | Skill (support engineer) | Weight | Evidence worth hearing | | --- | --- | --- | | Troubleshooting | 35% | Names the steps in order, says what each ruled out | | Customer care | 25% | A specific customer, what they said, what changed | | Written clarity | 20% | Explains a fault in plain words, no jargon wall | | Ownership | 20% | Says what they would change next time | The table sets out four sample skills for a support engineer, each with a weight and the evidence worth hearing. The weights are our example, not a standard. Set yours with the people who do the job. If you cannot say what a strong answer sounds like, an interview cannot test it, and a CV filter will fill the gap with proxies. ## Interview every applicant on the skills, not a shortlist A CV filter picks a shortlist, and a shortlist is where skills based hiring usually stops. Suppose someone reads 400 applications and chooses 30. The other 370 are judged on a page. If a skill is the test, every applicant needs a chance to show it. Tracy gives each applicant a first interview from one link. It starts on click at any hour, with no scheduling and no recruiter present, and the applicant needs only a browser, with no account. Questions come from your job description and you can edit them or write your own. Tracy reads the CV when the CV check is on, asks in the applicant's own field and follows up on what they said. The skills you choose become what she evaluates. Our guide to [how AI interviews work](/blog/how-do-ai-interviews-work) shows the flow end to end. ![Four steps: name the skills, every applicant interviews, Tracy follows up, a recording and transcript to read.](/blog/images/skills-based-hiring-1.png) *Tracy interviews every applicant on the skills you set and leaves a recording and transcript to read. Illustration with sample data* The skills come first; Tracy then interviews every applicant on them and leaves evidence to read. ## Follow up on the claim, not the CV line A claim on a CV is a starting point. "Data modelling" can mean a course or a career. The follow-up is what separates them, and it is what a form cannot ask. > **Tracy:** You list data modelling. What was the last model you built? > > **Applicant (fictional):** A churn model for a subscription app. > > **Tracy:** Which inputs did you leave out, and why? This is an illustration with sample content. A strong answer names a decision and its reason. A weak one repeats the job title. Either way the reviewer has the words, not an impression. ![Tracy asks an applicant about their last model, then which inputs they left out and why.](/blog/images/skills-based-hiring-2.png) *Tracy asks about the last thing the applicant built, then about a decision inside it. Illustration with sample data* The follow-up turns a listed skill into an example the reviewer can judge. ## Assess on your skills and keep a person in charge Each applicant is assessed on the skills you choose. For each question, the reviewer gets a written assessment with its reasoning, and behind every result sit the recording and a timestamped transcript. A person decides: they can move someone forward or invite them to interview again, and the history stays on the record. See [AI interview reports](/blog/ai-interview-reports) for what the report holds. ## Where this stops Tracy is an interview. For work samples, take-home tasks and reference checks, use your own process, and a person makes the decision. For roles where the skill is a deliverable, such as a portfolio or a design, you still want to see the work. For technical roles, a coding task can sit inside the same interview, with follow-ups on approach and trade-offs. Tracy publishes no bias audit, so do not describe the process as validated or fair on her say-so. What an interview gives you is evidence a person can read and challenge. If you hire a handful of people a year, a structured conversation by your own team is enough. The case for Tracy starts when applicants outnumber the hours you have. ## Pilot it on one role Run one role as a pilot, with a named owner and a stop condition. 1. Choose one role where you have already removed a requirement from the ad. 2. Write four skills and the evidence for each. Load the skills into Tracy with your questions. 3. Interview every applicant for four weeks. Let the owner read the reports for the people your CV screen would have rejected. 4. Stop if the owner would not move anyone from that group forward. 5. Measure applicants who finished, hours of reviewer time per hire, and hires from outside your usual CV pattern. On time spent, a Director at a top-tier early-stage VC says: "I was effectively spending around six hours a day on screening. With Tracy, I can now review and assess those candidates in approximately 1.5 hours." Tracy costs $1 per credit, one credit per started 20 minutes, and the first 3 results are free with no seats. Try one on the [Connect page](/connect), and see [pricing](/pricing). Which path fits you: - Many applicants for the same role: [volume hiring](/use-cases/volume-hiring). - Technical roles: [technical interviews](/use-cases/technical-interviews). - People teams setting up screening: [people teams](/use-cases/people-teams). Good looks like this: the shortlist contains people your old filter would not have picked, and the owner can say why each one is there. ## Common questions ### What is skills based hiring? Choosing who moves forward on evidence of what a person can do, such as work samples, tests or interview answers, instead of degrees, titles or the employers on their CV. ### How is skills based hiring different from traditional hiring? Traditional hiring screens on proxies, like a degree or a past title. Skills based hiring defines the skills first and checks them directly. In the Burning Glass and Harvard study, changing the ad alone moved few hires. ### What are skills based hiring best practices? Name three to five skills, weight them, write what good evidence looks like, check every applicant and not only a shortlist, score on the same scale, and keep a person making the decision. ### Can an interview check skills for every applicant? Yes. Tracy AI Interviewer gives each applicant a first interview on your skills, follows up on their answers and leaves a recording, a transcript and an assessment per question. --- See the method applied in one real interview. Send us one role and we will build the interview plan in days, with a recording and report you can check. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=skills-based-hiring&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=skills-based-hiring&utm_content=demo) ## Sources - [Burning Glass Institute and Harvard Business School Managing the Future of Work Project (2024), Skills-Based Hiring: The Long Road from Pronouncements to Practice](https://www.burningglassinstitute.org/research/skills-based-hiring-2024), retrieved 9 Nov 2026. - [Burning Glass Institute and Harvard Business School Managing the Future of Work Project (2024), Skills-Based Hiring: The Long Road from Pronouncements to Practice](https://www.burningglassinstitute.org/research/skills-based-hiring-2024), retrieved 9 Nov 2026. - [Burning Glass Institute and Harvard Business School Managing the Future of Work Project (2024), Skills-Based Hiring: The Long Road from Pronouncements to Practice](https://www.burningglassinstitute.org/research/skills-based-hiring-2024), retrieved 9 Nov 2026. - [Burning Glass Institute and Harvard Business School Managing the Future of Work Project (2024), Skills-Based Hiring: The Long Road from Pronouncements to Practice](https://www.burningglassinstitute.org/research/skills-based-hiring-2024), retrieved 9 Nov 2026. - [Tracy AI Interviewer home page, customer quotes](https://tracyinterview.ai/), retrieved 9 Nov 2026. - [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 5 Nov 2026. # Data Analyst Interview Questions for Hiring Teams > Data analyst interview questions for hiring teams: 15 questions on SQL and reasoning with follow-ups, strong and weak answers, and how to run them. - URL: https://tracyinterview.ai/blog/data-analyst-interview-questions - Published: 2026-11-09 - Author: Tracy team > **In short** **Tracy AI Interviewer** interviews every analyst applicant, with a SQL or debugging task inside the interview and follow-ups on their approach, so your analytics lead sees how someone reasons from data. > > Use the 15 data analyst interview questions below, each with a follow-up and what to listen for. > > A recording, a timestamped transcript and a written assessment per question sit behind every result. > > Next step: paste five of these questions into a Tracy interview. An analyst opening brings in two hundred applications with SQL, Excel and "storytelling with data" on every CV. Your analytics lead has an hour a day for screening. The CV shows which tools someone has used. It does not show whether they would question a number before presenting it. Data analyst interview questions should test that habit. These are for the analytics lead or hiring manager. They ask for reasoning about messy data, metrics, queries and communication, and each has a follow-up and a note on what a strong answer contains. Most work as a spoken conversation, so they suit a first-round interview before any take-home. Candidates can use them to practise. Tracy AI Interviewer runs the part that does not scale. She asks each applicant in their own field, follows up on what they just said and waits for the end of the thought, with a SQL or debugging task inside the same interview when you want one. For the adjacent role, see [data scientist interview questions](/blog/data-scientist-interview-questions), and for the process, [technical interviews](/use-cases/technical-interviews). ## Fifteen data analyst interview questions, grouped by what they test Pick one or two questions from each group. Ask the follow-up only after the first answer, and write down what a strong answer contains before the first candidate joins. Use invented tables and numbers in every case, so nobody needs your data. ### Group 1: questioning the data | Ask | Follow up with | Listen for | | --- | --- | --- | | 1. Weekly active users dropped 12 percent on a Monday. What do you check first? | What would make you say it is a tracking problem? | Checks the definition, tracking changes, a holiday or release, and segments before any story. Ranks by cost to check. | | 2. You get a table with no documentation. How do you learn what it holds? | What would you ask its owner? | Row count, keys, null rates, date range, a spot check against a source, then a short list of questions. | | 3. Tell me about a number you had to take back. | What check do you run now? | A real case, an owned mistake, a concrete habit. | ### Group 2: SQL in conversation | Ask | Follow up with | Listen for | | --- | --- | --- | | 4. A join doubles your revenue total. What happened? | How would you prove it? | Fan-out from a one-to-many join. Compares row counts before and after, aggregates first or joins on a unique key. | | 5. Find customers who bought in January but not in February. How would you write it? | What changes if there are ten million rows? | Anti-join or a left join with a null test, a clear date range. Thinks about indexes and sampling when asked. | | 6. When would you use a window function instead of a group by? | Give a case from your work. | Running totals, rankings, previous-row comparisons, anything that keeps row detail. A real example. | ### Group 3: metrics and causes | Ask | Follow up with | Listen for | | --- | --- | --- | | 7. Conversion rose after a redesign. Did the redesign cause it? | What else changed that week? | Asks about timing, traffic mix, seasonality and a control. Says what would change their mind. | | 8. Pick a metric your team uses. How could it mislead? | How would you tell your manager? | A real failure mode such as averages hiding a skew. Proposes a second metric and says it plainly. | | 9. An A/B test shows a 2 percent lift on a small sample. What do you say? | What would you want before acting? | Sample size, run time, interval or noise. Does not declare a winner from one number. | ### Group 4: tools and quality | Ask | Follow up with | Listen for | | --- | --- | --- | | 10. How do you check a spreadsheet or query somebody else wrote? | What do you do when you cannot reproduce their total? | Reads the logic, recomputes a slice, traces to the source. Reports differences without blame. | | 11. What would you automate in a weekly report, and what would you leave manual? | What breaks first when you automate it? | Names a specific step, a check that runs with it and a failure they have seen. | | 12. How do you decide which chart to use? | When did a chart mislead you? | Starts from the question. Names a bad truncated axis, a pie with eleven slices or a dual axis. | ### Group 5: working with the business | Ask | Follow up with | Listen for | | --- | --- | --- | | 13. A stakeholder asks for a dashboard. What do you ask before you build it? | What would make you say no? | Asks what decision it supports and who acts on it. Says no to a dashboard nobody opens. | | 14. Explain a regression result to someone who has never taken statistics. | What would you leave out? | Plain words, one example, honest about uncertainty. | | 15. Your analysis contradicts what the director believes. What do you do? | How do you know you are right? | Checks the work again, shows the evidence, stays open to a missing factor. | *Table 1. The 15 questions, each with a follow-up and what to listen for (written by us).* ## A strong answer questions the number, and a weak one explains it The difference shows up in the second answer. These pairs are written by us as examples of how it sounds, not quotes from real candidates. | | Weak answer | Strong answer | | --- | --- | --- | | Question 1, the Monday drop | "Users lost interest after the weekend." | "I would check whether the tracking changed on Friday, then look by platform. A drop on one platform points at a release." | | Question 4, the doubled total | "I would divide by two." | "A one-to-many join repeated each order. I would compare row counts and aggregate before joining." | | Question 7, the redesign | "Yes, conversion went up after launch." | "It might be the redesign. I would check traffic mix and the week before, and ask for a holdout." | *Table 2. Example weak and strong answers (illustrative, not from real candidates).* Listen for ownership. "The dashboard showed it" tells you about the tool. "I checked the row counts and found the join" tells you about the person. ## One follow-up chain shows where the real answer starts This is an illustration with sample data, not a real interview. The candidate is fictional. It shows the shape of what Tracy runs: an opening question from your list, the answer, and a follow-up built on it. It starts from question 4. > **Tracy:** A join doubles your revenue total. What happened? > > **Candidate (fictional):** Probably a one-to-many join. Each order repeated for each line item. > > **Tracy:** How would you prove that and not guess? > > **Candidate:** Count rows before and after the join, and count distinct order ids. If they differ, it is fan-out. > > **Tracy:** Tell me about the last time you found that in real work. > > **Candidate:** A returns table joined to orders. Refunds counted twice for two months until I compared row counts. The first answer is the textbook one. The last shows whether it ever happened. A Director at a top-tier early-stage VC says Tracy "sometimes brings up exactly those questions I might have forgotten to ask, which has been a pleasant surprise." ![Transcript card: join doubles revenue, candidate says one-to-many, Tracy asks how to prove it rather than guess.](/blog/images/data-analyst-interview-questions-1.png) *Tracy asks a follow-up built on the candidate's own answer. Your analytics lead sees whether the reasoning came from experience. Illustration with sample data* A follow-up built on the candidate's own answer shows whether they have done the work. ## Run the interview for every applicant without losing the follow-up A live screen for every applicant does not survive two hundred of them. A Director at a top-tier early-stage VC reviewed founder interviews and says: "I was effectively spending around six hours a day on screening. With Tracy, I can now review and assess those candidates in approximately 1.5 hours." - **Questions from the role.** Paste the analyst job description and Tracy drafts questions on metrics, SQL and communication that you edit or replace with your own. See the [product page](/product). - **A SQL or debugging task.** It sits in the same interview, with follow-ups on approach and trade-offs, so a join question ends with how they would check the row counts. - **CV check, as a setting.** With it on, Tracy can read the CV and ask about what it claims, so "built dashboards" becomes a question about one of them. See [how the CV check works](/blog/ai-interviewer-cv-check). - **No scheduling.** Each candidate opens a link at any hour, with no account; the join screen says the conversation is recorded and transcribed. See the [participant experience](/participant-experience). - **Evidence per result.** A recording, a timestamped transcript and a written assessment per question. See [AI interview reports](/blog/ai-interview-reports). Pricing is $1 per credit, one credit per started 20 minutes of an interview that produced a result, so a 34-minute interview costs $2. The first 3 results are free with no card, there are no seats, and a session with no answers is never charged. See [pricing](/pricing). Send us one role and we will build the interview plan in days. ![Recreated Tracy result report for a fictional analyst: recording with timestamped transcript, and three questions with what they said and quoted evidence.](/blog/images/data-analyst-interview-questions-2.png) *A recording and timestamped transcript sit behind every Tracy result. Your analytics lead reads each answer and its quote before deciding. Illustration with sample data* A reviewer can read what the candidate said to each question, beside the recording and transcript. ## Keep four checks with people 1. **Your own analyst.** Have one score five or six transcripts blind. If they disagree with the written assessments, fix the questions before inviting the next hundred. 2. **The query.** Tracy asks about approach and trade-offs, such as how they would check row counts. Whether the final SQL meets your standard is a reviewer's call. 3. **Integrity signals.** Tracy captures tab changes, gaze, face detection, delay before answers, fullscreen exits and external screens, and shows the evidence to a person. A signal is not proof. 4. **The decision.** A person decides from the evidence. For ten applicants, a call with one of your analysts is a reasonable choice; see [AI interview vs human interview](/blog/ai-interview-vs-human-interview). When two hundred apply, screening hours are the limit; see [how to screen hundreds of applicants](/blog/how-to-screen-hundreds-of-applicants). To try it, create a Tracy interview for this role from these questions and read the report on three applicants. ## Common questions ### What are good data analyst interview questions? Ask what they check first when a number moves, how they would debug a join, whether a result is causal, how a metric can mislead and how they explain it to a stakeholder. Keep the follow-up beside each. ### What SQL interview questions should I ask a data analyst? Ask about fan-out joins, finding customers who did one thing but not another, and window functions versus group by. Ask for the approach first and the syntax second. ### Where can candidates find data analyst interview questions and answers? Public lists such as the Reddit threads give candidates the questions, so expect rehearsed first answers. The follow-up is what separates them: ask for the last time it happened. ### Can an AI interviewer interview data analysts? Yes. Tracy asks questions in the candidate's field, can include a SQL or debugging task, follows up on approach and waits for the answer. Your team gets the recording, transcript and an assessment per question, and a person decides. --- Create a Tracy interview for this exact role from these questions, then read the recording and report; the button opens the Connect page. Or send us one role and we will build the interview plan, with a recording and report you can check. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=data-analyst-interview-questions&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=data-analyst-interview-questions&utm_content=demo) ## Sources - [Tracy AI Interviewer home page, customer proof](https://tracyinterview.ai/), retrieved 9 Nov 2026. - [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer Connect page](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. # DevOps Engineer Interview Questions and Follow-Ups for Engineering Leads > DevOps engineer interview questions for hiring teams: 15 questions with follow-ups, what strong answers sound like, and how to run them at volume. - URL: https://tracyinterview.ai/blog/devops-engineer-interview-questions - Published: 2026-11-09 - Author: Tracy team > **In short** **Tracy AI Interviewer** interviews every DevOps applicant on incidents, releases and trade-offs, and follows up on what each person says, so you see how they act under pressure. > > Use the 15 DevOps engineer interview questions below, grouped by what they test, each with a follow-up and what to listen for. > > Add a debugging task to the same interview. Every result keeps a recording, a timestamped transcript and a written assessment per question. > > Next step: paste these questions into a Tracy interview for one open role. Twenty DevOps applicants list Kubernetes, Terraform and CI/CD. The list says what they have touched. It does not say whether they rolled back at minute eight or kept digging. DevOps engineer interview questions for a first round have to test judgement under pressure, and the first answer is usually a recited method. These are for the engineering lead asking the questions, not for candidates preparing. Each has a follow-up and a note on what a strong answer contains. Tracy AI Interviewer runs the part that does not scale. She asks each applicant questions from your job description, follows up on what they just said and waits for the end of the thought. A recording, a timestamped transcript and a written assessment per question go to your team, who decide. For developers, see [software engineer interview questions](/blog/software-engineer-interview-questions). ## Fifteen DevOps engineer interview questions, grouped by what they test ### Group 1: incidents and on-call | Ask | Follow up with | Listen for | | --- | --- | --- | | 1. Walk me through the last production incident you led. | What was the first signal, and what did you do in the first ten minutes? | A timeline. Separates mitigation from root cause. Names who they told. | | 2. An alert fires at 3 a.m. and the dashboard looks normal. What do you do? | When do you wake someone? | Checks whether the alert itself is wrong. Has a rule for escalating. No heroics. | | 3. Tell me about an incident review that changed something. | How did you know the change worked? | A named change and a check afterwards. Blames the system, not a person. | ### Group 2: deployment and release | Ask | Follow up with | Listen for | | --- | --- | --- | | 4. How does a change get from merge to production where you work? | Where does it most often go wrong? | Names the stages, the gates and the rollback. Honest about the weak step. | | 5. A release needs a database migration. How do you ship it? | How would you roll it back? | Backward-compatible change, migration shipped apart from the code. Knows some data cannot be rolled back. | | 6. When do you stop a release? | Describe one time you did. | Criteria stated in advance and a real case. | ### Group 3: infrastructure as code and environments | Ask | Follow up with | Listen for | | --- | --- | --- | | 7. Two engineers apply changes to the same environment at once. What goes wrong? | What did you do the last time the code and the real environment disagreed? | State locking, review before apply, drift detection. A real case. | | 8. How do you keep secrets out of repositories and logs? | How do you rotate one in a live system? | A secret store, short-lived credentials, rotation without downtime. | | 9. A service needs a staging environment that matches production. What do you copy and what do you leave out? | What breaks if staging is too close to production? | Copies configuration, not customer data. Weighs cost against parity. | ### Group 4: reliability and observability | Ask | Follow up with | Listen for | | --- | --- | --- | | 10. How would you set a service level objective for an API? | What happens when the error budget is spent? | A user-facing measure, a target with a reason, a policy that changes the work. | | 11. A service's latency doubled after a release. How do you find the cause? | What would you rule out first? | Compares before and after. Checks deploy, dependency and load. Narrows quickly. | | 12. What do you alert on, and what do you leave out? | Tell me about an alert you deleted. | Alerts on symptoms and pages only when someone must act. | ### Group 5: cost, security and working with developers | Ask | Follow up with | Listen for | | --- | --- | --- | | 13. Cloud spend rose 40 percent in a month. Where do you look? | Who do you talk to before you change anything? | Ownership tags, the largest line items, a conversation with the service owner. | | 14. A developer asks for admin access to production to fix something fast. | What would you offer instead? | Time-boxed, audited access with the reason explained. Does not just say no. | | 15. Tell me about a tool or process you removed because it cost more than it gave. | What did the team lose? | An honest trade-off and a named owner of the decision. | *Table 1. The 15 questions, each with a follow-up and what to listen for (written by us).* ## A strong answer names what was done, and a weak one names a tool The difference shows up in the second answer. These pairs are written by us as examples of how it sounds, not quotes from real candidates. | | Weak answer | Strong answer | | --- | --- | --- | | Question 1, a production incident | "We found the bug and fixed it." | "I rolled back first to stop the damage. The cause took two days, and the review moved that check into the pipeline." | | Question 5, a migration | "I'd run it in the deploy script." | "Migration first, backward compatible, then the code that needs it. If the data cannot roll back, I say so before we start." | | Question 14, admin access | "No. Open a ticket." | "I'd grant it for an hour with audit logging, and fix the reason they needed it." | *Table 2. Example weak and strong answers (illustrative, not from real candidates).* Listen for what the person did. "We use blue-green deploys" tells you about the platform. "I rolled back at minute eight because the error rate had doubled" tells you about the engineer. ## One follow-up chain shows where the real answer starts This is an illustration with sample data, not a real interview. The candidate is fictional. It starts from question 11 and shows the shape of what Tracy runs. > **Tracy:** A service's latency doubled after a release. How do you find the cause? > > **Candidate (fictional):** I compare before and after: the deploy, traffic and dependencies. > > **Tracy:** What would you rule out first? > > **Candidate:** The release. Did latency move at the deploy minute, or later when traffic peaked? > > **Tracy:** Tell me about the last time that was the answer. > > **Candidate:** A connection pool setting. Latency rose 40 minutes after deploy, at peak traffic. I rolled back the flag first, then fixed the pool size. The first answer is a method anyone can recite. The third shows whether it was used on a live system. For a hands-on task, Tracy can add a debugging exercise to the same interview, with follow-ups on approach and trade-offs. ![Transcript card: latency doubled after a release, a candidate answer, and Tracy asking what they would rule out first.](/blog/images/devops-engineer-interview-questions-1.png) *Tracy asks a follow-up built on the candidate's own answer. You see whether an engineer has handled the situation. Illustration with sample data* A follow-up built on the candidate's own answer shows whether the method was ever used live. ## Run the interview for every applicant without losing the follow-up An engineering lead cannot hold a live call with every DevOps applicant, and a quiz does not show how someone handles a bad night. - **Questions from your job description.** Tracy drafts questions you can edit, from technical to behavioural, or use your own. See the [product page](/product). - **A task inside the interview.** Add a coding or debugging task, with follow-ups on approach and trade-offs. See [technical interviews](/use-cases/technical-interviews) and [AI technical interviews](/blog/ai-technical-interviews). - **Start on click.** Each applicant opens a link in a browser, with no account and no calendar, at any hour, and is told on the join screen that the conversation is recorded and transcribed. - **A result you can check.** A recording, a timestamped transcript and a written assessment per question sit behind every result. - **Scale.** Tracy has completed 600K+ interviews for 300+ clients, with 99% interview completion. It is $1 per credit, one credit per started 20 minutes of an interview that produced a result, so a 34-minute interview costs $2. The first 3 results are free with no card, and there are no seats. See [pricing](/pricing). Customisation takes days: send us one role and we will build the interview plan. ![Recreated Tracy result report for a fictional DevOps engineer: recording with timestamped transcript, and three questions with what they said and quoted evidence.](/blog/images/devops-engineer-interview-questions-2.png) *A recording and timestamped transcript sit behind every Tracy result. Your lead reads each answer and its quote before deciding. Illustration with sample data* A reviewer can read what the candidate said to each question, beside the recording and transcript. ## Keep three checks with people 1. **A senior engineer reads five or six transcripts blind.** If they disagree with the written assessments, change the questions before you invite more people. 2. **Hands-on proof for finalists.** An interview shows how someone reasons about incidents. A live exercise on your own stack shows the rest. 3. **The decision.** A person decides from the evidence. For three applicants, a call with your best engineer is a reasonable choice; see [AI interview vs human interview](/blog/ai-interview-vs-human-interview). The case for Tracy starts when applicants outnumber your engineers' hours. Pilot on one open role: 20 applicants, one reader. Track completion, hours from application to result and reader disagreement. See [AI interview reports](/blog/ai-interview-reports). ## Common questions ### What are good DevOps engineer interview questions? Questions about things the person has run: an incident, a release, a migration, an alert, a bill. Ask one or two from each group above and keep the follow-up beside each. ### What are DevOps engineer interview questions for experienced candidates? Use the same groups and ask for the decision behind each answer: what they traded away, what they would change, who they told. Seniority should show in the follow-up, not in harder trivia. ### How do you assess answers to DevOps engineer interview questions? Write down what a strong answer contains before the first interview. Look for a timeline, a named trade-off and a check made afterwards. Tool names alone say little. ### Can an AI interviewer interview DevOps engineers? Yes. Tracy asks questions from your job description, follows up on what each person said and can include a debugging task. A person reads the recording, transcript and assessments and decides. --- Create a Tracy interview for this exact role from these questions, then read the recording and report. Or send us one role and we will build the interview plan, with a recording and report you can check. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=devops-engineer-interview-questions&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=devops-engineer-interview-questions&utm_content=demo) ## Sources - [Tracy AI Interviewer home page](https://tracyinterview.ai/), retrieved 9 Nov 2026. - [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer Connect page](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. # Frontend Developer Interview Questions for Hiring Teams > Frontend developer interview questions for hiring teams: 15 questions with follow-ups, strong and weak answers, and how to run them at volume. - URL: https://tracyinterview.ai/blog/frontend-developer-interview-questions - Published: 2026-11-09 - Author: Tracy team > **In short** **Tracy AI Interviewer** interviews every frontend applicant on browser and performance trade-offs and follows up on each answer, so your engineers read how candidates reason, not a list of frameworks. > > Use the 15 frontend developer interview questions below, grouped by what they test, each with a follow-up and what to listen for. > > A recording, a timestamped transcript and a written assessment per question sit behind every result, and your team decides. > > Next step: paste five of these questions into a Tracy interview. A frontend opening draws a hundred applicants whose CVs list React, TypeScript and "performance optimisation". Two of your engineers can interview them between sprints. The CV says which tools someone has touched. It does not say whether they would measure a slow page before rewriting it. Frontend developer interview questions have to test that gap. These are written for the hiring manager or lead who runs the interview. They ask for judgement about browsers, state, accessibility and delivery, and each one has a follow-up and a note on what a strong answer contains. They suit mid-level and senior roles; for a five-year candidate, ask the follow-up every time. Candidates can use them to prepare too. Tracy AI Interviewer runs the part that does not scale. She asks each applicant in their own field, follows up on what they just said and waits for the end of the thought, with a coding task inside the same interview when you want one. For the wider engineering version, see [software engineer interview questions](/blog/software-engineer-interview-questions), and for the process, [technical interviews](/use-cases/technical-interviews). ## Fifteen frontend developer interview questions, grouped by what they test For each group, pick one or two questions. Ask the follow-up only after the first answer, and write down what a strong answer contains before the first candidate joins. Questions about experience should be answered with a real project, so ask for the project. ### Group 1: browsers and performance | Ask | Follow up with | Listen for | | --- | --- | --- | | 1. A page feels slow on a mid-range phone. Where do you start? | What would you measure before changing anything? | Measures first with a profiler trace or field data, and separates network from main-thread time. Names a finding, not a tool list. | | 2. What happens between typing a URL and seeing a page? | Which step have you actually made faster? | The order is roughly right and the depth sits where they have worked: caching, render-blocking resources, hydration. | | 3. A list of 5,000 rows scrolls badly. What are your options? | What does each option cost you? | Virtualisation, pagination, memoisation. Names the costs: find-in-page, screen readers, scroll position. | ### Group 2: JavaScript and state | Ask | Follow up with | Listen for | | --- | --- | --- | | 4. Describe a bug you caused with closures, async code or stale state. | How did you reproduce it? | A real case with reproduction steps, the fix and a test added afterwards. | | 5. Where should state live in an app you own? | When did you last move it, and why? | Local, shared and server state kept apart. Avoids "put everything in a store". | | 6. Two requests return out of order and the screen shows old data. What now? | How would you test that? | Cancels or ignores the stale response, uses request identifiers, tests with delayed mocks. | ### Group 3: CSS and accessibility | Ask | Follow up with | Listen for | | --- | --- | --- | | 7. How do you build a modal that works with a keyboard and a screen reader? | What breaks first in a custom one? | Focus moves in and returns, a label, Escape closes it. Prefers the native element where it fits. | | 8. A layout breaks at one width only. How do you find out why? | What stops it coming back? | Reads computed styles, prefers flexible sizing to more breakpoints, adds a visual or component test. | | 9. When do you reach for JavaScript instead of CSS? | A time you removed script for CSS? | A concrete case. Knows what CSS now does alone and what still needs script. | ### Group 4: frameworks and tooling | Ask | Follow up with | Listen for | | --- | --- | --- | | 10. Which framework feature do you avoid, and why? | When did avoiding it cost you? | A reasoned rule with an exception, not a taste. | | 11. How do you decide whether a dependency is worth adding? | What did you last remove? | Bundle size, maintenance, licence and the alternative of writing ten lines. | | 12. What should a component test cover? | What do you leave untested? | Behaviour a user sees, not implementation details. Names what is too costly to test. | ### Group 5: working with others | Ask | Follow up with | Listen for | | --- | --- | --- | | 13. A design cannot be built as drawn. What do you do? | How do you tell the designer? | Proposes an option with a trade-off and says it early, with a screenshot or a prototype. | | 14. You review a pull request with 900 changed lines. How? | What do you ask the author to do? | Asks to split it, reads tests first, comments on risk before style. | | 15. Tell me about something you shipped and later regretted. | What would have caught it earlier? | Owns the decision, names the missed signal, describes a check they now run. | *Table 1. The 15 questions, each with a follow-up and what to listen for (written by us).* ## A strong answer measures first, and a weak one names a tool The difference shows up in the second answer. These pairs are written by us as examples of how it sounds, not quotes from real candidates. | | Weak answer | Strong answer | | --- | --- | --- | | Question 1, the slow page | "I would add lazy loading and memoise everything." | "I would record a trace on a mid-range phone first. If the main thread is busy I look at scripts, if not at the network." | | Question 6, out-of-order requests | "I would add a loading spinner." | "The old response arrives last and overwrites the new one. I would cancel the old request and test it with a delayed mock." | | Question 7, the modal | "I would use a library for it." | "Focus goes in, comes back to the button, Escape closes it. I would check it with the keyboard before calling it done." | *Table 2. Example weak and strong answers (illustrative, not from real candidates).* Listen for ownership. "We used a library" tells you about the team. "I tested it with the keyboard and found the focus bug" tells you about the person. ## One follow-up chain shows where the real answer starts This is an illustration with sample data, not a real interview. The candidate is fictional. It shows the shape of what Tracy runs: an opening question from your list, the answer, and a follow-up built on it. It starts from question 1. > **Tracy:** A page feels slow on a mid-range phone. Where do you start? > > **Candidate (fictional):** With a trace on the phone. I want to know if it is the network or the main thread. > > **Tracy:** The trace shows the main thread busy for two seconds. What do you look at next? > > **Candidate:** The longest tasks. Last time it was a date library parsed on load, so I split it out and measured again. The first paint moved by a second. > > **Tracy:** What did you do to keep it from coming back? > > **Candidate:** A budget on the bundle in the build, so a pull request fails if it grows. The first answer is the textbook one. The fourth shows whether it ever happened. A Director at a top-tier early-stage VC says Tracy "sometimes brings up exactly those questions I might have forgotten to ask, which has been a pleasant surprise." ![Transcript card: slow page on a phone, candidate says trace first, Tracy asks what they look at next.](/blog/images/frontend-developer-interview-questions-1.png) *Tracy asks a follow-up built on the candidate's own answer. Your engineers see whether the work was done, not recited. Illustration with sample data* A follow-up built on the candidate's own answer shows whether they have done the work. ## Run the interview for every applicant without losing the follow-up A live call for every applicant does not survive a hundred of them. - **Questions from the role.** Paste the frontend job description and Tracy drafts questions on browsers, state and accessibility that you edit or replace with your own. See the [product page](/product). - **A coding or debugging task.** It sits in the same interview, with follow-ups on approach and trade-offs, for example why a search box should wait before it calls the server. - **CV check, as a setting.** With it on, Tracy can read the CV and ask about what it claims, so "performance optimisation" becomes a question about the trace, the bundle or the budget. See [how the CV check works](/blog/ai-interviewer-cv-check). - **No scheduling.** Each candidate opens a link, with no account, and is told on the join screen that the conversation is recorded and transcribed. See the [participant experience](/participant-experience). - **Evidence per result.** A recording, a timestamped transcript and a written assessment per question, so your engineer can check a keyboard-focus answer against what was said. See [AI interview reports](/blog/ai-interview-reports). Pricing is $1 per credit, one credit per started 20 minutes of an interview that produced a result, so a 34-minute interview costs $2. The first 3 results are free with no card, there are no seats, and a session with no answers is never charged. See [pricing](/pricing). Send us one role and we will build the interview plan in days. ![Recreated Tracy result report for a fictional frontend candidate: a summary, the recording beside a timestamped transcript, and a table of three questions with what the candidate said and quoted evidence.](/blog/images/frontend-developer-interview-questions-2.png) *A recording and timestamped transcript sit behind every Tracy result. Your engineer reads what the candidate said to each question, with quotes, before deciding. Illustration with sample data* A reviewer can open the recording and transcript, and read what the candidate said to each question, with quotes. ## Keep four checks with people 1. **Your own engineer.** Have one score five or six transcripts blind. If they disagree with the written assessments, fix the questions before inviting the next hundred. 2. **The component.** Tracy asks why the candidate built it that way and what it costs. Whether it meets your standard is a reviewer's call. 3. **Integrity signals.** Tracy captures tab changes, gaze, face detection, delay before answers, fullscreen exits and external screens, and shows the evidence to a person. A signal is not proof. See [AI interview cheating detection](/blog/ai-interview-cheating-detection). 4. **The decision.** A person decides from the evidence. For ten applicants, a call with one of your engineers is a reasonable choice; see [AI interview vs human interview](/blog/ai-interview-vs-human-interview). When a hundred apply, interview hours are the limit. Run one open frontend role and read the recording and report for three applicants before you widen it. ## Common questions ### What are good frontend developer interview questions? Ask about measuring a slow page, where state lives, keyboard and screen-reader behaviour, dependencies and how they review code. Keep the follow-up beside each, and ask for a real project. ### What are frontend developer interview questions for 5 years of experience? Ask the follow-ups every time. A five-year developer should name a project, the measurement, the fix and what they changed afterwards. Questions 1, 5, 10 and 14 separate that depth from a listed skill. ### How do you test frontend skills without a take-home? Put a short coding or debugging task inside the interview and ask about approach and trade-offs. Tracy can run that task in the same interview, with follow-ups on the candidate's own reasoning. ### Can an AI interviewer interview frontend developers? Yes. Tracy asks questions in the candidate's field, follows up on what they said and waits for the answer. Your team gets the recording, transcript and an assessment per question, and a person decides. --- Create a Tracy interview for this exact role from these questions, then read the recording and report; the button opens the Connect page. Or send us one role and we will build the interview plan, with a recording and report you can check. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=frontend-developer-interview-questions&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=frontend-developer-interview-questions&utm_content=demo) ## Sources - [Tracy AI Interviewer home page, customer proof](https://tracyinterview.ai/), retrieved 9 Nov 2026. - [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer Connect page](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. # Interview Questions for Legal Experts on AI Training Projects > Legal expert interview questions for AI training projects: 19 questions on citations, jurisdiction and drafting, with follow-ups and answers to listen for. - URL: https://tracyinterview.ai/blog/legal-expert-interview-questions - Published: 2026-11-09 - Author: Tracy team > **In short** **Tracy AI Interviewer** asks each legal expert questions in their own practice area, follows up on what they said and waits for the answer, so you see how they check a source and not just what they remember. > > The 19 questions below test how an expert applies a rule, verifies authority, handles jurisdiction and dates, reads drafting closely, and stays inside the limits of the job. > > Every result comes with a recording and a timestamped transcript, plus integrity signals and their evidence, for a person to review. Bar admission and other credentials are checked separately. Three weeks into a legal-evaluation project, a reviewer finds that an approved expert cited a case that does not say what the expert said it says. The expert had a good CV and the right practice area. Nobody had asked how they check a source. Legal expert interview questions for AI training projects have to ask exactly that. The job is to read a model's answer to a legal question, find what is wrong, unsupported or out of date, and explain why so a second reviewer would agree. Job seekers can use the questions to prepare, and should answer the follow-up too. Three limits. The questions test legal reasoning for review work, are not legal advice and state no law of any place. And they do not confirm admission to practise: Tracy does not verify a licence or a credential, so check bar admission with the body that holds the record. For the generalist version of this job, see our guide to [AI trainer interview questions](/blog/ai-trainer-interview-questions). For the rest of the process, see [expert vetting](/use-cases/expert-vetting). ## Legal review work depends on checking sources, so ask about the check The failure that matters most has a name in the research. Stanford researchers who tested AI legal research tools on a pre-registered set of over 200 open-ended legal queries separated two kinds of error. A response can be incorrect, or it can be "misgrounded": the law is described correctly but the source it cites does not support the claim ([source](https://hai.stanford.edu/news/ai-trial-legal-models-hallucinate-1-out-6-or-more-benchmarking-queries), 23 May 2024). A reviewer who only checks whether the statement of law is right will pass a misgrounded answer, which is why the second group below is the longest. The questions test applying a rule to facts, authority and citations, jurisdiction and time, reading and drafting, the limits of the job, and working method. A lawyer on your team should adapt them to the project's practice areas. ## Nineteen legal expert interview questions, grouped by what they test Choose one or two from each group. Ask the follow-up after the first answer, decide what a strong answer contains before the first expert joins, and give every expert the same opening questions. ### Group 1: applying a rule to facts | Ask | Follow up with | Listen for | | --- | --- | --- | | 1. A tenant stops paying rent after the landlord ignores a repair request. Pick a jurisdiction and talk me through how you would think about it. | Which single fact, if it changed, would change your answer? | Names the jurisdiction first. Separates facts given from facts missing. Says what they would confirm in the source. | | 2. Describe a fact pattern in your practice area where the obvious answer is wrong. | What did a less experienced colleague get wrong about it? | A specific trap. Says what the colleague skipped, not just that they erred. | | 3. How do you explain to a non-lawyer why "it depends" is the honest answer? | Where does "it depends" stop being an answer? | Names what it depends on. Knows when a hedge avoids the question. | ### Group 2: authority and citations | Ask | Follow up with | Listen for | | --- | --- | --- | | 4. A model cites a case to support its answer. Walk me through verifying it. | The case exists but does not say that. What do you do? | Reads the passage, not the headnote. Treats a real case with an unsupported proposition as an error. | | 5. How do you tell binding from persuasive authority in your jurisdiction? | Tell me about a time that distinction changed an answer. | Explains it in local terms. Gives a case where a model leaned on the wrong kind. | | 6. How do you check that a statute or regulation is the current version? | What do you do when you cannot reach the source? | Names where they look and checks the effective date. Marks the answer unverified, not guessed. | | 7. What makes you distrust a citation at first glance? | Which sign has misled you? | Concrete tells: a court that does not hear that case type, a year that does not fit, a quote that reads too neatly. Admits one that fooled them. | ### Group 3: jurisdiction and time | Ask | Follow up with | Listen for | | --- | --- | --- | | 8. A question never says where the dispute is. What do you do? | When is it fine to answer in general terms? | Asks, or states an assumption aloud. Answers generally only when the point holds everywhere. | | 9. Tell me about an area of law that changed recently in your practice. How would an earlier-trained model go wrong? | How would you score an answer that was right last year? | A real change and its date. Marks the answer out of date, with a note for the next reviewer. | | 10. A question assumes a rule that does not exist in your jurisdiction. How do you respond? | Which scores higher, a model that corrects the premise or one that plays along? | Corrects the premise first. Scores pushback higher and can explain the risk of playing along. | ### Group 4: reading and drafting | Ask | Follow up with | Listen for | | --- | --- | --- | | 11. Take this phrase: "the supplier will use reasonable efforts to deliver on time." How would you read it? | How would you redraft it for the supplier, then for the buyer? | Reads it in context and says what each side might argue. Offers two drafts for two clients. | | 12. Summarise a contract clause in two sentences for a client who is not a lawyer. | What did you leave out, and why? | Plain words that keep the obligation and the trigger. Names the omission and why it was safe. | | 13. You find two clauses in the same agreement that conflict. What do you do? | What if both parties already acted on one of them? | Looks for an order-of-precedence clause first. Raises conduct as a factor. | ### Group 5: confidentiality and limits | Ask | Follow up with | Listen for | | --- | --- | --- | | 14. What would you never share from a client matter in a task, a note or a chat with us? | How do you build a realistic scenario, then? | Builds an invented scenario from the shape of a matter. Removes names and details unasked. | | 15. Where is the line between legal information and legal advice, and why does it matter when you judge a model? | Give me an answer you would mark down for crossing it. | A practical line, such as telling one person what to do about their own dispute. A concrete example. | | 16. When should a model decline to answer or refer the person to a lawyer? | When would that be unhelpful? | Sees both failures. Refers when facts or stakes call for it, answers when it can help. | ### Group 6: working method | Ask | Follow up with | Listen for | | --- | --- | --- | | 17. You have ten minutes to grade a long answer. What do you do first? | What do you skip, and what makes you spend longer? | A fixed order: claim, authority, jurisdiction, date. Says which findings stop the clock. | | 18. You and another reviewer disagree about whether an answer is correct. How do you resolve it? | When do you hold your position? | Goes to the source together. Holds on text they can point to, moves on taste. | | 19. How many hours a week, in which time zone and for how long can you commit? What gets in the way? | What do you do when a client deadline lands in your project week? | Specific numbers and a plan for the bad week. Says early if it will not fit. | For a short call, take 1, 4, 6, 10, 15 and 18, and ask question 19 of everyone. For other practice areas, keep the groups and change the facts: an employment lawyer and a tax specialist should meet the same structure with different cases. ## A strong answer shows the work, and a weak one repeats the rule Weak answers are often accurate summaries of the law with no way to check them. These pairs are written by us as examples, not quotes from real experts. | Question | Weak answer | Strong answer | | --- | --- | --- | | 1. The unpaid rent | "The tenant can withhold rent if the landlord doesn't fix things." | "It depends on where the property is. I would ask that first, then whether the tenant gave written notice, and read the statute for that place before I said anything firm." | | 11. "Reasonable efforts" | "It means they have to try hard." | "It is ambiguous. A buyer will say a careful supplier's standard, a supplier will say it stops short of unreasonable cost. I'd check what the agreement says about cost and deadlines first." | Notice who owns the claim. "The law says" is a quotation. "I read it, and here is the date" is a person who did the work. ![A three-row transcript card. Tracy asks how the expert verifies a case a model cites. The expert says they find the case and read the passage the answer leans on. Tracy asks what they score if the case exists but the passage says something else.](/blog/images/legal-expert-interview-questions-1.png) *Illustration with sample data, not a real interview. A fictional expert answers a citation-check question, and Tracy follows up on it. Source: Tracy visual library, qa-followup template. Sample content written by the Qode Product Team.* The follow-up asks what the expert does when the citation exists but does not support the claim, which a prepared expert cannot recite. ## Ask every applicant the same questions without booking a single call A lawyer on your own team cannot interview every applicant when a project lands with a deadline. Tracy runs the interview for each expert, and each setting below is on the [product page](/product). - **Brief or your own questions.** Describe the role and Tracy drafts questions you can edit, or paste the 19 above. Style runs from quick to deep. - **CV check, as a setting.** With it on, Tracy can read the CV and ask about what it claims. - **Follow-ups that wait.** Tracy asks for the missing detail in the expert's field and waits until the thought is finished, which matters for questions 4 and 13. A director at an accelerator residency programme said: "Tracy sometimes brings up exactly those questions I might have forgotten to ask, which has been a pleasant surprise." - **Start on click.** Experts open a browser link with no account and no calendar, at any hour. The join screen says the conversation is recorded and transcribed. Candidates rate Tracy 92 on NPS. - **A result a reviewer can check.** A recording, a timestamped transcript and a written assessment for each question sit behind every result. - **Integrity signals, with evidence.** Tracy captures tab changes, gaze, face detection, delay before answers, fullscreen exits and external screens, and shows the evidence to a human reviewer. A signal is not proof, and your team decides. See [AI interview cheating detection](/blog/ai-interview-cheating-detection). Pricing is $1 per credit, one credit per started 20 minutes of an interview that produced a result, so a 25-minute interview takes two credits. The first 3 results are free with no card, there are no seats and an empty session is never charged. See the [pricing page](/pricing), or the [Connect page](/connect) for the API and MCP server. ![A recreated Tracy result report for a fictional legal reviewer. The recording sits beside a timestamped transcript, including the 5:48 answer about opening the case and reading the passage. Below, four fields are marked Stated, each with a quoted answer.](/blog/images/legal-expert-interview-questions-2.png) *Illustration, not a screenshot. A recreation of a Tracy result report for a fictional legal reviewer, with invented data. Source: Tracy visual library, result-traced screen recreation. All names, quotes and times are invented.* Each stated point links to the words the expert used, with the recording and transcript beside it, so a lab client can see why the expert was approved. ## Verify admission yourself, and keep a lawyer in the loop 1. **Admission and standing.** Check whether someone is admitted to practise, and in good standing, with the body that keeps the record. Tracy does not. 2. **A lawyer on your side.** Have one of your lawyers read five or six transcripts and score them blind. If they disagree with the written assessments on citation checks, change the questions before inviting more experts. 3. **The decision.** A person approves or declines from the evidence. If you approve a handful of attorneys a month, one conversation with your own lawyer is likely the better use of time ([AI interview vs human interview](/blog/ai-interview-vs-human-interview)). The case for Tracy starts when applicants outnumber your reviewers' hours. Pilot on one practice area in one jurisdiction with 20 applicants, and read [AI interview reports](/blog/ai-interview-reports) first. ## Common questions ### What should you ask a legal expert in an AI training interview? Ask how they apply a rule to facts, how they verify a citation, how they handle jurisdiction and dates, how they read a clause and where they stop. Choose one or two questions from each group above and keep the follow-up beside each one. ### How do you test legal reasoning without giving legal advice? Ask about method: what they check first, how they verify a source and what they would mark down in a model's answer. No question needs a statement of the law. ### Can an AI interviewer interview legal experts? Yes. Tracy AI Interviewer asks questions in the expert's own practice area, follows up on what they said and waits for the answer. Your team gets the recording, a timestamped transcript and a written assessment for each question, and a person decides. ### Does Tracy check bar admission or credentials? No. Tracy does not verify admission to practise, a degree or any other credential. Check these separately with the body that holds the record. ### How long should a legal expert interview be? Twenty to thirty minutes covers six questions with their follow-ups. That is our rule of thumb, not a researched figure. A 25-minute interview takes two credits, which is $2. --- Run these questions on your own applicants. The first 3 results are free, with no card. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=legal-expert-interview-questions&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=legal-expert-interview-questions&utm_content=demo) ## Sources - [Stanford HAI, 23 May 2024 (updated 30 May 2024): definition of a misgrounded legal answer](https://hai.stanford.edu/news/ai-trial-legal-models-hallucinate-1-out-6-or-more-benchmarking-queries), retrieved 9 Nov 2026. - [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 5 Nov 2026. - [Tracy AI Interviewer Connect page](https://tracyinterview.ai/connect), retrieved 5 Nov 2026. - [Tracy AI Interviewer home page, customer proof](https://tracyinterview.ai/), retrieved 9 Nov 2026. # Interview Questions for Medical Experts on AI Training Projects > Medical expert interview questions for AI training projects: 18 questions with follow-ups, strong and weak answers, and how to run them at volume. - URL: https://tracyinterview.ai/blog/medical-expert-interview-questions - Published: 2026-11-09 - Author: Tracy team > **In short** **Tracy AI Interviewer** asks each medical expert questions in their own specialty, follows up on what they said and waits for the answer, so you see how they reason and not only what they know. > > Use the 18 questions below, grouped by what they test: reasoning with missing information, reviewing a model's answer, scope and safety, rubric writing, specialty depth and project fit. > > Every result keeps a recording and a timestamped transcript, with integrity signals and their evidence, for a person to review. Licences and credentials are checked separately. A clinical-review project lands on a Monday and hundreds of applicants have CVs that name the right specialty. A CV says what someone trained in. It does not say whether they can explain why a model's answer is unsafe, or whether the person on the call is the person who will do the work. Medical expert interview questions for AI training projects have to test that gap. These ask for clinical reasoning applied to review work: reading a model's answer, writing the criteria for judging it, and knowing when to stop. Each has a follow-up and a note on what a strong answer contains. Job seekers can use them to prepare, and should answer the follow-up too, because that is where the interview starts. Two limits. The questions test reasoning for review work, give no medical advice and are not a clinical standard. And they do not check a licence: Tracy does not verify a licence or a degree, so check credentials separately against the register in the expert's country. For the generalist version of this job, see our guide to [AI trainer interview questions](/blog/ai-trainer-interview-questions). For the rest of the process, see [expert vetting](/use-cases/expert-vetting). ## The work asks for judgement on a model's answer, not only medical knowledge A public healthcare benchmark, HealthBench, shows the work at scale. Its preprint, submitted in May 2025, says responses are evaluated with conversation-specific rubrics created by 262 physicians ([source](https://arxiv.org/abs/2505.08775)). Specialty matters as well: a cardiologist and a dermatologist should not get the same case question. The questions test reasoning with missing information, reviewing a model's answer, scope and safety, rubric writing and agreement, specialty depth and currency, and project fit (hours, time zone, steadiness over a long batch, care with patient information). A clinician on your team should adjust each one to the specialty and the task. ## Eighteen medical expert interview questions, grouped by what they test Pick one or two questions from each group. Ask the follow-up after you hear the first answer, write down what a strong answer contains before the first expert joins, and give every expert the same opening questions. ### Group 1: reasoning with missing information | Ask | Follow up with | Listen for | | --- | --- | --- | | 1. A patient says their chest has felt tight for two days. Before you say anything useful, what do you want to know? | Which answer would change your next step the most? | Asks about onset, what it feels like, context and who the patient is. Orders the questions by impact. Says what a text conversation cannot settle. | | 2. Tell me about a case where your first impression was wrong. What changed your mind? | What would you have needed to see earlier? | A real case with a turning point. Owns the miss. Names a check they now do. | | 3. Two reasonable clinicians manage the same kind of case differently. Describe one in your field. | What should a model say about that disagreement? | Separates real uncertainty from error. Wants options and reasons, not an invented consensus. | ### Group 2: reviewing a model's answer | Ask | Follow up with | Listen for | | --- | --- | --- | | 4. A model told a patient: "This is usually harmless and will pass by itself." What do you check first? | What would make you reject it even if every word is true? | Looks at what is missing: when to seek care, who the answer is for. Names a reason to reject that is not a factual error. | | 5. What is the difference between an answer that is wrong and one that is unsafe? | Give me an example from your own reviewing. | Draws a defensible line. Knows a correct answer can be unsafe. Real example, patient details removed. | | 6. How do you grade an answer that is correct but written for a clinician when a patient asked? | Would your score change if the asker were a nurse? | Grades against the asker, not a textbook. Says what changes with the reader. | | 7. A model says "current guidelines recommend" something. How do you check it? | Which version, and for which country? | Opens the source and reads the sentence. Checks date, population and strength of wording. | ### Group 3: scope, safety and patient information | Ask | Follow up with | Listen for | | --- | --- | --- | | 8. When should an answer tell the person to see a clinician instead of answering? | Give a case where that advice would be too cautious. | Sees both failures. A referral can be the useful answer or a way of saying nothing. | | 9. What do you do when a task asks you to judge something outside your specialty? | How would your lead know the flag was fair? | Flags it, gives a one-sentence reason, offers a nearby task. Does not guess to keep the queue moving. | | 10. What would you never put into a task, a note or a chat with us about a patient? | How do you write a rubric example, then? | Builds an invented case. Removes identifiers without being asked. | ### Group 4: rubrics and agreement with other reviewers | Ask | Follow up with | Listen for | | --- | --- | --- | | 11. Pick a question patients often ask in your field. Write three criteria for judging an answer. | Which of your three could two reviewers read differently? | Criteria scored yes or no. Notices their own vague wording. | | 12. A rubric line says "recommends appropriate follow-up." What is wrong with it? | Rewrite it so a second reviewer scores it the same way. | Says "appropriate" hides the judgement. Splits it into a trigger and a time frame. | | 13. You and another reviewer score the same answer far apart. What do you do? | When do you hold your score? | Compares reasons, not numbers. Holds on a fact they can show, moves on taste. | ### Group 5: specialty depth and currency | Ask | Follow up with | Listen for | | --- | --- | --- | | 14. What is a common belief in your specialty that is out of date or contested? | How would you spot a model repeating it? | Something specific, with how settled it is. Describes the tell in a model's wording. | | 15. How do you keep up with your field, and what did you last change because of it? | How did you judge that source? | A named source and a concrete change. | | 16. Where does practice differ between countries or health systems in your field? | How would you mark an answer that fits one system and not another? | A real first-hand difference. Marks an answer that does not say where it applies. | ### Group 6: working on the project | Ask | Follow up with | Listen for | | --- | --- | --- | | 17. Walk me through twenty reviews in a two-hour block. How do you keep your standard steady? | How would you notice it slipping? | A routine with breaks or spot checks. Knows when they get faster and worse. | | 18. How many hours a week, in which time zone and for how long can you commit? What would interrupt it? | What happens when a clinic shift runs over? | Specific numbers and a plan for the bad week. Says early if it will not work. | For a six-question call, take 2, 4, 7, 9, 12 and 16, and ask question 18 of everyone. These are also the medical specialist interview questions to adapt: keep the group, change the case to the specialty. ## A strong answer names what it would check, and a weak one stops at the fact A weak expert is often right and runs out of reasons when the question moves from a fact to a decision. These pairs are written by us as examples, not quotes from real experts. | Question | Weak answer | Strong answer | | --- | --- | --- | | 4. What do you check first in the model's reassurance? | "It's accurate, so I'd score it high." | "It's accurate and never says when to get care. For a patient asking about a symptom, that gap gets marked down." | | 2. What changed your mind about the case you got wrong? | "I don't think I've had one." | "A case I labelled routine. The detail I skimmed was in the history, and I now read the history first." | Listen for ownership. "We found" describes the team. "I checked, and this is what I do now" describes the person. ![A three-row transcript card. Tracy asks what the expert does when a task is outside their specialty. The expert says they flag it and send it back with a note. Tracy asks how their lead would know the flag was fair and not a way to skip.](/blog/images/medical-expert-interview-questions-1.png) *Illustration with sample data, not a real interview. A fictional expert answers a scope question, and Tracy asks the follow-up that answer calls for. Source: Tracy visual library, qa-followup template. Sample content written by the Qode Product Team.* A follow-up built on the expert's own answer shows whether they have actually done the work. ## Run the same interview for every applicant without losing the follow-up A live call for every applicant does not survive a project that draws hundreds of them. Tracy runs the interview for each expert, and each setting below is on the [product page](/product). - **Brief or your own questions.** Describe the role and Tracy drafts questions you can edit, or paste the 18 above. Style runs from quick to deep. - **CV check, as a setting.** With it on, Tracy can read the CV and ask about what it claims. - **Follow-ups that wait.** Tracy asks for the missing detail in the expert's field and waits for the end of a thought. A director at an accelerator residency programme said: "Tracy sometimes brings up exactly those questions I might have forgotten to ask, which has been a pleasant surprise." - **Start on click.** Each expert opens a browser link with no account and no calendar, at any hour. The join screen says the conversation is recorded and transcribed. Candidates rate Tracy 92 on NPS. - **A result you can check.** A recording, a timestamped transcript and a written assessment for each question sit behind every result, so a reviewer can go from a statement to the moment it was said. - **Integrity signals, with the evidence.** Tracy captures tab changes, gaze, face detection, delay before answers, fullscreen exits and external screens, and shows the evidence to a human reviewer. A signal is not proof, and your team decides. See [AI interview cheating detection](/blog/ai-interview-cheating-detection). Pricing is $1 per credit, one credit per started 20 minutes of an interview that produced a result, so a 30-minute interview takes two credits. The first 3 results are free with no card, there are no seats, and a session with no answers is never charged. See the [pricing page](/pricing), or the [Connect page](/connect) for the API and MCP server. ![A recreated Tracy result report for a fictional clinical reviewer. The recording sits beside a timestamped transcript, including the 8:40 answer about an accurate reply that never says when to seek care. Below, three fields are marked Stated, each with a quoted answer.](/blog/images/medical-expert-interview-questions-2.png) *Illustration, not a screenshot. A recreation of a Tracy result report for a fictional clinical reviewer, with invented data. Source: Tracy visual library, result-traced screen recreation. All names, quotes and times are invented.* A reviewer can read each stated point as the expert said it, with the recording and transcript beside it. ## Keep three checks with people 1. **The credential check.** Verify a licence or degree separately, against the right register, before the interview or at approval. A strong interview from someone with no valid licence is still a reject. 2. **A clinician on your side.** Have one of your clinicians read five or six transcripts and score them blind. If they disagree with the written assessments, change the questions before you invite the next hundred experts. 3. **The decision.** A person decides from the recording, transcript, assessments and any integrity signals. Tracy rejects no one. If you vet ten experts a month, a clinician on a video call is likely cheaper than any tool ([AI interview vs human interview](/blog/ai-interview-vs-human-interview)). The case for Tracy starts when applicants outnumber your reviewers' hours. Pilot on one specialty with 15 to 30 applicants, and read [AI interview reports](/blog/ai-interview-reports) first. ## Common questions ### What should you ask a medical expert in an AI training interview? Ask about reasoning with missing information, how they review a model's answer, when an answer should refer or stop, how they write a rubric criterion and how they keep up with their field. Pick one or two questions from each group above and keep the follow-up beside each one. ### How do you test clinical reasoning without giving medical advice? Ask about process, not outcomes: what they would want to know first, what they check in a model's answer, and what makes an answer unsafe. You never need to state a treatment. ### Can an AI interviewer interview medical experts? Yes. Tracy AI Interviewer asks questions in the expert's own field, follows up on what they said and waits for the answer. Your team gets the recording, a timestamped transcript and a written assessment for each question, and a person decides. ### Does Tracy check a medical licence? No. Tracy does not verify a licence, a degree or any credential. Check credentials separately against the register that applies in the expert's country. ### How long should the interview be? Twenty to thirty minutes covers six questions with their follow-ups. That is our rule of thumb, not a researched figure. A 30-minute interview takes two credits, which is $2. --- Run these questions on your own applicants. The first 3 results are free, with no card. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=medical-expert-interview-questions&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=medical-expert-interview-questions&utm_content=demo) ## Sources - [HealthBench, arXiv 2505.08775, submitted 13 May 2025: rubrics created by 262 physicians](https://arxiv.org/abs/2505.08775), retrieved 9 Nov 2026. - [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 5 Nov 2026. - [Tracy AI Interviewer Connect page](https://tracyinterview.ai/connect), retrieved 5 Nov 2026. - [Tracy AI Interviewer home page, customer proof](https://tracyinterview.ai/), retrieved 9 Nov 2026. # Product Manager Interview Questions With Follow-Ups for Hiring Teams > Product manager interview questions for hiring teams: 15 questions with follow-ups, what strong answers sound like, and how to run them at volume. - URL: https://tracyinterview.ai/blog/product-manager-interview-questions - Published: 2026-11-09 - Author: Tracy team > **In short** **Tracy AI Interviewer** interviews every product manager applicant from your job description and follows up on what each person says, so you see how they decide, not only what they have shipped. > > Use the 15 product manager interview questions below, grouped by what they test, each with a follow-up and what to listen for. > > Every result keeps a recording, a timestamped transcript and a written assessment per question for a person to review. > > Next step: paste these questions into a Tracy interview for one open role. A product role opens and 300 people apply. Every CV says "owned the roadmap" and "drove adoption". None says whether the person ever decided what not to build. Product manager interview questions for a first round have to find that out, and a single answer rarely does. These are written for the person asking the questions. Candidates preparing for an interview will find them useful too, but the notes on what to listen for are for the interviewer. Each question has a follow-up and a note on what a strong answer contains. Tracy AI Interviewer runs the part that does not scale. She asks each applicant questions drawn from your job description, follows up on what they just said and waits for the end of the thought. A recording, a timestamped transcript and a written assessment per question go to you, and you decide. For engineering roles, see [software engineer interview questions](/blog/software-engineer-interview-questions), and for the wider workflow, [hiring at volume](/use-cases/volume-hiring). ## Fifteen product manager interview questions, grouped by what they test ### Group 1: product sense and prioritisation | Ask | Follow up with | Listen for | | --- | --- | --- | | 1. Pick a product you use every day. What would you change first, and for whom? | Who would be unhappy with that change? | Names a user and a problem before a feature. Accepts that the change costs someone something. | | 2. You can fund one of three projects: a retention fix, an enterprise feature a big customer asked for, or a platform cleanup. How do you choose? | What would you want to learn before Friday? | States criteria before the choice. Names the cheapest evidence. Says what waits. | | 3. Tell me about something you decided not to build. | Who pushed back, and what happened? | A real case, a named stakeholder, an outcome someone could check. | ### Group 2: metrics and evidence | Ask | Follow up with | Listen for | | --- | --- | --- | | 4. A feature you launched has 20 percent adoption. Is that good? | What would you compare it with? | Asks for the target and who the 20 percent are. Counts the users it was built for. | | 5. Weekly active users rose and revenue is flat. What do you check? | Which explanation would change the roadmap? | Segments, pricing, mix. Ties the answer to a decision, not a dashboard. | | 6. Tell me about a metric you chose and later stopped trusting. | What replaced it? | A real case, how they noticed, what they do differently now. | ### Group 3: engineering and design | Ask | Follow up with | Listen for | | --- | --- | --- | | 7. An engineer says your date cannot be met. What do you do? | When do you cut scope and when do you move the date? | Asks what drives the estimate. Trades scope in the open. Does not escalate first. | | 8. How do you write a requirement so two engineers build the same thing? | Read me one sentence from a spec you wrote. | Observable acceptance criteria, edge cases, stated non-goals. | | 9. Describe a disagreement with a designer. | What did each of you give up? | The argument is about the user, not taste. Names what moved. | ### Group 4: customers and discovery | Ask | Follow up with | Listen for | | --- | --- | --- | | 10. Tell me about the last customer conversation that changed your plan. | What did you ask that got that answer? | Specific and recent. Open questions. Separates what customers say from what they do. | | 11. Ten customers ask for the same feature. Do you build it? | What would make you say no? | Asks about the problem behind the request, who asked and how much they use or pay. | | 12. How do you find out why people stop using something? | What would you do with only five conversations? | Mixes usage data and interviews. Honest about small samples. | ### Group 5: ownership and stakeholders | Ask | Follow up with | Listen for | | --- | --- | --- | | 13. Tell me about a launch that missed. | What did you tell the company, and when? | Owns it, gives a timeline, names a process change. | | 14. A senior leader asks for a feature you think is wrong. What happens? | What would change your mind? | Disagrees with evidence, commits once it is decided, says what would move them. | | 15. What did you do last quarter that nobody asked you to do? | What did it cost you? | A concrete action with an owner and an honest cost. | *Table 1. The 15 questions, each with a follow-up and what to listen for (written by us).* ## A strong answer names a decision, and a weak one names a framework The difference shows up in the second answer. These pairs are written by us as examples of how it sounds, not quotes from real candidates. | | Weak answer | Strong answer | | --- | --- | --- | | Question 2, one of three projects | "I would use RICE and take the highest score." | "Retention first, because the enterprise request is one account. I'd tell sales it waits a quarter and say why." | | Question 4, 20 percent adoption | "It is okay for a new feature." | "Of whom? If it was built for admins, I'd count admins only, then compare with the last feature we shipped to them." | | Question 13, a launch that missed | "The market was not ready." | "I wrote the requirement badly. I told the company that week, and now every spec names what we won't do." | *Table 2. Example weak and strong answers (illustrative, not from real candidates).* Listen for ownership. A framework name tells you what someone has read. "I decided this, and this is what happened" tells you what they did. ## One follow-up chain shows where the real answer starts This is an illustration with sample data, not a real interview. The candidate is fictional. It shows the shape of what Tracy runs: an opening question from your list, an answer, and a follow-up built on it. It starts from question 4. > **Tracy:** A feature you launched has 20 percent adoption. Is that good? > > **Candidate (fictional):** It depends on the target. For a niche feature, maybe. > > **Tracy:** What would you compare it with? > > **Candidate:** Adoption among the users it was built for. If it was for admins, I would count only admins. > > **Tracy:** Tell me about a time you made that correction. > > **Candidate:** We shipped bulk export and saw 6 percent. Counting only accounts above 50 seats it was 31 percent, so we stopped calling it a failure and moved the entry point. The first answer is the safe one. The third shows whether it ever happened. A Director at a top-tier early-stage VC says Tracy "sometimes brings up exactly those questions I might have forgotten to ask, which has been a pleasant surprise." ![Transcript card: a 20 percent adoption question, a candidate answer, and Tracy asking what they would compare it with.](/blog/images/product-manager-interview-questions-1.png) *Tracy asks a follow-up built on the candidate's own answer. You see whether a product manager has made the call before. Illustration with sample data* A follow-up built on the candidate's own answer shows whether they have done the work. ## Run the interview for every applicant without losing the follow-up A live call for every applicant stops working once a product role draws a few hundred. - **Questions from your job description.** Paste it and Tracy drafts questions you can edit, from quick to deep and from technical to behavioural, or use your own. See the [product page](/product). - **CV check, as a setting.** With it on, Tracy can read the CV and ask about what it claims, so "led the roadmap" becomes a question about one decision. - **Start on click.** Each applicant opens a link in a browser, with no account and no calendar, at any hour, and is told on the join screen that the conversation is recorded and transcribed. See the [participant experience](/participant-experience). - **A result you can check.** A recording, a timestamped transcript and a written assessment per question sit behind every result. It is $1 per credit, one credit per started 20 minutes of an interview that produced a result, so a 34-minute interview costs $2. The first 3 results are free with no card, there are no seats, and a session with no answers is never charged. See [pricing](/pricing). Customisation takes days: send us one role and we will build the interview plan. ![Recreated Tracy result report for a fictional product manager: a recording, a transcript with the 5:58 answer about adoption, and three questions with what they said and quoted evidence.](/blog/images/product-manager-interview-questions-2.png) *A recording and timestamped transcript sit behind every Tracy result. You check what the candidate said before you decide. Illustration with sample data* A reviewer can read what the candidate said to each question and open the recording and transcript behind it. ## Keep three checks with people 1. **A product leader reads five or six transcripts blind.** If they disagree with the written assessments, fix the questions before you invite the next hundred. 2. **A work sample for finalists.** Interview evidence shows how someone reasons. It does not replace a case or a portfolio review. 3. **The decision.** A person reads the evidence and decides. For five applicants a month, a call with the hiring manager is a reasonable choice; see [AI interview vs human interview](/blog/ai-interview-vs-human-interview). The case for Tracy starts when applicants outnumber the hours your product leaders have. Pilot on one role: 20 applicants, one reader. Track completion, hours from application to result and how often the reader disagrees with an assessment. See [AI interview reports](/blog/ai-interview-reports). ## Common questions ### What are good product manager interview questions? Questions that make the candidate decide something: what to build first, what a metric means, what to cut, what to tell a senior leader. Ask one or two from each group above and keep the follow-up beside each. ### What are good product manager interview questions for experienced candidates? Use the same groups and raise the stakes. Ask for decisions with a cost, a launch that missed, a stakeholder they disagreed with, and a number they owned. Ask for the follow-up detail: who, when, what changed. ### How do you judge answers to product manager interview questions? Write what a strong answer contains before the first interview. Look for a named decision, a stated trade-off and ownership of the result. A rehearsed model answer usually breaks at the second follow-up. ### Can an AI interviewer interview product managers? Yes. Tracy asks questions from your job description, follows up on what each person said and waits for the answer. Your team gets the recording, transcript and an assessment per question, and a person decides. --- Create a Tracy interview for this exact role from these questions, then read the recording and report. Or send us one role and we will build the interview plan, with a recording and report you can check. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=product-manager-interview-questions&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=product-manager-interview-questions&utm_content=demo) ## Sources - [Tracy AI Interviewer home page](https://tracyinterview.ai/), retrieved 9 Nov 2026. - [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer Connect page](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. # QA Engineer Interview Questions With Follow-Ups for Hiring Teams > QA engineer interview questions for hiring teams: 15 questions with follow-ups, what strong answers sound like, and how to run them at volume. - URL: https://tracyinterview.ai/blog/qa-engineer-interview-questions - Published: 2026-11-09 - Author: Tracy team > **In short** **Tracy AI Interviewer** interviews every QA applicant from your job description and follows up on what each person says, so you see how they test, not what they memorised. > > Use the 15 QA engineer interview questions below, grouped by what they test, each with a follow-up and what to listen for. > > Every result keeps a recording, a timestamped transcript and a written assessment per question for a person to review. > > Next step: paste these questions into a Tracy interview for one open role. Fifty QA applicants list Selenium, Cypress and test plans. The list does not say what they do when a test fails one run in twenty. QA engineer interview questions for a first round have to expose how someone tests, and a definition recited from a course will not. These are for the hiring manager asking the questions. Each has a follow-up and a note on what a strong answer contains. Tracy AI Interviewer runs the part that does not scale. She asks each applicant questions from your job description, follows up on what they just said and waits for the end of the thought. A recording, a timestamped transcript and a written assessment per question go to your team, who decide. For developers, see [software engineer interview questions](/blog/software-engineer-interview-questions), and for the wider workflow, [hiring at volume](/use-cases/volume-hiring). ## Fifteen QA engineer interview questions, grouped by what they test ### Group 1: test design | Ask | Follow up with | Listen for | | --- | --- | --- | | 1. Here is a login form with email, password and "remember me". How would you test it? | Which three tests do you run first? | Covers boundaries, invalid input, state and security. Orders by risk. | | 2. A feature has ten settings with two values each. How do you decide what to test? | What would you leave untested? | Risk-based thinking, combinations rather than all 1,024. Says what is left. | | 3. How do you write a test case so someone else can run it? | Show me a bad one you rewrote. | Preconditions, steps, an expected result, one idea per case. | ### Group 2: finding and reporting bugs | Ask | Follow up with | Listen for | | --- | --- | --- | | 4. Tell me about the best bug you found. | How did you know it was a bug and not a misunderstanding? | A real case, a reproduction, a check against the spec. | | 5. A developer says "cannot reproduce". What now? | What do you attach next time? | Environment, build, data, logs, a recording. Stays calm. | | 6. How do you decide a bug's severity? | When did you disagree with a product owner about one? | Impact and frequency in user terms. Argues with evidence. | ### Group 3: automation | Ask | Follow up with | Listen for | | --- | --- | --- | | 7. What do you automate first, and what do you leave manual? | Tell me about a test you stopped automating. | Stable, repeated, high-value checks. Counts maintenance cost. | | 8. A test fails one run in twenty. What do you do? | When do you delete it? | Treats it as a defect. Checks timing, shared data and order. Quarantines with a deadline. | | 9. A check could be a unit, API or UI test. How do you choose? | Where does most of your automation sit? | The lowest layer that can catch the bug. Few UI tests. | ### Group 4: release and risk | Ask | Follow up with | Listen for | | --- | --- | --- | | 10. A release is due tomorrow and testing is half done. What do you tell the team? | What would you ask them to decide? | Reports what is covered and what is not. Lets the owner choose. | | 11. What do you check before release beyond the tests? | What did you miss once? | Rollback, monitoring, migrations. An honest miss. | | 12. How do you test something with no specification? | Who do you ask first? | Explores, asks users and developers, writes assumptions down. | ### Group 5: collaboration and judgement | Ask | Follow up with | Listen for | | --- | --- | --- | | 13. How do you work with developers before code is written? | Describe a requirement you questioned. | Raises ambiguity early. Uses examples. | | 14. A teammate's tests all pass but customers report problems. What do you look at? | What would you change in the suite? | Asks what the tests miss. Checks user paths and production feedback. | | 15. Tell me about a time you decided not to test something. | What happened? | A reasoned trade-off. Owns the result. | *Table 1. The 15 questions, each with a follow-up and what to listen for (written by us).* ## A strong answer names the risk, and a weak one names the tool The difference shows up in the second answer. These pairs are written by us as examples of how it sounds, not quotes from real candidates. | | Weak answer | Strong answer | | --- | --- | --- | | Question 1, the login form | "I'd test valid and invalid logins." | "Lockout after repeated failures, reset links that expire, and what 'remember me' keeps after the browser closes. Those carry the most risk." | | Question 8, a flaky test | "Rerun it until it passes." | "A test that fails one in twenty is a defect. I'd check timing and shared data, quarantine it and give it a deadline." | | Question 10, half-finished testing | "We'll push the date." | "Here is what is covered and what is not. The untested part touches payments, so I'd ask the product owner to decide." | *Table 2. Example weak and strong answers (illustrative, not from real candidates).* Listen for how the person tests, not what they memorised. A definition of boundary value analysis tells you what they read. A boundary they found in your own form tells you how they work. ## One follow-up chain shows where the real answer starts This is an illustration with sample data, not a real interview. The candidate is fictional. It starts from question 8 and shows the shape of what Tracy runs. > **Tracy:** A test fails one run in twenty. What do you do? > > **Candidate (fictional):** I rerun it and check whether it is flaky. > > **Tracy:** What would you check if it keeps failing at that rate? > > **Candidate:** Timing and shared data first. A wait that is too short, or tests that depend on order. > > **Tracy:** Tell me about the last flaky test you fixed. > > **Candidate:** A checkout test that read the cart before the API had answered. I replaced a fixed wait with a check on the response, and it stopped failing. The first answer is a reflex. The third shows whether they found a cause. ![Transcript card: a test fails one run in twenty, a candidate answer, and Tracy asking what to check if it keeps failing.](/blog/images/qa-engineer-interview-questions-1.png) *Tracy asks a follow-up built on the candidate's own answer. You see how a tester works, not what they recite. Illustration with sample data* A follow-up built on the candidate's own answer shows whether they found a cause. ## Run the interview for every applicant without losing the follow-up A QA opening can draw hundreds of applicants, many early in their careers, and a live call for each does not fit your week. - **Questions from the role.** Paste the job description and Tracy drafts questions, such as how to test a login form, that you edit or replace with your own. See the [product page](/product). - **No scheduling.** Each applicant opens a link in a browser, at any hour, with no account. The join screen says the conversation is recorded and transcribed. Steven Perry, Account Executive at IT Resources, an IT staffing firm in Tampa, US, says "By the time they finish with this platform, they're even laughing and joking." See the [participant experience](/participant-experience). - **Evidence behind every result.** Your lead gets a recording, a timestamped transcript and a written assessment per question, so a flaky-test answer can be checked against what was said. Pricing is $1 per credit, one credit per started 20 minutes of an interview that produced a result, so a 34-minute interview costs $2. You pay nothing for the first 3 results (no card), there are no seats, and a session with no answers is never charged. See [pricing](/pricing). Send us one role and we will build the interview plan in days. ![Recreated Tracy result report for a fictional QA engineer: a recording, a transcript with the 5:08 answer about flaky tests, and three questions with what they said and quoted evidence.](/blog/images/qa-engineer-interview-questions-2.png) *A recording and timestamped transcript sit behind every Tracy result. Your lead reads the candidate's own words before deciding. Illustration with sample data* A reviewer can read what the candidate said to each question and open the recording and transcript behind it. ## Keep three checks with people 1. **A senior tester reads five or six transcripts blind.** If they disagree with the written assessments, fix the questions before you invite more people. 2. **A hands-on exercise for finalists.** Give them a small feature to test on your own product. An interview shows how someone thinks, and the exercise shows how they find things. 3. **The decision.** A person decides from the evidence. For a handful of applicants, a call with your test lead is a reasonable choice; see [AI interview vs human interview](/blog/ai-interview-vs-human-interview). The case for Tracy starts when applicants outnumber your testers' hours. Start with one role and 20 applicants, and have one tester read every result. Note how many finish, how long it takes from application to result, and how often the tester disagrees with an assessment. See [AI interview reports](/blog/ai-interview-reports). ## Common questions ### What are good QA engineer interview questions? Questions that make the candidate test something live: a form, a flaky failure, a release with half the testing done. Ask one or two from each group above and keep the follow-up beside each. ### What are good QA engineer interview questions for freshers? Use groups 1, 2 and 5. Ask how they would test a login form or an everyday object, how they would report a bug, and how they learn a new feature. Do not ask for tools they have not used; listen for method and curiosity. ### How do you assess answers to QA engineer interview questions? Write what a strong answer contains before the first interview. Look for risk-based choices, a reproducible bug report and a stated trade-off. A list of tool names is not evidence of how someone tests. ### Can an AI interviewer interview QA engineers? Yes. Tracy asks questions from your job description, follows up on what each person said and waits for the answer. Your team gets the recording, transcript and an assessment per question, and a person decides. --- Create a Tracy interview for this exact role from these questions, then read the recording and report. Or send us one role and we will build the interview plan, with a recording and report you can check. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=qa-engineer-interview-questions&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=qa-engineer-interview-questions&utm_content=demo) ## Sources - [Tracy AI Interviewer home page](https://tracyinterview.ai/), retrieved 9 Nov 2026. - [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer Connect page](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. # AI Interview Rules in 2027: Colorado, California, Illinois, NYC and the EU (Nov 2026) > Which AI hiring rules start on 1 January 2027, which already apply, and what each official text says, with dates. An explanation, not legal advice. - URL: https://tracyinterview.ai/blog/ai-interview-rules-2027 - Published: 2026-11-06 - Author: Tracy team Fact-checked on 3 Nov 2026 against the official texts in the Sources list. > **In short** > - A buyer can check a written assessment per question against a transcript timed to the recording, with integrity signals shown as evidence and a person deciding. Two AI hiring rules start on 1 January 2027 and others already apply; this page makes no compliance claim for Tracy. > - Two rules start on **1 January 2027**: Colorado's SB 26-189 and California's privacy-agency ADMT rules. California's employment regulations have applied since **1 October 2025**. > - Explained from the official text, not legal advice. Whether a rule covers your hiring is a question for your counsel. "AI hiring laws 2027" suggests one date. The rules that touch an AI interview carry several, and not all of them fall in 2027. This guide lists each rule that touches an AI interview, with its date and its official words, so you can see what already applies, what starts on 1 January 2027 and what to ask any vendor. It explains the texts. It does not tell you whether a rule covers your hiring, because that depends on facts only you and your counsel have. If you are still deciding what an AI interview is, start with [what an AI interviewer is](/blog/what-is-an-ai-interviewer), or see the [numbers on AI in hiring](/blog/ai-in-hiring-statistics). ## The dates at a glance Each date is when the rule takes effect or, for the California privacy agency, when compliance is due. | Rule | Applies from | What the official text is about | | --- | --- | --- | | Colorado SB 26-189 | 1 Jan 2027 | Notice, an explanation after an adverse outcome and human review on request, for automated decision-making technology used in consequential decisions, which include employment | | California Civil Rights Council regulations on automated-decision systems | 1 Oct 2025 | Existing anti-discrimination law applied to automated-decision systems, and four years of records | | California Privacy Protection Agency (CPPA) ADMT rules | 1 Jan 2027 | Pre-use notice and opt-out or human appeal for automated decisionmaking technology used for significant decisions, hiring included | | Illinois HB 3773 (Public Act 103-0804) | 1 Jan 2026 | AI with a discriminatory effect in employment decisions, and notice to employees | | Illinois Artificial Intelligence Video Interview Act | 1 Jan 2020 | Notice, explanation and consent before AI analyses a recorded video interview | | New York City Local Law 144 | 5 Jul 2023 | Bias audit, public summary and notices for automated employment decision tools | | EU AI Act, Article 5(1)(f) | 2 Feb 2025 | A ban on AI systems that infer emotions in the workplace, with medical and safety exceptions | Some summaries say two state rules start on 1 January 2027. That is close, but the pair is Colorado's SB 26-189 and California's privacy-agency rule, not California's employment regulations. ## Colorado: SB 26-189 starts on 1 January 2027 Colorado's date has moved twice. The 2024 act, SB 24-205, was set for 1 February 2026. A 2025 special-session bill, SB 25B-004, moved it to 30 June 2026. Then in May 2026 the legislature repealed and reenacted the law as SB 26-189, which the governor signed on 14 May 2026. Older articles that give either earlier date are out of date. The signed act says: "Except as otherwise provided in subsection (2) of this section, this act takes effect January 1, 2027." Subsection (2) lists a few provisions, including the appropriation and the safety clause, that took effect on passage. The duties described below start on 1 January 2027, and the act applies to consequential decisions made on or after that date. The official bill summary, which we read on the Colorado General Assembly bill page, defines a consequential decision as one that relates to an individual's "access to, eligibility for, or compensation related to education, employment" and other listed areas, so hiring is in the frame. What the official summary says a deployer, meaning the organisation using the technology, must do: - Notice. "mandating that deployers provide clear and conspicuous notice to consumers at the point of interaction with a covered ADMT." - Explanation. "A deployer is required to provide a consumer with a plain language description of a covered ADMT's role within 30 days after the covered ADMT makes a consequential decision that results in an adverse outcome for the consumer." - Human review. "The act also grants consumers the right to request meaningful human review and reconsideration following a covered ADMT making a consequential decision resulting in an adverse outcome." - Records. "Both developers and deployers are required to retain records necessary to demonstrate compliance with the act for at least 3 years." The summary adds that the attorney general must adopt rules on the post-adverse-outcome disclosures by 1 January 2027, and that before 1 January 2030 the attorney general must give a 60-day notice and chance to cure where a cure is possible. On timing, a law firm note published on 11 May 2026, before the bill was signed, reported that a federal magistrate judge had stayed enforcement of the earlier law on 27 April 2026 and that the attorney general had said it "does not intend to enforce [the CO AI Law] or any legislation replacing or amending [the CO AI Law] until after the rulemaking process has concluded." We did not verify the current position against the court docket or the attorney general, so treat 1 January 2027 as the date the statute takes effect and check with the attorney general before you plan around an enforcement date. ## California: two sets of rules, two dates Two California bodies issued rules about automated systems, with different dates. ### Civil Rights Council regulations: already in force The Civil Rights Council's page says: "The effective date of the regulations is October 1, 2025." (The same page carries a second sentence about the same regulations and an April 2026 date under a different heading, the contractor-nondiscrimination entry. We read the 1 October 2025 date as the one that applies, and you should confirm with the Council.) The Council's press release says the regulations "clarify the application of existing antidiscrimination laws" to tools like these. Among other things they aim to: - "Make it clear that the use of an automated-decision system may violate California law if it harms applicants or employees based on protected characteristics, such as gender, race, or disability." - "Ensure employers and covered entities maintain employment records, including automated-decision data, for a minimum of four years." - "Affirm that automated-decision system assessments, including tests, questions, or puzzle games that elicit information about a disability, may constitute an unlawful medical inquiry." We read the Council's press release, not the final text, because the final-text PDF is a scanned image. ### CPPA ADMT rules: compliance from 1 January 2027 The privacy agency's rules are a separate set under the California Consumer Privacy Act. The agency announced on 23 September 2025 that "The regulations go into effect January 1, 2026" with more time for some parts, including automated decisionmaking technology. We read the regulation text itself. Section 7200(b) says: "A business that uses ADMT for a significant decision prior to January 1, 2027, must be in compliance with the requirements of this Article no later than January 1, 2027." Two definitions decide who this reaches. Section 7001(ddd) lists "employment or independent contracting opportunities or compensation" among significant decisions, and (ddd)(4) says that includes "Hiring;". Section 7001(e) defines ADMT as "any technology that processes personal information and uses computation to replace human decisionmaking or substantially replace human decisionmaking." It adds that substantially replacing means using the output without human involvement, and that human involvement requires a reviewer to "Know how to interpret and use the technology’s output to make the decision;", to review the output and other relevant information, and to "Have the authority to make or change the decision based on their analysis". The article also asks for a pre-use notice ("must provide consumers with a Pre-use Notice") and an opt-out ("A business must provide consumers with the ability to opt-out of the use of ADMT to make a significant decision concerning the consumer, except as set forth in subsection (b)"), where subsection (b) includes an appeal to a human reviewer who can overturn the decision. We did not check how the CCPA treats job applicants as "consumers", and the conditions for each exception are longer than we quote here. Take both questions to counsel. ## Illinois: two laws Illinois HB 3773 (Public Act 103-0804). The bill status page records "Effective Date January 1, 2026" and the synopsis says it is a civil rights violation, in recruitment, hiring and other employment decisions, "for an employer to use artificial intelligence that has the effect of subjecting employees to discrimination on the basis of protected classes" or to use zip codes as a proxy, and "for an employer to fail to provide notice to an employee that the employer is using artificial intelligence." The synopsis says "employee"; we did not read the act to see whether applicants are covered. We read the bill page, not the public act. The Artificial Intelligence Video Interview Act (820 ILCS 42). This one has applied since 1 January 2020 and speaks directly to recorded video. Section 5 says an employer that asks applicants to record video interviews and uses AI analysis of them, when "considering applicants for positions based in Illinois," must, before asking, "Notify each applicant before the interview that artificial intelligence may be used to analyze the applicant's video interview and consider the applicant's fitness for the position," explain "how the artificial intelligence works and what general types of characteristics it uses to evaluate applicants," and obtain consent. It adds: "An employer may not use artificial intelligence to evaluate applicants who have not consented to the use of artificial intelligence analysis." Section 15 requires deletion of an applicant's interviews within 30 days of a request. Whether a given live interview counts as a "recorded video interview" with "artificial intelligence analysis" is a legal question, and we do not answer it. We read an archived copy of the official ilga.gov page, and later amendments were not checked. For the difference between a recorded one-way interview and a live two-way one, see [one-way versus two-way video interviews](/blog/one-way-vs-two-way-video-interviews). ## New York City: Local Law 144 The city's Department of Consumer and Worker Protection (DCWP) says Local Law 144 of 2021 "prohibits employers and employment agencies from using an automated employment decision tool unless the tool has been subject to a bias audit within one year of the use of the tool, information about the bias audit is publicly available, and certain notices have been provided to employees or job candidates." The city rule's page gives "Effective date : July 5, 2023", and DCWP says it "will begin enforcement of this law and rule on July 5, 2023." The law names the employer or employment agency that uses the tool. A vendor's audit helps, but the duty in the quoted sentence sits with the user. ## The EU: the emotion-recognition ban Article 5(1)(f) of the EU AI Act prohibits "the use of AI systems to infer emotions of a natural person in the areas of workplace and education institutions, except where the use of the AI system is intended to be put in place or into the market for medical or safety reasons." Article 113 says "Chapters I and II shall apply from 2 February 2025", and Article 5 sits in Chapter II. The Service Desk marks Article 113 as amended by the Digital Omnibus on AI (consolidated text as at 27 July 2026); the date for point (f) is unchanged. A later date, 2 December 2026, applies to other Article 5 points, not to point (f). The text does not mention interviews. Whether a product "infers emotions" is a legal and technical question, so ask the vendor in writing (question 6 below). ## What the rules have in common The texts above keep asking for four things: - Telling the person. Colorado, California's CPPA rules, both Illinois laws and New York City each ask for some form of notice. - A person who can decide. Colorado gives a right to human review on request, and the CPPA definition of ADMT turns on whether a human reviewer with authority is involved. - Records. California says four years for automated-decision data; Colorado says at least 3 years for developers and deployers. - Checking for bias. New York City asks for an audit and a public summary; California and Illinois treat discriminatory effect as a violation of existing law. For how an AI interview report can support a reviewer, see [what reviewers should see in an AI interview report](/blog/ai-interview-reports). ## What to ask any AI interview vendor Ask them in writing. Good answers do not make anyone compliant, but they show what you are buying. 1. Does the tool score, rank or recommend, or does it record what was said? Colorado's definition of ADMT covers "predictions, recommendations, classifications, rankings, scores", and New York City's law is about tools used in decisions. 2. Who decides, and can a person see the evidence and overrule? This is the human-involvement test in the CPPA definition and the human-review right in Colorado. 3. What does the person see before the interview, and can we edit the notice? Colorado, the CPPA, Illinois and New York City each ask for notice. 4. Can you give us documentation of intended uses, known limitations and training data categories? Colorado asks developers for it from 1 January 2027. 5. Has the tool had a bias audit in the past year, and where is the public summary? This is the New York City question. 6. Does it infer emotions or personality traits from a face or a voice? This is the EU question, and the Illinois video act asks employers to explain "what general types of characteristics" the AI uses. 7. How long are recordings kept, who else sees them and how do we delete one on request? Illinois gives 30 days to delete on request. 8. What records can we export, and for how long? California says four years; Colorado at least 3. 9. How do you handle a person who asks to see or correct their data, or asks for a human review? Colorado grants these rights. For a wider list of what to check, see [how AI interviews work for employers and candidates](/blog/how-do-ai-interviews-work). ## What Tracy's own pages show Tracy AI Interviewer runs a two-way video interview in the browser. Its published product and home pages say this on the points above. - The decision. The product page says: "Tracy does not reject anyone. A person decides." The home page says: "A person reviews the evidence and decides." This helps with question 2. - Recording and transcript. The product page says the interview is "recorded on both sides, with a transcript timed to the recording", so a reviewer reads what was said beside the recording. This helps with question 2. - A written assessment per question. The product page says Tracy writes "an assessment for each question with the reasoning behind it". The home page says: "You read the words, then you decide." A reviewer can check each assessment against the transcript. - Integrity signals, as signals. During the interview Tracy captures signals such as where the candidate is looking, response timing, tab switching, leaving fullscreen and an external screen. Nothing is decided during the interview: afterwards the signals are assessed together and the evidence goes to the reviewer. Tracy's [cheating-detection guide](https://tracyinterview.ai/blog/ai-interview-cheating-detection) says a signal calls for review and does not establish that someone cheated. A signal is not a verdict, and the person reviewing decides. - A notice before the interview. The product page says: "People are told about the recording before it starts." That is a recording notice. It does not claim to meet any notice rule above, and what your notice must say is for you and your counsel to settle. This page makes no claim that Tracy complies with any rule above, has had a bias audit, or has been assessed under any of them. It does not answer questions 4 to 6 for Tracy, so ask us in writing, as you would any vendor. A demo shows a real result. For the reasoning behind keeping a person in the decision, see [AI interview versus human interview](/blog/ai-interview-vs-human-interview). ## When a vendor answer is not enough A vendor can tell you what its tool does. It cannot tell you whether a rule covers your hiring, which roles and locations count, or what your notice must say. Those answers depend on where candidates are, who you hire and how you use the output, and they belong with your counsel. If New York City's law covers a tool you use, the audit duty in the DCWP sentence sits with you as the user. ## Common questions ### Which AI hiring rules start on 1 January 2027? Two, as far as the official texts show: Colorado's SB 26-189, which takes effect on 1 January 2027, and the California Privacy Protection Agency's automated decisionmaking technology rules, which require a business using that technology for a significant decision, hiring included, to comply by 1 January 2027. ### Is California's AI hiring rule already in force? California has two separate sets. The Civil Rights Council's automated-decision-system employment regulations have been in effect since 1 October 2025. The privacy agency's ADMT rules apply from 1 January 2027. ### Does New York City's Local Law 144 still apply? Its city rules took effect on 5 July 2023, and the city's Department of Consumer and Worker Protection says it began enforcement that day. We saw no change on its page on 3 November 2026. ### Does the EU AI Act ban AI interviews? No text we read bans interviews as such. Article 5(1)(f) bans AI systems that infer emotions of a person in the workplace, except for medical or safety reasons, and it has applied since 2 February 2025. ### Does Tracy comply with these laws? This page makes no such claim. It shows what Tracy's own pages state: a person reviews the evidence and decides, every interview is recorded on both sides with a transcript timed to the recording, each question gets an assessment, integrity signals are signals for a person to review, and people are told about the recording before it starts. ### Is this legal advice? No. It explains official texts as they read on 3 November 2026. Whether a rule covers your hiring is a question for your counsel. --- **[Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=ai-interview-rules-2027&utm_content=try-free)** and read a real result with the recording and transcript beside it, or [book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=ai-interview-rules-2027&utm_content=demo) to ask us your vendor questions in person. ## Sources All sources retrieved 3 Nov 2026. - Colorado General Assembly, [SB 26-189 bill page](https://leg.colorado.gov/bills/sb26-189): the enacted bill summary (consequential decision, notice, explanation, human review, records, enforcement), signed 14 May 2026. Opens in a browser; it refuses automated requests. The summary wording quoted above appears only here. - Colorado General Assembly, [SB 26-189 signed act (PDF)](https://leg.colorado.gov/bill_files/116489/download): section 5, effective date and applicability (some provisions on passage, the duties on 1 Jan 2027). - Proskauer, [Major Developments Put Colorado's AI Law on Ice Ahead of Implementation](https://www.proskauer.com/blog/major-developments-put-colorados-ai-law-on-ice-ahead-of-implementation), 11 May 2026: a law firm report of the stay of enforcement and the attorney general's stated intent. Secondary source, written before signing. - California Civil Rights Council, [rulemaking actions](https://calcivilrights.ca.gov/civilrightscouncil/rulemaking-actions/): effective date of the automated-decision-system regulations. - California Civil Rights Council, [press release of 30 Jun 2025](https://calcivilrights.ca.gov/2025/06/30/civil-rights-council-secures-approval-for-regulations-to-protect-against-employment-discrimination-related-to-artificial-intelligence/): what the regulations do. - California Privacy Protection Agency, [announcement of 23 Sep 2025](https://cppa.ca.gov/announcements/2025/20250923.html): regulations effective 1 Jan 2026 with ADMT compliance from 1 Jan 2027. - California Privacy Protection Agency, [text of regulations (PDF)](https://cppa.ca.gov/regulations/pdf/ccpa_updates_cyber_risk_admt_appr_text.pdf): ADMT definition, significant decision including hiring, sections 7200, 7220 and 7221. - Illinois General Assembly, [HB 3773 bill status](https://web.archive.org/web/2026/https://ilga.gov/ftp/legislation/103/BillStatus/HTML/10300HB3773.html): Public Act 103-0804, effective 1 Jan 2026, synopsis (archived copy of ilga.gov). - Illinois General Assembly, [Artificial Intelligence Video Interview Act, 820 ILCS 42](https://web.archive.org/web/2024/https://www.ilga.gov/legislation/ilcs/ilcs3.asp?ActID=4015&ChapterID=68): sections 5, 15 and 20 (archived copy of ilga.gov). - NYC Department of Consumer and Worker Protection, [Automated Employment Decision Tools](https://www.nyc.gov/site/dca/about/automated-employment-decision-tools.page): what Local Law 144 requires and enforcement from 5 Jul 2023. - NYC Rules, [Automated Employment Decision Tools (Updated)](https://rules.cityofnewyork.us/rule/automated-employment-decision-tools-updated/): adopted rule, effective 5 Jul 2023. - European Commission AI Act Service Desk, [Article 5](https://ai-act-service-desk.ec.europa.eu/en/ai-act/article-5): prohibited practices, including emotion inference in the workplace. - European Commission AI Act Service Desk, [Article 113](https://ai-act-service-desk.ec.europa.eu/en/ai-act/article-113): application dates (page marked as amended by the Digital Omnibus on AI; point (f) date unchanged). - Tracy AI Interviewer, [home page](https://tracyinterview.ai/): a person reviews the evidence and decides; you read the words, then you decide; integrity signals. - Tracy AI Interviewer, [product page](https://tracyinterview.ai/product): recording notice, two-way video, recording and transcript timed to it, an assessment per question, "Tracy does not reject anyone. A person decides." - Tracy AI Interviewer, [AI interview cheating detection](https://tracyinterview.ai/blog/ai-interview-cheating-detection): integrity signals flagged for review, not as proof of cheating. # How an AI Interviewer Uses the CV to Ask Better Questions > What a CV can and cannot tell you, why one script for every applicant wastes the interview, and how Tracy AI Interviewer can ask questions based on each resume. - URL: https://tracyinterview.ai/blog/ai-interviewer-cv-check - Published: 2026-11-06 - Author: Tracy team Fact-checked on 4 Nov 2026 against the tracyinterview.ai pages listed in Sources. > **In short** > - **Tracy AI Interviewer can read each candidate's CV, ask 3 to 5 questions that come from it and from the job's must-haves, and follow up on every answer,** so the interview tests what the CV claims instead of repeating a script. > - A good CV question names a specific thing on the page and asks for the decision behind it. The examples below are ours, not Tracy output. > - CV check is a setting you can switch off, and a person decides what happens next. A CV filter reads a CV for keywords and ranks it. An interview can do something different with the same page: ask the candidate about it. Tracy AI Interviewer does this when CV check is on. It receives the job description and the candidate's CV along with your question set, and adds questions drawn from that CV and from the job's must-have requirements. ## What can a CV tell you, and what can't it? A CV tells you what someone says they did: employers, titles, dates, tools, results. That is a useful list of things to ask about. It does not tell you how much of the work was theirs, what they would do differently, or whether the stated result came from their decision or from the team around them. Two candidates can list the same tool at the same employer. One built the system and the other used it twice. A keyword filter sees the same line twice. A conversation tells them apart in two or three questions. Treat the CV as a list of claims to test. It gives you questions to ask. ## Why does one script for every applicant waste the interview? A fixed question set gives everyone the same questions, which is fair. It also leaves the most useful material unused. The candidate's own CV holds the claims most worth testing, and a generic script never touches them. A polished general answer from a strong candidate sounds the same as a polished general answer from someone who read about the topic. A question about their own project cannot be answered from a blog post. Use both. Keep the role's questions for comparison across candidates, and add a few questions that come from each CV. ## How does Tracy's CV check work? On the [product page](https://tracyinterview.ai/product), Tracy describes it in one line: CV check adds 3 to 5 questions from each CV and the must-haves. In practice: 1. **Setup.** Tracy drafts the questions from the job description or your brief. You can edit them, replace them with your own and set the style from quick to deep and from behavioural to technical. 2. **CV check on.** Tracy receives the job description, the candidate's CV and the question set. It asks 3 to 5 tailored questions based on the CV and the job's must-have requirements, in addition to the role's questions. 3. **Follow-ups.** Each follow-up depends on the answer before it, so a CV question can open into two or three more. Tracy waits for the end of a thought before it speaks, with longer pauses allowed on hard questions. 4. **Review.** The interview is a two-way video call, recorded on both sides, with a transcript timed to the recording. CV check is a setting, and it only runs when you switch it on. ## What does a good CV question look like? A good CV question has three parts: it names something specific on the page, it asks about the candidate's own part in it, and it leaves room to follow up. The six questions below are invented examples that show the pattern. Tracy's own questions will differ with each CV and job. | CV line (invented) | Example question | What it tests | | --- | --- | --- | | "Led migration of billing service to a new database" | What did you decide yourself on that migration, and what was decided for you? | Ownership | | "Reduced support backlog by 30%" | Which change did the most for that 30%, and how did you know? | Cause and measurement | | "Managed a team of 8" | Tell me about a time one of the eight was not delivering. What did you do first? | People management in practice | | "Experience with Python and SQL" | Describe a query you wrote that was slow. How did you find the problem? | Depth behind a tool list | | Gap of 14 months between roles | What were you working on during that time, and what would you say you learned? | An invitation to explain, asked neutrally | | "Owned on-call for the payments API" | Tell me about the worst page you got. What did you change afterwards? | Ownership under pressure | Three habits make these work. Ask about the decision, not the title. Ask what they would change, because people who did the work have views. Ask the neutral version of any question about a gap, a short stay or a career change, so the candidate can explain without being put on the defensive. The follow-up does the rest. When the candidate says "I chose a composite index," the next question is why, and what it cost on writes. ## How do you review a CV-based interview? After the interview, Tracy writes an assessment for each question with the reasoning behind it, plus an overall assessment. Warm-up, notice period and salary questions are left out. Every point links to the moment in the recording, and the full transcript sits beside the video. For a CV question, that means you can read what was asked, read the candidate's answer in their own words and play the moment. If the assessment says the candidate gave a specific account of the work, you can check that against the transcript in a minute. If it says the answer stayed general, you can hear it for yourself. Tracy does not reject anyone. A person reviews the evidence and decides. The [AI screening interviews guide](/blog/ai-screening-interviews) covers where this fits in a hiring process, and [how AI interviews work](/blog/how-do-ai-interviews-work) walks through the interview from invitation to result. For the candidate's side, [How to prepare for an AI interview](/blog/how-to-prepare-for-an-ai-interview?utm_source=blog&utm_medium=seo&utm_campaign=ai-interviewer-cv-check&utm_content=internal-link) covers what to expect and what to ask the employer. ## How do you turn CV check on and edit the questions? CV check sits in the interview setup, next to the questions Tracy drafted from your brief. Turn it on for the interviews where the CV is part of the evidence. You can edit any drafted question, replace the set with your own, or leave the role questions as they are and let Tracy add the CV ones. If your team starts interviews from an AI app, Tracy runs as an MCP server, and the [Workable guide](/blog/tracy-workable-integration) shows one way to invite candidates from there. ## What should you watch for on fairness? CV questions are personal by design, so a few practices keep them fair. - **Ask the same must-haves of everyone.** The job's must-have requirements are the common thread. Whatever else differs between candidates, each should be asked about the same must-haves, in whatever form their CV makes natural. - **Do not turn CV details into a trap.** A question about a gap or a short stay is a chance to explain. Keep it neutral, and remember that "I don't have a good answer" can say as much about the question as about the candidate. - **Read for specifics.** A fluent answer with no specifics is a weaker signal than a halting one with names and numbers. The transcript lets you see the difference. - **Humans decide.** The assessment is a written summary with reasoning. The hiring decision stays with your team, and the recording is there to check the assessment against. The [AI behavioural interviews guide](/blog/ai-behavioral-interviews) goes deeper on behavioural questions and how to read the answers. ## When should you leave CV check off? Leave it off when the CV is not the evidence. Some cases: - **Roles where the work sample is the evidence,** such as a trade, a task-based role or an entry-level role where a CV has little on it. - **Interviews where identical questions matter more than tailored ones,** for example when you want to compare many candidates on exactly the same prompts. - **Research or customer interviews,** where there is no CV to read. - **Candidates with sparse CVs,** where questions from the page would be thin and the role's questions do more. Turning it off costs you nothing in the rest of the interview: the questions, follow-ups, recording, transcript and assessment still run. ## Common questions ### Can an AI interviewer ask questions based on a resume? Yes. With CV check on, Tracy AI Interviewer can add 3 to 5 questions drawn from each candidate's CV and the job's must-have requirements, then follow up on what the candidate says. ### Does Tracy reject candidates based on their CV? No. The CV is used to shape the questions. Tracy writes an assessment for each question with its reasoning, and a person reviews the evidence and decides. ### Can I edit the CV questions before the interview? You can edit the questions Tracy drafts from the job description, or replace them with your own, and CV check is a setting you turn on or off per interview. ### When should CV check be off? Leave it off when the CV is not the evidence for the role, for example roles where people show skill by doing a task, or when you want every candidate to get exactly the same questions. ### What does the recruiter see afterwards? The recording of both sides, a transcript timed to the recording, and a written assessment for each question with the reasoning behind it, plus an overall assessment. --- **[Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=ai-interviewer-cv-check&utm_content=try-free)** with CV check on one of your own roles, or [book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=ai-interviewer-cv-check&utm_content=demo). ## Sources All sources retrieved 4 Nov 2026. - Tracy AI Interviewer, [product page](https://tracyinterview.ai/product): questions from the job description or your brief, edit or replace them, CV check adds 3 to 5 questions from each CV and the must-haves, follow-ups that depend on the answer, two-way recorded video with a timed transcript, an assessment per question with reasoning, "Tracy does not reject anyone. A person decides.", API and MCP. - Tracy AI Interviewer, [home page](https://tracyinterview.ai/): CV check, follow-ups, assessment per question with timestamps, "A person reviews the evidence and decides." - Tracy AI Interviewer, [volume hiring](https://tracyinterview.ai/use-cases/volume-hiring): "Turn on CV check to add 3 to 5 questions from each CV." # How to Vet Domain Experts at Scale for AI Training Projects > Tests show answers; follow-ups show expertise. A layered way to vet domain experts for AI training work at scale, with a follow-up chain, integrity signals and where each layer fits. - URL: https://tracyinterview.ai/blog/vet-domain-experts-at-scale - Published: 2026-11-06 - Author: Tracy team Fact-checked on 3 Nov 2026 against the public platform pages and announcements listed in Sources. > **In short** > - **Tracy AI Interviewer can take the conversation layer of expert vetting,** asking field-specific follow-ups of every applicant and leaving reviewers the recording, a transcript timed to it and integrity signals with evidence. > - Tests show answers. Follow-ups show whether the expertise is real and the person is the one who earned it. Use both, plus identity and trial-task checks. Vetting domain experts for AI training projects breaks in a predictable way as you scale. Ten applicants, and a domain lead talks to all ten. A thousand applicants, and you lean on a form, a test and a credential check. Each of those can pass someone who can write a fluent answer and cannot do the work, and fail someone who can do it and writes plainly. This guide sets out a layered way to vet experts at volume, shows what a follow-up adds that a test cannot, and says where Tracy AI Interviewer fits and where it does not. ## Why does vetting break at scale? Public pages from the platforms that do this show the shape of the work. Outlier, an AI training platform, says its onboarding takes 30 to 90 minutes and "includes creating your account, selecting your areas of expertise, completing skill screenings so we can match you with the right projects, and identity verification". Its minimum qualification, it says, is typically undergraduate-level expertise, so a screening has to do the separating. The tasks are demanding. Outlier lists writing a challenging prompt, creating grading rubrics and rating and ranking answers, and asks experts to write a difficult question in their field, "one that would confuse an AI model into writing an incorrect answer." DataAnnotation describes remote contract roles "matched to your skills", across fields from finance and law to data science. For the same field, one expert writes a prompt a model gets wrong and a rubric two graders score alike. Another writes a vague one. You find out which after the project has started, unless the vetting asked. The stakes are visible in the market. Mercor, which says it connects human expertise with AI labs and enterprises, announced a [$350 million Series C at a $10 billion valuation](https://mercor.com/blog/series-c/) on 27 Oct 2025. The AI Insider reported in September 2025 that micro1 secured a [$35M Series A at a $500M valuation](https://theaiinsider.tech/2025/09/18/micro1-secures-35m-series-a-at-500m-valuation-to-scale-human-expertise-for-ai-training/). We cite both as market context: they show how much money now rides on expert supply. They say little about how a platform vets, apart from micro1's AI recruiter, covered under the alternatives below. ## What should vetting check? Five things, and each is a place a test alone falls short. | What to check | What a test shows | What a follow-up adds | | --- | --- | --- | | Domain depth | A right or wrong answer | How far the knowledge goes past the first answer, and where it stops | | Reasoning | A final answer | The path to it, and what would change it | | Prompt and rubric writing | A written sample | Whether the expert can defend the rubric against a case they did not plan for | | Catching model errors | Whether they flagged a planted error | Why it is wrong, and how they would make a grader catch it | | Integrity | A score, from whoever answered | Whether the person can keep going on their own answer, live | Our guide to [AI trainer interview questions](/blog/ai-trainer-interview-questions) has 25 questions across these areas. For a data scientist as the expert, see the [data scientist question bank](/blog/data-scientist-interview-questions). ## What does a layered vetting process look like? No single layer carries the load. A process that holds at volume stacks five. 1. **Identity and credentials.** Confirm the person is who they say they are and holds what they claim. This is its own step, and an interview does not replace it. 2. **A screen.** A short test or questionnaire that removes the clearly unqualified, cheaply, for everyone. 3. **A conversation in the expert's field.** A two-way interview where each follow-up builds on the last answer. This is where depth and ownership show. 4. **A paid trial task.** Real project material, reviewed closely. It shows how they work, not only how they talk. 5. **Ongoing quality checks.** Review early outputs, and keep sampling after onboarding. Layer 3 is the costly one at volume, because each conversation takes a domain lead's time. Tracy can run that layer: the same conversation with every expert who passes layer 2, each starting when they click the link at any hour, with your leads reading the results instead of holding the calls. ## What does a follow-up chain add? An example Here is an example we wrote to show the shape of a chain, in statistics. It is not a transcript of a real interview. The opening question gives the expert a model answer to assess. > **Interviewer:** A model answered: "The sample mean of n independent Cauchy draws converges to the true mean by the law of large numbers." Is that right? > > **Expert:** No. The Cauchy distribution has no mean, so there is no true mean to converge to. > > **Interviewer:** Why does the law of large numbers not apply? > > **Expert:** It needs a finite expectation. For Cauchy, the integral for the mean does not converge absolutely. > > **Interviewer:** What does the sample mean of 100 Cauchy draws look like? > > **Expert:** It has the same Cauchy distribution as a single draw. Averaging does not narrow it, which is why a few extreme values keep showing up however many you take. > > **Interviewer:** How would you write a rubric item so a grader catches this error? > > **Expert:** I'd make the check "states that the mean does not exist" a required item, and add a note that "converges by the law of large numbers" is a fail, even when the final sentence sounds right. The first answer is where a test would stop. The second shows the expert knows the condition behind the rule, and the third whether they can say what actually happens, which is hard to fake from a half-remembered summary. The fourth checks the skill the project needs most: turning an error into something a grader can score. Now the same opening with a weaker expert, also written by us. | Question | Weaker answer | What it shows | | --- | --- | --- | | Is the model's statement right? | "Yes, the law of large numbers says averages converge." | Recalls the rule, misses its condition. | | Why would it not apply? | "Maybe the sample is too small?" | Guessing at a cause. | | What does the mean of 100 draws look like? | "Close to zero, probably." | Cannot say what actually happens. | The weaker expert might have scored well on a multiple-choice test of the same topic. Three follow-ups tell you more than the test did. ## How do you check the expert is the one answering? Fluent written answers are easy to get from a model now. Two checks work together. First, the follow-up itself. A question built on what the expert just said cannot be pre-written, and an expert who is relaying someone or something else slows down when the next question goes a layer deeper. Our guide to [AI interview cheating detection](/blog/ai-interview-cheating-detection) covers this in full. Second, integrity signals with evidence. Tracy captures the face and where the expert is looking, response timing, tab switching, leaving fullscreen and an external screen, and nothing is decided during the interview: afterwards the signals are assessed together and the evidence goes to your reviewer. Tracy's [cheating-detection guide](https://tracyinterview.ai/blog/ai-interview-cheating-detection) names five signals flagged for a closer look: tab changes, abnormal eye tracking, live face detection, the average delay before an answer and external screens. The guide says a signal calls for review and does not establish that someone cheated. A person looks at the moments flagged and decides. Do not treat a signal as proof: people read, look away and lose a connection for ordinary reasons. ## How does Tracy fit in the layers? Tracy AI Interviewer (see [what an AI interviewer is](/blog/what-is-an-ai-interviewer)) runs layer 3. It runs [two-way video interviews](/blog/one-way-vs-two-way-video-interviews). With CV check on, Tracy reads the expert's CV and the project's must-have requirements and asks 3 to 5 questions tailored to them, on top of your own. Follow-ups are in the expert's own field and build on what was said, and Tracy waits for the end of a thought, so a long technical answer is not cut off. Questions are generated from the project description and can be edited or replaced, so the field-specific questions you already use can go in as written. Four details matter at volume: - Experts join from a browser with nothing to install, whenever it suits them, which suits experts in many time zones. - People are told about the recording before it starts. - Each expert is interviewed when they click the link, with no scheduling, and a candidate who drops can rejoin the same session within 10 minutes. - Tracy connects through an API and an MCP server, so interviews can be created and read from the tools a team already uses, and enterprise teams can ask for a customised setup. For published prices across tools, see [AI interview cost](/blog/ai-interview-cost). Your domain leads get the recording of both sides with a transcript anchored to it by timestamp and a written assessment per question with its reasoning, so they can check a claim against what the expert said. Tracy does not reject anyone: a person reviews the evidence and decides. Our guide to [AI interview reports](/blog/ai-interview-reports) shows what a reviewer should look for. ## When a different route is better - **If you would rather have a marketplace source and screen experts for you.** Marketplaces such as Mercor say they connect experts with AI labs and enterprises, and micro1, according to The AI Insider, recruits its experts through its own AI recruiter, Zara. You trade control of the vetting for speed, which suits a team that needs experts this month and has no pipeline of its own. - **If you need to see the expert do the task.** Run a paid trial on real material. An interview shows how they think; a trial shows what they produce. - **If you need identity documents or credentials checked.** Use a dedicated check. Tracy is not a credential check. - **For a senior reviewer or a lead role.** Put a domain lead in the conversation. Tracy fits the first conversation with everyone, not the final one. - **For people who have not agreed to be recorded.** Tracy is not for talking to someone who has not agreed to be recorded. For the technical side of a coding or system-design check, see [AI technical interviews](/blog/ai-technical-interviews). If your problem is volume before vetting begins, [screening when there are too many applicants](/blog/how-to-screen-candidates) is the place to start. ## Common questions ### How do you vet a domain expert for AI training work? Layer the checks: confirm identity and credentials, screen with a test, then hold a conversation in the expert's own field where each follow-up builds on the last answer. Finish with a paid trial task. The conversation is the layer that shows depth and whether the expert did the thinking. ### Why are tests not enough to vet experts? A test shows an answer. It cannot ask why, or what would change the answer, and a fluent written answer can come from a person, a search or a model. A follow-up on the expert's own answer is harder to feed. ### Can an AI interviewer vet domain experts? It can run the conversation layer with every applicant: read the CV and tailor questions to it, ask your questions, follow up in the expert's field, and give your reviewers the recording, a timestamped transcript and a written assessment per question. Your domain leads review the evidence and decide. ### How do you check an expert is not using AI during the interview? Follow-ups on their own answer help expose rehearsed or fed answers. Tracy flags integrity signals such as tab changes and abnormal eye tracking for a closer look. They are signals, not a verdict, and a person decides. ### What should you do after the interview? Give the shortlist a paid trial task on real project material and review the first outputs closely. Keep checking quality after onboarding, not only before. --- **[Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=vet-domain-experts-at-scale&utm_content=try-free)** and run the conversation layer with your next expert applicants, or [book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=vet-domain-experts-at-scale&utm_content=demo). ## Sources All sources retrieved 3 Nov 2026. - Outlier, [home page](https://outlier.ai/): common tasks, minimum qualifications, onboarding with skill screenings and identity verification. - DataAnnotation, [home page](https://www.dataannotation.tech/): remote contract roles matched to skills. - Mercor, [Unlocking human potential in the AI economy](https://mercor.com/blog/series-c/), 27 Oct 2025: Series C and valuation, as the company states them. - The AI Insider, [Micro1 Secures $35M Series A at $500M Valuation to Scale Human Expertise for AI Training](https://theaiinsider.tech/2025/09/18/micro1-secures-35m-series-a-at-500m-valuation-to-scale-human-expertise-for-ai-training/), 18 Sep 2025: funding of micro1 and its AI recruiter, Zara, as reported. - Tracy AI Interviewer, [product page](https://tracyinterview.ai/product): two-way video interview, questions from the project description or your own, CV check, follow-ups, recording notice, recording and transcript timed to it, an assessment per question, integrity signals, API and MCP. - Tracy AI Interviewer, [home page](https://tracyinterview.ai/): a person reviews the evidence and decides; integrity signals with their evidence. - Tracy AI Interviewer, [pricing](https://tracyinterview.ai/pricing): API and MCP connections and a customised setup for Enterprise. - Tracy AI Interviewer, [Connect page](https://tracyinterview.ai/connect): MCP server and API connection to the tools a team already uses. - Tracy AI Interviewer, [AI interview cheating detection](https://tracyinterview.ai/blog/ai-interview-cheating-detection): five integrity signals flagged for a closer look, and that a signal calls for review without establishing cheating. # Do Candidates Like AI Interviews? What Candidates Should Know, What Employers Should Check > Do candidates like AI interviews? What candidates should know, what employers should check, and why candidates rate Tracy 92 on NPS. - URL: https://tracyinterview.ai/blog/ai-interview-candidate-experience - Published: 2026-11-05 - Author: Tracy team > **In short** Do candidates like AI interviews? Candidates rate Tracy 92 on NPS. > > Candidates open a browser link, see a notice that the conversation is recorded and transcribed, and get follow-ups on their own answers. > > A person reads the evidence and decides, and a candidate who drops can rejoin within 10 minutes. > > Next step: take an interview yourself, as a candidate would. ## Do candidates like AI interviews? Some do and some do not, and the answer depends on the process around the interview as much as on the AI. The only candidate data on this page is Tracy's own NPS, so what follows is advice for employers, not a claim about how candidates feel in general. Tell candidates who reads the result and what is recorded. Say whether there is a timer. Let them start when it suits them and give them a way back in if a connection drops. That makes candidate experience a design question for the employer, and this guide explains what a candidate meets, in order, and what to check before you send the link. ![Candidate journey: join screen, conversation, follow-ups on answers, human review.](/blog/images/ai-interview-candidate-experience-1.png) *Tracy shows candidates a notice first and puts a person at the end, so they know who reads the result. Illustration, not a screenshot. Source: Tracy AI Interviewer. Illustration with sample data* A candidate sees a notice first, talks, gets follow-ups and a person reviews the result. ## What a candidate sees, step by step The [participant experience page](/participant-experience) is the guide to what a candidate sees. In short: - **A browser link.** The candidate opens it and starts. There is no account to create and nothing to install. - **A notice on the join screen.** It says the conversation is recorded and transcribed. Nothing is hidden before the first question. - **A start on click, at any hour.** There is no scheduling and no recruiter present. The interview is ready in under 10 seconds, so a candidate can do it after work or on a weekend. - **Questions in their own field.** Tracy reads the CV when the CV check is on, asks about the work the person has done, follows up on what they said and waits for the answer. - **A way back in.** A candidate who drops can rejoin within 10 minutes, a limit the customer can configure. A reminder goes out after a day, and failed interviews are chased automatically. How the interview itself runs is covered in [how AI interviews work](/blog/how-do-ai-interviews-work), and the first-round format in [AI screening interviews](/blog/ai-screening-interviews). Three real screens follow: the device check before the camera and microphone are allowed, the same check when it is ready, and the page a candidate sees at the end. The end page says the answers are saved and that Tracy does not decide anything. ![Device check with camera and microphone off, Choose Allow, and a note that recording starts on join.](/blog/images/ai-interview-candidate-experience-3.png) *Screenshot of the Tracy AI Interviewer join flow, 9 Nov 2026. Source: Tracy AI Interviewer join flow, 9 Nov 2026.* Device check with camera and microphone off, Choose Allow, and a note that recording starts on join. ![Device check ready: blurred camera preview, microphone meter, Join the interview button, spot held 10 minutes.](/blog/images/ai-interview-candidate-experience-4.png) *Screenshot of the Tracy AI Interviewer join flow, 9 Nov 2026. Source: Tracy AI Interviewer join flow, 9 Nov 2026.* Device check ready: blurred camera preview, microphone meter, Join the interview button, spot held 10 minutes. ![End page: Thank you, you're all done, with a What happens next note.](/blog/images/ai-interview-candidate-experience-5.png) *Screenshot of the Tracy AI Interviewer join flow, 9 Nov 2026. Source: Tracy AI Interviewer join flow, 9 Nov 2026.* End page: Thank you, you're all done, and a What happens next note saying Tracy does not decide anything. ## What candidates say Steven Perry, Account Executive at IT Resources, an IT staffing firm in Tampa, US, describes the experience from the employer side: "By the time they finish with this platform, they're even laughing and joking." That is one customer's observation. ## What candidates should know before they start If you have been sent an AI interview link, these points answer most worries. - **It is recorded and transcribed.** The join screen tells you. Both sides are kept, so the employer can review what was said. - **A person decides.** The interview produces a report. A recruiter decides: they can move you forward or invite you to interview again, and the history stays on the record. - **Answer in your own words.** Follow-up questions come from what you said, so a specific example works better than a general claim. - **Check your connection first.** If you drop, rejoin within 10 minutes. The employer sets the limit, so read the invitation. - **You can ask the employer.** If something is unclear, such as how the recording is used, ask the person who sent the invitation. For the rules that apply to employers using automated tools, see [AI interview rules in 2027](/blog/ai-interview-rules-2027). ## What should employers check about candidate experience? A weak experience costs you completed interviews and the people you wanted. Before you send the link, test each point as a candidate would. The table lists each check, why it matters to the candidate and how Tracy handles it. | Check | Why it matters to the candidate | How Tracy handles it | | --- | --- | --- | | Join-screen notice | They know the conversation is recorded before they speak | The join screen says it is recorded and transcribed | | Access | An account or an install loses people before question one | A browser link, no account | | Timing | Evening and weekend candidates need a start that suits them | Starts on click, at any hour, no scheduling | | Dropped connection | A lost connection should not end the application | Rejoin within 10 minutes, configurable per customer | | Question quality | Generic questions feel like a form | Questions in the person's field, with follow-ups | | Who decides | Candidates want a person behind the result | A person reviews the evidence and decides | | Evidence | A result you cannot check cannot be defended | Evidence per result: recording, transcript, written assessment per question | Create an interview for one role and take it as a candidate would. ![Four checks: join-screen notice, link start at any hour, rejoin within 10 minutes, a person decides.](/blog/images/ai-interview-candidate-experience-2.png) *Tracy gives employers four things to check before they send the link. Illustration with sample data. Source: Tracy AI Interviewer product documentation.* Four checks cover the notice, the start, the way back and the decision before you send the link. ## Why a person deciding matters Tracy does the interview and the report. The customer decides, and a recruiter can reverse a rejection or invite a candidate back. Tracy publishes no bias audit, so the safeguard on offer is the evidence: a recording of both sides, a timestamped transcript and a written assessment per question with its reasoning, which a second reviewer can check. Integrity signals follow the same rule. Tracy captures tab changes, gaze, live face detection, delay before answers, fullscreen exits and external screens, and shows them as evidence for a person to interpret. They are not proof of anything and nothing is decided in the interview. See [AI interview cheating detection](/blog/ai-interview-cheating-detection) for how to read them, and [AI interview vs human interview](/blog/ai-interview-vs-human-interview) for what each does best. ## Candidate experience at volume Volume is where scheduling queues, silence after the application and long waits can hurt experience. A start on click removes the queue, and a candidate can finish while still interested. Which path fits you: - Hiring at volume: [volume hiring](/use-cases/volume-hiring) - Screening applications: [applications](/use-cases/applications) - Cost of the first interviews: [pricing](/pricing) For the steps around the interview, see [how to screen candidates](/blog/how-to-screen-candidates). ## Common questions ### Do candidates like AI interviews? It depends on the experience. A clear notice, an easy start, questions about their own work and a person deciding make a difference. ### Do job candidates dislike AI interviews? Some do, mostly when the process is opaque or rigid. Say that the conversation is recorded, let people start when it suits them, allow a rejoin, and make clear that a person decides. ### Is an AI interview recorded? With Tracy, yes. The join screen tells the candidate that the conversation is recorded and transcribed, and the employer reviews the recording, transcript and written assessment. ### What if a candidate's connection drops? They can rejoin within 10 minutes. The customer can configure that limit, and a reminder goes out after a day. ### Do candidates need an account or an app? No. They open a browser link and start, at any hour. ### Can candidates use AI in an interview? Tracy captures integrity signals such as tab changes, gaze, delay before answers and external screens, and shows them as evidence to a person. Nothing is decided in the interview. --- Create an interview for one role and take it as a candidate would. You get the interview plan, built in days, with a recording and report you can check. The first 3 results are free, with no card. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=ai-interview-candidate-experience&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=ai-interview-candidate-experience&utm_content=demo) ## Sources - [Tracy AI Interviewer home page, customer quotes](https://tracyinterview.ai/), retrieved 9 Nov 2026. - [Tracy AI Interviewer participant experience](https://tracyinterview.ai/participant-experience), retrieved 9 Nov 2026. - [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 5 Nov 2026. - [Tracy AI Interviewer Connect page](https://tracyinterview.ai/connect), retrieved 5 Nov 2026. # How to Prepare for an AI Interview (Nov 2026) > How to prepare for an AI interview: get your examples ready, set up your space and know what to ask. Tracy AI Interviewer is the worked example. - URL: https://tracyinterview.ai/blog/how-to-prepare-for-an-ai-interview - Published: 2026-11-05 - Updated: 2026-11-06 - Author: Tracy team > **In short** > - Preparing for an AI interview mostly means preparing specific examples. Tracy AI Interviewer waits for the end of your thought and asks for the missing detail, so an answer that already says what you did, why and what happened leaves less to chase. > - You start from a link in your browser, with nothing to install. Tracy has three ways in: an invitation link, an apply page and the employer's own system. If your connection drops you can rejoin the same session within 10 minutes. > - A person at the company reads the result and decides after the interview. Fact-checked on 5 Nov 2026 against Tracy's own pages. Sources are listed at the end. If you have been invited to an AI interview, the useful preparation is small: a few examples you can say out loud, a working camera and microphone, and a few questions for the employer. Tracy AI Interviewer is the worked example below. The tips are ours, not Tracy's. The last section is for people who send invitations. ## What you will see, step by step Tracy's [candidate page](https://tracyinterview.ai/participant-experience?utm_source=blog&utm_medium=seo&utm_campaign=how-to-prepare-for-an-ai-interview&utm_content=internal-link) shows the real screens for the invitation path, in the order a candidate meets them. Applicants who start from an apply page answer a few basic questions before the interview, and an employer's own system can hand a candidate the link. In short, on the invitation path: 1. **The invitation email.** It gives the length, the format and the last day to take the interview, and there is no sign-up. 2. **The welcome page.** Before anything starts, it says the interview is recorded and transcribed, who sees it, and that you can stop at any time. 3. **The device check.** The room tests your camera, your microphone and your connection. Recording starts when you join. 4. **The conversation.** Tracy asks the questions and follows up. If the room cannot hear you, it says so and offers Change mic. 5. **The last page.** It confirms your answers are saved and says what happens next. 6. **The host reads the result.** The host asks who finished and gets a link to the report. Video interviews add some prompts. Tracy asks for a full-screen window, checks for a face on camera, shows a pop-up if your face leaves the frame, asks you to disconnect external monitors and asks you to keep the tab open. Integrity signals are on unless turned off. Tracy reports what it sees there. It does not score or reject anyone; a person decides. What the person at the company reads is in [AI interview reports](/blog/ai-interview-reports?utm_source=blog&utm_medium=seo&utm_campaign=how-to-prepare-for-an-ai-interview&utm_content=internal-link), and [How AI interviews work](/blog/how-do-ai-interviews-work?utm_source=blog&utm_medium=seo&utm_campaign=how-to-prepare-for-an-ai-interview&utm_content=internal-link) explains the format. ## Before the day 1. **Read the job post twice.** Underline the three to five things it asks for most. 2. **Pick three examples from your own work** that show those things. For each, write down what you did, why you chose that approach and what happened, with a number if you have one. 3. **Say them out loud.** The interview is spoken, and an example that reads well on paper can come out vague when you say it. Record yourself on your phone and listen for the places where you go general. 4. **Prepare for the follow-up, not only the question.** For each example, ask what a curious person would ask next: why that approach, what it cost, what you would change. [AI behavioral interviews](/blog/ai-behavioral-interviews?utm_source=blog&utm_medium=seo&utm_campaign=how-to-prepare-for-an-ai-interview&utm_content=internal-link) has more on rehearsed stories. Tracy waits for the end of your thought, about half a second on simple questions and up to about two seconds on hard technical ones, and asks for the missing detail. 5. **If the role is technical, practise a trade-off.** Tracy can introduce a coding task and probe the trade-offs, so practise explaining a choice aloud while you work. 6. **Check the language.** The interview runs in one language, chosen at setup. If your invitation does not say which, ask. 7. **Set up your space.** The device check tests the camera, the microphone and the connection, so have all three ready, in a quiet room. Close the apps and tabs you do not need, and open the link a few minutes early. ## During the interview **Answer with specifics.** Names of tools, sizes, trade-offs, what you would change. A short illustration, written by us and not taken from a real interview: > Tracy: Tell me about a project that slipped. > Candidate (general): We had some delays, but we managed it in the end. > Tracy (follow-up): What caused the delay, and what did you change to recover? > Candidate (prepared): A vendor API change cost us two weeks. I moved the integration work earlier and cut one report from the scope, so we shipped five days late instead of two weeks. **Think out loud rather than going quiet (our advice).** Tracy allows longer pauses on hard questions. A long pause can also be marked as a signal for the reviewer: the example on Tracy's product page marks 14 seconds before a polished answer. The pages we checked give no threshold. **Stay in the interview tab, on one screen (our advice).** The candidate screens above ask you to stay in fullscreen, disconnect external monitors and keep the tab open, and say that switching apps or tabs is flagged as unusual activity. Signals are marked on the recording for a person to review. [AI interview cheating detection](/blog/ai-interview-cheating-detection?utm_source=blog&utm_medium=seo&utm_campaign=how-to-prepare-for-an-ai-interview&utm_content=internal-link) explains what signals are and are not. **Ask if you are unsure how something works (our advice).** You can ask Tracy about the process. Tracy notes requests to repeat a question as one of its signals for the reviewer. ## If something goes wrong Connections drop. With Tracy you can rejoin the same session within 10 minutes. If an interview is left unfinished, a reminder goes out after a day. If you cannot get back in, tell the employer who invited you and say when it happened. ## Questions to ask the employer An invitation may not answer all of these, so ask before you start: - What is the interview for: a first screen, or part of the final decision? - How long is the recording kept? - What happens next, and by when will I hear back? - Who makes the decision? - What is the alternative if I cannot do a video interview, or need an adjustment? Qode's privacy policy gives indicative retention periods by type of data. Its row for customer-managed candidate data says per customer instructions, and no row names interview recordings, so the employer is the one to ask. ## If you are the one sending invitations If you invite by email, Tracy's invitation already gives the length, the format and the last day. Add what only you know: - What the interview is for, and who reads the result and decides. - How long you keep the recording. - Who to contact if something goes wrong. For the other half of the invitation, [AI screening interviews](/blog/ai-screening-interviews?utm_source=blog&utm_medium=seo&utm_campaign=how-to-prepare-for-an-ai-interview&utm_content=internal-link) covers what to tell candidates about the AI interviewer. ## Common questions ### How do I prepare for an AI interview? Pick three examples from your own work and practise them out loud. Give what you did, why and what happened in each answer, since Tracy asks for the missing detail. Set up your camera, microphone and connection before you open the link. ### What should I practise first? Your examples, out loud. Writing them down is not the same as saying them, and the interview is spoken. ### How do follow-up questions work? Tracy waits for the end of your thought and asks for the missing detail. An answer that already says what you did, why and what happened leaves less to ask for. ### What will I see before the interview starts? On the invitation path with Tracy, an invitation email that gives the length, the format and the last day, a welcome page that says the interview is recorded, and a device check for the camera, the microphone and the connection. Recording starts when you join. If you start from an apply page, you answer a few basic questions first. ### What if my connection drops during the interview? With Tracy you can rejoin the same session within 10 minutes. If an interview is left unfinished, a reminder goes out after a day. ### Will anyone see my recording? A person at the company reads the result, with the recording and a transcript behind it. Tracy's welcome page says up front that the interview is recorded and who sees it. Ask the employer how long they keep it. Qode's privacy policy gives indicative retention periods by type of data. Its row for customer-managed candidate data says per customer instructions, and no row names interview recordings. --- If you send interviews rather than take them, **[Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=how-to-prepare-for-an-ai-interview&utm_content=try-free)** and run one yourself first, or [book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=how-to-prepare-for-an-ai-interview&utm_content=demo). ## Sources All sources retrieved 5 Nov 2026. - Tracy AI Interviewer, [Home page](https://tracyinterview.ai/): each person starts when they are ready, at any hour, in the browser; Tracy listens, waits for the end of the thought and asks for the missing detail; signals marked on the recording for a person to review, nothing decided during the interview. - Tracy AI Interviewer, [Product](https://tracyinterview.ai/product): no install for the person being interviewed; a two-way video interview where Tracy speaks, listens and answers questions about the process; recorded on both sides with a transcript timed to the recording; a wait of about half a second on simple questions and up to about two seconds on hard technical ones; a "Long pause" signal example of 14 seconds; rejoin the same session within 10 minutes; a reminder after a day for unfinished interviews; one interview language chosen at setup; a coding task for technical roles; repeated requests for a question noted as a signal; "Tracy does not reject anyone. A person decides." - Tracy AI Interviewer, [What candidates see](https://tracyinterview.ai/participant-experience): the invitation email with length, format and last day and no sign-up; the welcome page saying the interview is recorded, who sees it and that candidates can stop at any time; the device check; the conversation and the Change mic prompt; the fullscreen, second screen, face on camera and tab prompts; integrity signals on unless turned off; Tracy only reports what it sees and never scores or rejects a participant; the last page. - Tracy AI Interviewer, [Hiring at volume](https://tracyinterview.ai/use-cases/volume-hiring): if people drop they can rejoin within 10 minutes, and a reminder goes out after a day. - Qode, [Privacy policy](https://tracyinterview.ai/privacy): indicative retention periods by type of data; the row for customer-managed candidate data says per customer instructions. The table has no row for interview recordings. # How to Screen Hundreds of Applicants Without a Recruiter on Every Call > Four ways to screen hundreds of applicants for high-volume roles, what to check in any tool, when a simple method is enough, and a four-step Tracy setup from a job description. - URL: https://tracyinterview.ai/blog/how-to-screen-hundreds-of-applicants - Published: 2026-11-05 - Updated: 2026-11-11 - Author: Tracy team Fact-checked on 4 Nov 2026 against Tracy's public pages and the survey listed in Sources. > **In short** > - **Tracy AI Interviewer interviews every applicant who clicks the link, at any hour with no scheduling,** reads each CV to tailor questions and follows up on what each person says, so you cut on evidence instead of at the CV. > - Four ways to screen at volume: a CV filter, phone screens, one-way video or forms, and an AI two-way interview. Each has a size where it is enough. > - Eight questions to put to any tool before you buy are listed below. When hundreds of people apply for hourly, staffing or high-volume roles, the first round is where the time goes. A recruiter cannot phone everyone, so most teams cut at the CV and call whoever is left. Tracy's [volume-hiring page](https://tracyinterview.ai/use-cases/volume-hiring) describes the result: recruiters call a fraction of the applicants, so the shortlist depends on who applied first. ## Why does shortlisting by CV lose people? A CV is a good first filter for requirements that are facts: a licence, a location, a start date. It is a weak filter for how well someone can do the work, because it shows what a person chose to write down. Two applicants with the same CV can answer the same question very differently, and the CV does not tell you which one you are looking at. Volume makes it worse. Robert Half reported that in a survey fielded in November 2025, 67% of US HR leaders said reviewing AI-generated applications had slowed their hiring. The survey measured opinion across more than 2,000 US hiring managers, the exact question is not published, and it is about reviewing applications, not about screening interviews. The pressure is plain: more applications and the same recruiter hours. When the pile is too big, teams cut it by rule of thumb: keywords, school names, the first fifty received. Each of those rules drops some people who could have done the job, and nobody finds out who. A first round that gives each applicant a real conversation removes the need to guess. ## What are the four ways to screen at volume? | Method | What every applicant gets | Where it breaks | | --- | --- | --- | | CV filter | Nothing: a person reads the CV, or a rule filters it. | Shows what was written, not how the person thinks. Rules drop people silently. | | Phone screens | A short call with a recruiter. | A recruiter can call only a fraction of a large pile, and scheduling takes time. | | One-way video or forms | The same fixed questions, answered when convenient. | No follow-up, so a vague answer stays vague. Someone still has to watch or read every response. | | AI two-way interview | A conversation with an interviewer that listens and asks follow-ups. | Needs applicants who agree to be recorded, and a person still reviews the evidence and decides. | These are our descriptions of how each method usually works, not a ranking. Most teams combine them: a knockout form for hard requirements first, then one conversation method for everyone who passes. For a longer comparison of the interview formats, see [one-way versus two-way video interviews](/blog/one-way-vs-two-way-video-interviews) and [how to screen candidates](/blog/how-to-screen-candidates). ## When is a simpler method enough? You do not need an interview tool for every hiring problem. - **A small pool.** If a recruiter can call every applicant in a day or two, phone screens are fine and a tool adds setup for little gain. - **Hard requirements only.** If the decision comes down to a licence, a shift pattern or the right to work, a short form with knockout questions answers it faster than any interview. - **A fixed script and no judgement needed.** If you need the same five factual answers from everyone, a one-way video or a form collects them, and a person can skim the results. - **A role where the test is doing the work.** For a trade or a physical skill, a short practical test tells you more than any conversation. The simple methods start to fail when the pile is large, the answers need to be read for meaning, and the people cut at the CV are the ones you would have wanted to hear from. ## What should you check in any screening tool? Put these eight questions to any vendor, Tracy included. 1. **Every applicant is interviewed, not a sample.** Ask what happens to applicant number 400. 2. **Any hour, no scheduling.** Applicants for hourly work often apply in the evening or at the weekend. 3. **Questions that fit the role.** Can you start from the job description, edit the questions and bring your own? 4. **Follow-ups.** Does the next question depend on the last answer, or is it a fixed list read out? 5. **Evidence a reviewer can check.** Is there a recording, a transcript tied to it, and a written reason for each assessment? 6. **Integrity signals as signals.** Any signal of looking away or switching tabs is a prompt for a person to look, not proof of anything. 7. **Connection to your stack.** Can it create and read interviews from the tools your team already uses, by API or MCP? 8. **A price model you can predict.** Per result, per seat or per month, and what is charged when an applicant does not show up. For a market view of vendors against these points, see [AI interview tools for high-volume hiring](/blog/ai-interview-tools-for-high-volume-hiring). To compare Tracy with one named alternative, see [Tracy vs HeyMilo](/compare/tracy-vs-heymilo). ## How does Tracy AI Interviewer screen hundreds of applicants? Tracy AI Interviewer runs a two-way video interview for each applicant on its own. Taking the eight questions in order: - **Every applicant, any hour.** Each applicant starts when they click the link, at any hour, with no recruiter and no calendar. Tracy's page describes the same depth for applicant one and applicant five hundred. We write this as volume over a hiring round, and we do not give a figure for how many interviews can run at once. - **Questions from the job description.** Tracy drafts questions from the job description or your brief. You can edit them, replace them with your own and set the style from quick to deep and from behavioural to technical. - **A CV check.** Tracy can read each CV and add 3 to 5 questions tailored to it and to the job's must-have requirements, on top of your questions. - **Follow-ups that wait.** Each follow-up depends on the answer before it. Tracy waits for the end of a thought before it speaks, with longer pauses allowed on hard questions, so an applicant who is thinking is not cut off. For technical roles Tracy can introduce a coding task and probe the trade-offs. - **Evidence to check.** Both sides are recorded, with a transcript timed to the recording. Each question gets a written assessment with the reasoning behind it, so a reviewer reads the applicant's own words and decides. Our guide to [AI interview reports](/blog/ai-interview-reports) shows what to expect. - **Integrity signals.** Tracy records signals such as where the applicant looked, response timing, tab switches, leaving fullscreen and an external screen, and gives the reviewer the evidence. A signal is not a verdict, and our post on [AI interview cheating detection](/blog/ai-interview-cheating-detection) explains the limits. - **Your stack.** Tracy connects through an API and an MCP server, so interviews can be created and read from tools a team already uses. You can start an interview from a personal invite link, an apply page or your own system. [Tracy with Workable](/blog/tracy-workable-integration) is one worked example, and enterprise teams can ask for a customised setup. - **Price model.** You pay for results, not seats. Your first 3 results are free, then it is $1 per credit, one credit per started 20 minutes. A session with no answers is never charged. Enterprise pricing is set to your volume. See [AI interview cost](/blog/ai-interview-cost) for how prices compare across tools. The interview is recorded on both sides, and Tracy does not reject anyone: a person decides. For how an interview runs from the applicant's side, read [how AI interviews work](/blog/how-do-ai-interviews-work), and for the wider category, [AI screening interviews](/blog/ai-screening-interviews). ## What does a Tracy setup look like in four steps? This example is a high-volume role, written by us as an illustration, not a customer story. It follows the steps on Tracy's volume-hiring page. 1. **Start from the job description.** Paste the posting or a short brief for a warehouse team lead. Tracy drafts the questions. Edit them, add two of your own about shift handling, and turn on CV check so each applicant also gets 3 to 5 questions from their own CV. 2. **Invite everyone, or link from your apply page.** Send the link to every applicant, or put it on the apply page so people start when they are ready. If an applicant's connection drops, they can rejoin within 10 minutes, and a reminder goes out after a day. What the applicant sees, and how they can prepare, is in [How to prepare for an AI interview](/blog/how-to-prepare-for-an-ai-interview?utm_source=blog&utm_medium=seo&utm_campaign=how-to-screen-hundreds-of-applicants&utm_content=internal-link). 3. **Let applicants interview when they choose.** Evening and weekend applicants start without waiting for a slot. Nobody on your team joins the call. 4. **Read results as they land.** Open each result: an assessment per question, the transcript and the recording behind it, and any integrity signals with their evidence. Review the interviews in the order you like, then decide who moves on. Start with the three free results on a real role before you commit to a full round. ## Where is Tracy the wrong tool? - **Applicants who have not agreed to be recorded.** Tracy runs recorded video interviews, and it is not for talking to someone who has not agreed to be recorded. Use a call. - **A short list of hard requirements.** If a form can decide, a form is cheaper and quicker. - **A small pool you can phone yourself.** A recruiter who can call everyone in two days gains little from a tool. - **A hands-on skill test.** If the proof is doing the work, set a practical task and use the interview for the conversation about it. - **A phone line or text chat.** Tracy is a video interview, not a phone line or a text chatbot. - **A final decision.** Tracy writes up what people said, and the decision stays with you. ## Common questions ### Can you screen hundreds of applicants without cutting at the CV? Yes, if the first round does not need a person on every call. Options are a short knockout form for hard requirements, one-way video or forms for a fixed set of questions, or an AI interviewer that talks with each applicant and follows up. Pick by how much you need to learn before a person gets involved. ### What is the fastest way to screen high-volume applicants? The fastest first round is one every applicant can start by themselves, with nothing to schedule. A knockout form does that for yes-or-no requirements. For anything that needs a real answer, an on-demand interview such as Tracy AI Interviewer does it without a calendar, and a person reads the evidence afterwards. ### Is it fair to shortlist by CV when hundreds apply? It is quick, but a CV tells you what someone wrote, not how they explain it, and a very large pile is often cut by order of arrival or keyword. If you cut by CV, write down the rule first and keep a small sample of rejected CVs to check it. ### What should I check before choosing a screening tool for volume? Check that every applicant gets an interview, at any hour, with no scheduling; that questions fit the role; that follow-ups depend on the answer; that a reviewer can check the evidence; that integrity signals are signals, not verdicts; that it connects to your stack; and how it charges. ### When is Tracy not the right tool for screening? When applicants have not agreed to be recorded, when you only need a yes-or-no on a few requirements, when the pool is small enough for a recruiter to call, or when the real test is a hands-on demonstration. Tracy runs recorded two-way video interviews, and a person makes the decision. --- **[Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=how-to-screen-hundreds-of-applicants&utm_content=try-free)** on a role with hundreds of applicants, or [book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=how-to-screen-hundreds-of-applicants&utm_content=demo). ## Sources All sources retrieved 4 Nov 2026. - Tracy AI Interviewer, [Hiring at volume page](https://tracyinterview.ai/use-cases/volume-hiring): interviews started on click at any hour, no calendar, the volume problem, CV check, rejoin and reminder, results per question, the same depth for each applicant. - Tracy AI Interviewer, [product page](https://tracyinterview.ai/product): setup from a brief, three ways to start, two-way video, follow-ups that wait, coding task with a probe, assessment per question, integrity signals, API and MCP. - Tracy AI Interviewer, [home page](https://tracyinterview.ai/): interviews that follow up in the person's field, evidence behind each result. - Tracy AI Interviewer, [pricing](https://tracyinterview.ai/pricing): 3 free results, $1 per credit per started 20 minutes, no seats, no-answer session not charged, enterprise set to volume with API and MCP. - Tracy AI Interviewer, [Connect page](https://tracyinterview.ai/connect): MCP server address and the AI apps and ATS guides. - Robert Half, [Survey: 67% of HR leaders report AI-generated applications are slowing hiring](https://press.roberthalf.com/2026-03-10-Robert-Half-survey-67-of-HR-leaders-report-AI-generated-applications-are-slowing-hiring), 10 Mar 2026: survey of more than 2,000 US hiring managers, fielded November 2025; the exact question is not published. # AI Behavioral Interviews: STAR Follow-Ups for Every Applicant > How an AI interviewer runs a behavioral interview: the STAR questions, the follow-ups that test a rehearsed story, and how to read the answers against the recording and transcript. - URL: https://tracyinterview.ai/blog/ai-behavioral-interviews - Published: 2026-11-04 - Updated: 2026-11-11 - Author: Tracy team > **In short** > - A STAR answer is easy to rehearse, so the second and third question carry most of a behavioral interview. Tracy AI Interviewer asks them for every applicant, at any hour, and each follow-up depends on the answer before it. > - You write the behavioral questions or let Tracy draft them from your brief. When a story sounds too tidy, replay the exact moment: the transcript sits beside the video, and each assessment point opens the part of the recording it rests on. > - The research on structured interviews is about human-run ones. Use it for the design rules, not as a claim about AI accuracy. Fact-checked on 4 Nov 2026 against Tracy's own pages and the original research summaries. Sources are listed at the end. Behavioral interviewing asks people what they did, not what they would do. The common frame is STAR: the situation, the task, the action and the result. Candidates learn it, and a prepared story has all four parts. This page covers what STAR misses, the follow-ups that test a story, and how to run them with [Tracy AI Interviewer](https://tracyinterview.ai/) for every applicant, not only the few a recruiter has time to call. ## What STAR misses A STAR answer gives you a shape. It does not tell you whether the story is the candidate's own. Three gaps to look for: - **Whose action was it?** "We" covers a lot of ground. The answer may be a team's work with the candidate on the edge of it. - **What did it cost?** Tidy stories end in a result with no trade-off, though real projects usually closed off another choice. - **What would they change?** Ask for it: people who did the work usually have something they would redo. These are our observations about how interviews go, not a measured finding. They are why a first round works better as a conversation than as a form. [One-way and two-way video interviews](/blog/one-way-vs-two-way-video-interviews) explains why a recording cannot ask the second question. ## Follow-ups that test a story Pick follow-ups that ask for something only the person who did the work would know. | Gap | Follow-up | | --- | --- | | Whose action | "What did you personally decide?" | | The trade-off | "What did you give up by doing it that way?" | | The disagreement | "Who disagreed, and what did they want instead?" | | The failure | "What went wrong, and when did you notice?" | | The result | "How would you know it worked? What was the number or the signal?" | | Hindsight | "What would you do differently now?" | Tracy's product page shows the same move in an example from its site, which uses invented people: a data engineer says reads were the bottleneck, and Tracy asks, "How much faster were reads, and what did it cost you on writes?" ## An example chain This is an illustration we wrote, not a transcript of a real candidate. > Interviewer: Tell me about a time you handled a difficult stakeholder. > Candidate: Our head of sales wanted a feature shipped before it was ready. I talked it through and we reached a compromise. > Interviewer: What did you personally say that changed their mind? > Candidate: I showed them the support tickets from the last release. > Interviewer: What did the compromise cost you? > Candidate: We shipped half the feature and lost two weeks of the roadmap. > Interviewer: How did you know the half release was the right call? The chain asks for a source, a cost and a test, and each answer lands on the transcript. ## Running it with Tracy 1. **Set the goal and the questions.** Start from the job description or your own behavioral questions. Tracy drafts questions that you can edit or replace. 2. **Pick the language.** Each interview is a two-way video interview in one language, chosen at setup. 3. **Invite everyone.** Send a personal invite, publish an apply page, or let your own system hand over the link. Candidates have nothing to install and start whenever it suits them, so every applicant gets the same follow-ups, not only the few a recruiter can call. People are told about the recording before it starts. 4. **Let the follow-ups run.** Tracy listens for what was said, asks for the missing detail, and answers questions about the process. It waits for the end of a thought, so a candidate who is working out a hard example is not cut off. With CV check on, it also asks 3 to 5 questions tailored to the CV, which tests the stories against what the CV claims. 5. **Open the result.** The assessment for each question comes with its reasoning, and the recording plays beside a transcript timed to it. Response timing and tab switches show up as marks on the recording, for a person to check. [How a first round works with an AI interviewer](/blog/ai-screening-interviews) covers the steps in more detail, and [software engineer interview questions](/blog/software-engineer-interview-questions) shows how follow-ups sit alongside technical questions. ## Reading the answers Write down what a good answer contains before you send the interview: for example, the candidate's own action, a trade-off and a result. Then read each assessment against the transcript. Each point jumps to the moment it came from, so you can check it without replaying the whole interview. [AI interview reports](/blog/ai-interview-reports) goes through what a reviewer should look for. Tracy does not reject anyone: a person decides. ## What the research says Two research summaries are worth knowing, and both describe human-run interviews. A 1998 meta-analytic summary in Psychological Bulletin (Schmidt and Hunter) put the average validity of structured interviews for predicting job performance at .51, compared with .38 for unstructured interviews. Validity here is a correlation, not a percentage of correct predictions. The figures are statistically corrected estimates from older studies, and a 2022 re-analysis argues that such corrections for range restriction can overcorrect. A 2022 re-analysis in the Journal of Applied Psychology (Sackett and colleagues) found structured interviews to be the top-ranked selection procedure for predicting job performance. The authors also report that most of the procedures that ranked high in earlier summaries remain high in rank, with mean validity estimates reduced by .10 to .20 points. That points to a design rule: ask every applicant the same questions and read the answers against the same evidence. The research is silent on AI interviewers, so it proves nothing about their accuracy. ## When to use people instead - **Senior roles where the interview is also a negotiation.** A person reads the room and sells the role. - **A final-round conversation with the team.** Fit with a team is judged by the team. - **Candidates who have not agreed to be recorded.** Tracy records every interview and is not for them. ## What it costs The first 3 results are free. After that it is $1 per credit, charged per started 20 minutes for interviews that produced a result, and a session where nobody answers is never charged. There are no seats. If you want to see the options for the first round before you decide, [how to screen candidates when there are too many applicants](/blog/how-to-screen-candidates) is the place to start. ## Common questions ### What is a behavioral interview? An interview that asks candidates to describe what they actually did in past situations, usually structured as situation, task, action and result (STAR). ### Can an AI run a behavioral interview? Yes. Tracy AI Interviewer asks your questions or ones it drafts from your brief. Each follow-up depends on the answer before it, and a vague answer gets a sharper question. ### Why are follow-ups needed on STAR answers? A STAR answer is easy to prepare. The follow-up asks for what a prepared answer leaves out: the decision, the disagreement, the cost, the part the candidate did not do. ### Does research support structured interviews? A 1998 meta-analytic summary put the average validity of structured interviews at .51 against .38 for unstructured ones, and a 2022 re-analysis ranked structured interviews first among selection procedures. Both concern human-run interviews. ### What does the reviewer get after the interview? A written assessment for each question with its reasoning, a transcript timed to the recording, and integrity signals with their evidence. A person reads it and decides. ### How do I start? Start from the job description or your own behavioral questions, and Tracy drafts questions you can edit. The first 3 results are free. --- **[Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=ai-behavioral-interviews&utm_content=try-free)** and run a behavioral interview with follow-ups on your first candidates at no cost, or [book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=ai-behavioral-interviews&utm_content=demo) to see one worked through. ## Sources All sources retrieved 4 Nov 2026. - Tracy AI Interviewer, [Home page](https://tracyinterview.ai/): the follow-up example from the site (invented people), integrity signals, first 3 results free. - Tracy AI Interviewer, [Pricing](https://tracyinterview.ai/pricing): free results, $1 per credit, billing, no-answer sessions, no seats. - Tracy AI Interviewer, [Product](https://tracyinterview.ai/product): questions drafted from a brief, CV check, three ways in, no install, people told about the recording before it starts, one interview language chosen at setup, follow-ups that depend on the answer before it, waiting for the end of a thought, points linked to the recording, transcript beside the video, "Tracy does not reject anyone. A person decides.". - Frank L. Schmidt and John E. Hunter, [The Validity and Utility of Selection Methods in Personnel Psychology](https://doi.org/10.1037/0033-2909.124.2.262), Psychological Bulletin, 1998: structured interview validity .51 versus .38. - Sackett, Zhang, Berry and Lievens, [Revisiting meta-analytic estimates of validity in personnel selection](https://doi.org/10.1037/apl0000994), Journal of Applied Psychology, 2022: structured interviews top-ranked. # AI Interview vs Human Interview: What Each Is Good At > What the field evidence says about AI and human interviewers, where each is the better choice, and how to combine them: AI for the first conversation, people for the decision. - URL: https://tracyinterview.ai/blog/ai-interview-vs-human-interview - Published: 2026-11-04 - Updated: 2026-11-11 - Author: Tracy team > **In short** > - Use an AI interviewer for the first conversation with every applicant, and people for the later rounds and the decision. Tracy AI Interviewer is built for that split: everyone who applies is interviewed, at any hour, with follow-ups on what was said. > - The best field evidence is narrow. One experiment found more job offers with an AI voice agent in entry-level call-centre hiring in the Philippines, with humans deciding. It does not cover skilled roles, and it was not about Tracy. > - People are better where the job of the interview is judgment, rapport or the final call. Fact-checked on 4 Nov 2026 against the original papers, surveys and Tracy's own pages. Sources are listed at the end. An interview has several jobs: gathering the same evidence from everyone, probing what a person says, judging fit with a team, and making a decision. AI and people do those differently. This page sets out what the evidence supports for each, keeps what people think apart from what was measured, and says where a human interviewer is the better pick. For the basics of the AI side, see [how AI interviews work](/blog/how-do-ai-interviews-work). ## What the evidence says ### Measured outcome: one field experiment The strongest study we found is a natural field experiment by Brian Jabarian (Carnegie Mellon University) and Luca Henkel (Erasmus University Rotterdam), a working paper, read here in its arXiv version 2 of 11 September 2026. It covers about 70,000 entry-level customer service applications in the Philippines between March and June 2025. Applicants interviewed by an AI voice agent received job offers 12% more often than those interviewed by a human recruiter: 9.73% versus 8.70%. Human recruiters made every hiring decision. What to keep in mind: - **It was a voice agent**, not video, and not Tracy. - **It was high-volume, entry-level hiring in one country.** It says little about skilled or technical roles. - **It is a working paper**, not shown to be peer reviewed. The paper discloses that on 1 August 2025, after data collection ended, Brian Jabarian accepted an unpaid role as Chief Economist at PSG Global Solutions, the recruitment firm that supplied the data. ### Measured outcome: what applicants chose In the same experiment, when applicants were allowed to choose, 78% picked an AI voice agent over a human recruiter. It shows what applicants chose in one setting and country; it does not measure general preference. The paper also notes that applicants who choose the AI voice agent have lower language and analytical scores than those who choose a human recruiter. ### Measured outcome: who finished Not every applicant finishes. In a preprint (arXiv version 2, 7 August 2026) on one Junior Frontend Engineer posting that drew 34,493 unique applicants, about one in four of the 25,536 applicants assigned to the AI-interview track (25.07 percent) completed a 30 to 40 minute AI interview. That is a completion rate for invited applicants to one posting, on one platform, from a preprint not shown to be peer reviewed. Two of its four authors are affiliated with micro1, the platform that built the AI interviewer being studied. The paper links the lower completion to the 30 to 40 minute length, and the figure cannot be read as a general completion rate for AI interviews. ### Perception: what people say they want In a Pew Research Center survey of 11,004 US adults in December 2022, 71% opposed AI making a final hiring decision, 7% favored it and 22% were not sure. The survey asked in 2022 about the final decision, so it says little about first-round interviews. It is still a good reason to keep a person on the decision. ### The yardstick for any interview A 1998 meta-analytic summary in Psychological Bulletin (Schmidt and Hunter) put the average validity of structured interviews for predicting job performance at .51, compared with .38 for unstructured interviews. Validity here is a correlation, not a percentage of right calls, and the .51 describes human-run structured interviews, not AI interviewers. A 2022 re-analysis in the Journal of Applied Psychology (Sackett and colleagues) found structured interviews to be the top-ranked selection procedure for predicting job performance. The authors also report that most of the procedures that ranked high in earlier summaries remain high in rank, with mean validity estimates reduced by .10 to .20 points. The result supports asking everyone the same questions and reading answers against the same evidence. It gives no ground for calling an AI interview accurate because a human-run structured interview was. ## What each does well | Job of the interview | AI interviewer (Tracy) | Human interviewer | | --- | --- | --- | | The same first conversation for everyone | Yes: every applicant gets a consistent first round at the same depth, and a conversation that adapts to their answers, whenever they join | Hard to keep identical across many people | | Follow-up on a vague answer | Yes: each follow-up depends on the answer before it, and a vague answer gets a sharper question | Yes, and can read a room | | Written result with evidence | Yes: a written assessment per question with its reasoning, beside a transcript timed to the recording | Depends on the interviewer's notes | | Reading the CV and tailoring questions | Can: with CV check on, 3 to 5 questions tailored to the CV and the role's must-haves | Yes, as far as the interviewer's time allows | | Reach across everyone who applies | Yes: each applicant starts on click, at any hour, with no scheduling | Limited by calendars | | Integrity evidence | Signals such as tab switching or an external screen, marked on the recording for a person to review | By observation, not recorded evidence unless notes are kept | | Fits the tools you use | API and MCP connections, three ways to start an interview, a customised setup for enterprise teams | Depends on the team's process | | Fit with a specific team | Not the first round's job | Yes | | Rapport and selling the role | Not its job | Yes | | The final decision | No: Tracy does not reject anyone, and a person decides | Yes | The table is our assessment of the jobs, with Tracy's column resting on its own published description. The human column is general reasoning, not a measured comparison. ## Where AI falls short - **Skilled roles are not covered by the best evidence.** The experiment above was entry-level. If you hire engineers or experts, run your own comparison on a small batch, and read the results next to a human interview. [AI technical interviews](/blog/ai-technical-interviews) covers what a technical first round can and cannot show. - **People who have not agreed to be recorded.** Tracy records every interview with a transcript, tells people about the recording before it starts, and is not for talking to someone who has not agreed. - **Rapport.** A conversation meant to sell the role or judge team fit belongs to a person. ## Where people fall short - **Consistency at volume.** Ten interviewers asking ten different sets of questions give you ten different data sets. - **Reach.** A recruiter's calendar limits how many people get a conversation, which is the problem [how to screen candidates when there are too many applicants](/blog/how-to-screen-candidates) is about. These two points are our reasoning, not measurements. ## How to combine them 1. **AI for the first conversation.** Set the questions around the must-haves of the role. With Tracy you start from the job description or bring your own questions, and candidates join from a link in the browser with nothing to install, so the first round can cover everyone, not a shortlist. 2. **A person reads the results.** Each interview ends in a written assessment per question with its reasoning, with the recording, transcript and integrity signals behind it. [AI interview reports](/blog/ai-interview-reports) covers how to review them quickly. 3. **People run the later rounds and decide.** Keep the decision with a person, as most US adults in the 2022 Pew survey wanted. [How a first round works with an AI interviewer](/blog/ai-screening-interviews) walks through the steps. ## What it costs With Tracy the first 3 results are free, then $1 per credit, charged per started 20 minutes when the result is ready. A session with no answers is never charged. There are no seats. ## Common questions ### Is an AI interview better than a human interview? It depends on the job of the interview. In one field experiment, applicants interviewed by an AI voice agent received offers 12% more often than those interviewed by a human recruiter, but that was entry-level call-centre hiring in the Philippines, and humans made every decision. It does not show AI is better in general. ### What is an AI interview best at? Giving every applicant a consistent first conversation, at a time that suits them, with a written result to read afterwards. ### What are human interviewers best at? The decision, the later rounds, and any conversation that depends on rapport or judgment about a particular team. ### Do candidates prefer AI or human interviewers? When applicants to entry-level customer service jobs in the Philippines were allowed to choose, 78% picked an AI voice agent, in a 2025 field experiment. It shows what applicants chose in one setting and says nothing general about preference. ### Should AI make the hiring decision? In 2022, most Americans said no. In a Pew Research Center survey of 11,004 US adults in December 2022, 71% opposed AI making a final hiring decision. With Tracy AI Interviewer a person decides. ### Can I use both? Yes, and the evidence you can lean on points that way: an AI interview for the first conversation with everyone, then people for the later rounds and the decision. --- **[Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=ai-interview-vs-human-interview&utm_content=try-free)** and interview your first applicants at no cost, or [book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=ai-interview-vs-human-interview&utm_content=demo) to see how a result reads. ## Sources All sources retrieved 4 Nov 2026. - Tracy AI Interviewer, [Home page](https://tracyinterview.ai/): follow-ups, "You read the words, then you decide", integrity signals, "Teams can also connect through the API". - Tracy AI Interviewer, [Pricing](https://tracyinterview.ai/pricing): free results, $1 per credit, billing, no-answer sessions, no seats, Enterprise with a customised setup and API and MCP connections. - Tracy AI Interviewer, [Product](https://tracyinterview.ai/product): questions from the job description or your own, CV check, three ways in, no install, people told about the recording before it starts, recorded with a transcript timed to the recording, follow-ups that depend on the answer before, assessment per question, integrity signals, "Tracy does not reject anyone. A person decides.". - Tracy AI Interviewer, [Hiring at volume](https://tracyinterview.ai/use-cases/volume-hiring): each applicant starts on click, at any hour, with no calendar; a consistent first round at the same depth for applicant one and applicant five hundred. - Brian Jabarian and Luca Henkel, [Voice AI in Firms: A Natural Field Experiment on Automated Job Interviews](https://arxiv.org/abs/2607.28222), arXiv working paper, version 2, 11 September 2026: offers 9.73% versus 8.70%, 78% choice, the author disclosure. - Ada Aka, Emil Palikot, Ali Ansari and Nima Yazdani, [Better Together: Quantifying the Benefits of AI-Assisted Recruitment](https://arxiv.org/abs/2507.08029), arXiv preprint, version 2, 7 August 2026: 25.07 percent of invited applicants completed a 30 to 40 minute AI interview; two authors are affiliated with micro1. - Frank L. Schmidt and John E. Hunter, [The Validity and Utility of Selection Methods in Personnel Psychology](https://doi.org/10.1037/0033-2909.124.2.262), Psychological Bulletin, 1998: structured interview validity .51 versus .38. - Sackett, Zhang, Berry and Lievens, [Revisiting meta-analytic estimates of validity in personnel selection](https://doi.org/10.1037/apl0000994), Journal of Applied Psychology, 2022: structured interviews top-ranked. - Pew Research Center, [Americans' views on use of AI in hiring](https://www.pewresearch.org/internet/2023/04/20/americans-views-on-use-of-ai-in-hiring/), chapter 1 of AI in Hiring and Evaluating Workers: What Americans Think, 20 April 2023: 71% oppose AI making a final hiring decision. # AI Screening Interviews: How a First Round Works With an AI Interviewer > What happens in an AI screening interview, from invite to report, and what a good first round finds out. Tracy AI Interviewer is the worked example. - URL: https://tracyinterview.ai/blog/ai-screening-interviews - Published: 2026-11-04 - Updated: 2026-11-11 - Author: Tracy team > **In short** > - Tracy AI Interviewer interviews everyone who applies, not a shortlist: each person starts when they click the link, at any hour, with no scheduling. It can read each CV and tailor questions to it, and it follows up on what people actually said. > - A result comes with the video recording, a transcript timed to it, a written assessment per question with its reasoning, and integrity signals marked on the timeline. You read the words, then you decide. > - The steps are: set the questions, invite people, let them join when it suits them, read one result per person. API and MCP connections let a team run it from the tools it already uses. Fact-checked on 4 Nov 2026 against Tracy's own pages and the original survey release. Sources are listed at the end. An AI screening interview is a first-round conversation that software runs, so that every applicant gets one instead of only the few a recruiter has time to call. This guide walks through what happens from invite to report, using [Tracy AI Interviewer](https://tracyinterview.ai/) as the worked example, and says what a good first round has to find out. ## Why teams move the first round to an AI interviewer In a Robert Half survey fielded in November 2025, 67% of US HR leaders said reviewing AI-generated applications had slowed their hiring. The release says the survey has responses from more than 2,000 US hiring managers and does not break out the group behind the 67%, so read it as the release's own wording. It reports how leaders see their hiring; it does not measure delay or cover interviews. When more applications arrive than a team can talk to, most applicants never get a conversation. A screening interview moves the conversation itself to the start, so everyone who applies can be asked the same base questions. If you are still deciding how to cut a long list, [how to screen candidates when there are too many applicants](/blog/how-to-screen-candidates) covers the options before the interview. ## What a first round has to find out Before you pick a tool, write down what the first round is for. A good one finds out three things: - Whether the person has done the work the role needs, in enough detail that you can tell. - Whether they can explain their own decisions as well as name them. - Whether they want the job as described. A form collects a claim. A conversation can ask a second question when the first answer is vague, and with Tracy it can start from the CV: with CV check on, it asks 3 to 5 questions tailored to the CV and the role's must-haves, on top of the role's own questions. The difference between [one-way and two-way video interviews](/blog/one-way-vs-two-way-video-interviews) also matters here. ## How a first round runs with Tracy ### 1. Set the goal and the questions Start from the job description or your own brief and Tracy drafts the questions. You can edit them, replace them with your own, and set the style from quick to deep and from behavioural to technical. When you set up from ChatGPT, nothing is created until you approve the draft, so you read it, ask for changes and say yes when it is right. Write the questions around the must-haves of the role. Keep the list short enough that a candidate can answer each one properly. ### 2. Choose the language Each interview is a two-way video interview in one language, which you choose at setup. A microphone check runs before the first question. ### 3. Invite people There are three ways in: send a personal invite link, publish an apply page, or let your own system ask for an interview and hand the person the link. Candidates have nothing to install and start when it suits them. People are told about the recording before it starts. Tracy is not for talking to someone who has not agreed to be recorded. On top of that notice, tell candidates in the invitation that an AI interviewer runs the conversation, so they hear it before they click. [How AI interviews work](/blog/how-do-ai-interviews-work) covers what candidates see. ### 4. The conversation Tracy runs a two-way video conversation. Its home page gives a typical response time under one second, which is Tracy's own figure. Tracy listens for what was said and asks for the missing detail, and each follow-up depends on the answer before it. A vague answer gets a sharper question, and people can ask questions about the process. It waits for the end of a thought before it speaks, with longer pauses allowed on hard questions, so it does not cut candidates off. Each interview is its own session that starts on click, so applicants do not queue for a slot and a recruiter does not need to be present; a candidate who drops can rejoin the same session within 10 minutes. Tracy's product page shows a follow-up in an example from its site, which uses invented people: > Tracy: Walk me through the index you chose for the ledger table. > Nina: A B-tree on account and date. Reads were the bottleneck. > Tracy (follow-up): How much faster were reads, and what did it cost you on writes? > Nina: About four times faster on reads. Writes slowed by a fifth. The follow-up asks for a number and a cost, which a vague first answer would not have given. ### 5. Read the result Each interview ends in a result you can check against the recording. It holds a written assessment for each question with its reasoning, the video with a transcript timed to it, and integrity signals such as tab switching, leaving fullscreen or an external screen, marked on the recording timeline for you to review. A signal marks a place to look. How to read many of them quickly is the subject of [AI interview reports](/blog/ai-interview-reports). Tracy does not reject anyone: a person decides. Tracy has API and MCP connections, so a team can set up interviews, invite people and read results from tools it already uses, and enterprise teams can ask for a customised setup. ## What it costs The first 3 results are free. After that, a credit is $1 and credits never expire. Billing is 1 credit per started 20 minutes of interview, so a 34-minute interview is 2 credits. You pay for interviews that produced a result, and a session with no answers is never charged. There are no seats; you pay per result. No card is needed to start. Plan for volume before you send a public link: check your credit balance against the number of people you expect. ## Where an AI screening interview is the wrong tool - **A person has not agreed to be recorded.** Tracy records and transcribes every interview, so it does not fit. - **The decision needs a human relationship.** A first round gathers evidence; it does not replace the conversation where a hiring manager and a candidate decide whether they want to work together. - **You have a handful of applicants.** If you can speak with everyone yourself, you may not need a tool. ## Mistakes to avoid 1. **Treating the result as a verdict.** The assessment tells you what people said. Someone on your team still decides. 2. **Asking too many questions.** Each answer gets thinner, and the follow-ups have less room. 3. **Skipping the preview.** Take the interview yourself first, so you know what your candidates will go through. 4. **Running out of credits mid-campaign.** Check the balance before a large send, so applicants you wanted are not left waiting. For the next step, [AI behavioral interviews](/blog/ai-behavioral-interviews) shows how follow-ups test a rehearsed story. ## Common questions ### What is an AI screening interview? A first-round conversation run by software instead of a recruiter. The hiring team sets the questions, each applicant joins from a link, and the team reads a written result afterwards. ### Does an AI screening interview decide who gets hired? It should not. Tracy AI Interviewer asks the questions and writes up what people said. It does not reject anyone: a person reads the result and decides. ### Do candidates need to install anything? With Tracy, no. There is no install for the person being interviewed: they start from a link in the browser, on video, and are told about the recording before it starts. ### How is a screening interview different from a one-way video interview? A one-way recording cannot ask a second question. In a two-way interview the interviewer can follow up on a vague answer while the candidate is still there. ### What does a screening interview cost with Tracy? The first 3 results are free. After that it is $1 per credit, with 1 credit per started 20 minutes of interview. A session where nobody answers is never charged. ### What should I check before using one? Whether candidates are told what they are joining, whether the interview can ask follow-ups, whether the result links back to the recording and transcript, and who makes the decision. --- **[Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=ai-screening-interviews&utm_content=try-free)** and interview your first applicants at no cost, or [book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=ai-screening-interviews&utm_content=demo) to see a full report first. ## Sources All sources retrieved 4 Nov 2026. - Tracy AI Interviewer, [Home page](https://tracyinterview.ai/): the example follow-up from the site (invented people), typical response time, "You read the words, then you decide", integrity signals, "Teams can also connect through the API", first 3 results free. - Tracy AI Interviewer, [Pricing](https://tracyinterview.ai/pricing): $1 per credit, billing per started 20 minutes, no-answer sessions never charged, no seats, no card, Enterprise with a customised setup and API and MCP connections. - Tracy AI Interviewer, [Product](https://tracyinterview.ai/product): questions drafted from a brief, CV check, three ways in, no install, people told about the recording before it starts, two-way video interview, transcript timed to the recording, follow-ups, assessment per question, integrity signals, "Tracy does not reject anyone. A person decides.", API and MCP. - Tracy AI Interviewer, [Hiring at volume](https://tracyinterview.ai/use-cases/volume-hiring): everyone who applies starts on click at any hour with no recruiter and no calendar, rejoin within 10 minutes. - Tracy AI Interviewer, [ChatGPT help](https://tracyinterview.ai/help/chatgpt): the interview draft is reviewed and approved before it is created. - Robert Half, [Survey: 67% of HR leaders report AI-generated applications are slowing hiring](https://press.roberthalf.com/2026-03-10-Robert-Half-survey-67-of-HR-leaders-report-AI-generated-applications-are-slowing-hiring), 10 March 2026: the 67% figure, a November 2025 survey; the release reports it for US HR leaders and says the survey has responses from more than 2,000 US hiring managers. # How Do AI Interviews Work? (For Employers and Candidates) > How an AI interview works from both sides: what the employer sets up, what the candidate sees, what is recorded, who decides, and what disclosure means. - URL: https://tracyinterview.ai/blog/how-do-ai-interviews-work - Published: 2026-11-04 - Updated: 2026-11-11 - Author: Tracy team > **In short** > - With Tracy AI Interviewer the employer sets an interview up once and every applicant is interviewed when they click the link, at any hour, with no scheduling. It can read the CV and tailor questions to it, and each follow-up depends on the answer before it. > - The employer then reviews one result per person. Every point links to the moment in the recording, and integrity signals come with the evidence behind them. Tracy does not reject anyone. > - People are told about the recording before it starts. Whether an AI is running it should be said earlier, in the invitation. Fact-checked on 4 Nov 2026 against Tracy's own pages, the survey report and the bill texts. Sources are listed at the end. An AI interview is a job interview in which software asks the questions. The rest of this page follows one interview from both sides: what the employer does before and after, and what the candidate sees in between. The employer side uses [Tracy AI Interviewer](https://tracyinterview.ai/) as the example, because its steps are published. If you want a definition first, start with [what an AI interviewer is](/blog/what-is-an-ai-interviewer). ## For employers: setup, questions, result ### Setup You start from the job description or your own brief, and Tracy drafts questions that you can edit or replace with your own. When you set up from ChatGPT, nothing is created until you approve the draft. With CV check on, Tracy also reads each candidate's CV and asks 3 to 5 questions tailored to it and the role's must-haves. Each interview is a two-way video interview in one language, which you choose at setup. ### The interview You send a personal invite, publish an apply page, or let your own system hand the person a link. Each candidate starts from the link in a browser at any hour, so nobody waits for a slot and a recruiter does not need to be present. Tracy listens to what was said and each follow-up depends on the answer before it, so a vague answer gets a sharper question. It waits for the end of a thought before it speaks, with longer pauses allowed on hard technical questions. Its typical response time is under one second, which is Tracy's own figure. For a closer look at the first-round shape, see [how an AI screening interview works](/blog/ai-screening-interviews). ### The result Each interview ends in a written assessment for every question, with its reasoning, next to a transcript timed to the recording. A tab switch or an external screen is flagged against the recording, and a person reviews it. Tracy also has API and MCP connections for teams that want to set up interviews and read results from the tools they already use. [AI interview reports](/blog/ai-interview-reports) explains what a reviewer should look for in one. ### The decision Tracy does not reject anyone: a person decides. To keep that true in your process, decide in advance who reads the results and what they are allowed to conclude from them. ## For candidates: what to expect For a candidate who has been sent an AI interview, this is what a Tracy interview looks like. - **You get a link.** You open it in a browser. There is nothing to install. - **You choose the moment.** You start when you click the link, at any hour, so there is no call to schedule. - **You are told about recording.** People are told about the recording before it starts. - **It is a video interview.** You talk to the interviewer, and a microphone check runs before the first question. - **Expect follow-ups.** A vague answer gets a sharper question, such as how much faster something was and what it cost you. Concrete detail helps; a rehearsed summary invites another question. You can also ask the interviewer questions about the process. Ask the employer who reads your answers and who decides. With Tracy the interview does not reject anyone: a person decides. ## Disclosure: what candidates are told In Greenhouse's 2026 Candidate AI Interview Report, 70% of US candidates who had been evaluated by AI said it had not been clearly disclosed before their most recent AI interview. That is a vendor survey of what candidates recall, from Greenhouse, which sells hiring software including Voice AI interviews. It is not a measure of how often employers fail to disclose. The report prints the question ("Before your most recent AI interview, were you told upfront that AI would evaluate you?") but not its field dates. It surveyed 2,950 active job seekers in five countries, 1,200 of them in the US, and does not print the size of the group of US candidates who had an AI evaluation, which is the base of the 70%. Employers should say it in the invitation. With Tracy, people are told about the recording before it starts. The invitation is where to say that an AI interviewer runs the conversation, what it will ask about and who reads the result. Some laws now ask for notice. The Illinois HB 3773 (Public Act 103-0804) bill status records an effective date of 1 January 2026, and its synopsis makes it a civil rights violation for an employer to use artificial intelligence that has the effect of subjecting employees to discrimination on the basis of protected classes, or to fail to provide notice to an employee that the employer is using artificial intelligence. The synopsis says "employee"; we read an archived copy of the bill page, not the act, so check whether applicants are covered. Colorado's SB 26-189, as enacted, applies to consequential decisions made on or after 1 January 2027 (a few provisions take effect on passage); before using covered automated decision-making technology to materially influence an employment decision, an employer must give the person a clear and conspicuous notice (a prominent public notice at points of interaction is one way to do it), and must later explain the technology's role if the outcome is adverse. These are explanations of the statutes, not legal advice; check how they apply to you. ## Where AI interviews fall short - **Someone who has not agreed to be recorded.** Tracy is not for that. - **The final call.** An interview result is evidence for a decision, and a person makes the decision. - **A role that needs a person in the room.** If the first conversation is about rapport, run it yourself. When you compare an AI interview with a person doing the same job, [AI interview vs human interview](/blog/ai-interview-vs-human-interview) sets out what each does well. ## What it costs For employers: the first 3 results are free, then $1 per credit, charged per started 20 minutes when the result is ready. A session where nobody answers is never charged. If you are weighing tools on price, [AI interview cost](/blog/ai-interview-cost) sets the models side by side. ## Common questions ### How does an AI interview work? The employer sets the questions and sends a link. The candidate joins from a browser and has a spoken conversation with the AI interviewer, which can ask follow-up questions. Afterwards the employer reads a written result. ### Is an AI interview recorded? With Tracy AI Interviewer, yes. The interview is recorded on both sides with a transcript, and people are told about the recording before it starts. ### Does the AI decide whether I get the job? With Tracy, no. It does not reject anyone: it asks the questions and writes up what people said, and a person decides. ### Do I need to download anything? With Tracy, no. There is no install for the person being interviewed: you join from a link in the browser, and a microphone check runs before the first question. ### Can the employer tell me an AI is interviewing me? They should, and some laws require notice. Illinois and Colorado both have rules on telling people that AI is used. This is an explanation, not legal advice. ### What does the employer get after the interview? With Tracy, a written assessment per question with its reasoning, a transcript timed to the recording, and integrity signals with their evidence, for a person to read and decide. --- **[Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=how-do-ai-interviews-work&utm_content=try-free)** and take a practice interview yourself before you invite anyone, or [book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=how-do-ai-interviews-work&utm_content=demo) to see a finished result. ## Sources All sources retrieved 4 Nov 2026. - Tracy AI Interviewer, [Home page](https://tracyinterview.ai/): two-way video, follow-ups, typical response time, "You read the words, then you decide", integrity signals, "Teams can also connect through the API". - Tracy AI Interviewer, [Pricing](https://tracyinterview.ai/pricing): free results, $1 per credit, billing per started 20 minutes, no-answer sessions never charged, Enterprise with API and MCP connections. - Tracy AI Interviewer, [Product](https://tracyinterview.ai/product): questions drafted from a brief, CV check, three ways in, one interview language chosen at setup, no install, people told about the recording before it starts, recorded on both sides with a transcript timed to the recording, follow-ups, microphone check, "Tracy does not reject anyone. A person decides.". - Tracy AI Interviewer, [Hiring at volume](https://tracyinterview.ai/use-cases/volume-hiring): each applicant starts when they click the link, at any hour, with no recruiter and no calendar. - Tracy AI Interviewer, [ChatGPT help](https://tracyinterview.ai/help/chatgpt): the interview draft is approved before it is created. - Greenhouse, [2026 Candidate AI Interview Report](https://www.greenhouse.com/blog/2026-candidate-ai-interview-report): the 70% figure, US candidates who experienced AI evaluation. - Greenhouse, [2026 Candidate AI Interview Report, full PDF](https://cdn.prod.website-files.com/66bf84e70c6fc57dbbe0187e/69f0e4b59dbbc57d8439ed9c_Greenhouse%202026%20Candidate%20AI%20Interview%20Report.pdf): the printed question wording and the sample (2,950 job seekers, 1,200 in the US). - Greenhouse, [Voice AI interviewing](https://www.greenhouse.com/voice-ai-interviewing): Greenhouse sells Voice AI, a structured voice interview. - Colorado General Assembly, [SB 26-189 as enacted](https://leg.colorado.gov/bill_files/116489/download): clear and conspicuous notice before covered automated decision-making technology materially influences a consequential decision (section 6-1-1704), the 1 January 2027 effective date and applicability, provisions that take effect on passage. - Illinois General Assembly, [HB 3773 bill status](https://web.archive.org/web/2026/https://ilga.gov/ftp/legislation/103/BillStatus/HTML/10300HB3773.html): Public Act 103-0804, effective 1 January 2026, synopsis (archived copy of ilga.gov). # AI in Hiring Statistics: 15 Numbers to Know in 2026 > Current AI in hiring statistics on candidate use, AI interviews, disclosure, application fraud and human review, with dates and survey methods. - URL: https://tracyinterview.ai/blog/ai-in-hiring-statistics - Published: 2026-11-03 - Author: Tracy team ![Among US job seekers surveyed by Greenhouse in October 2026, 77% used AI in their search, 60% changed or considered changing their writing, and 81% were rejected without a conversation.](/blog/images/hiring-statistics-job-seekers.png) *Source: [Greenhouse, October 2026](https://www.greenhouse.com/newsroom/6-in-10-job-seekers-are-rethinking-how-they-write-to-avoid-sounding-like-ai-even-as-most-use-it-in-their-search). US respondents, n=1,200. These are separate experiences, not parts of a total.* > **The short version** > - AI is now present on both sides of hiring. Separate Greenhouse surveys found 77% of US job seekers had used it and 63% had faced an AI interview. > - Disclosure and human review are the fault lines. In one Greenhouse survey, 70% of people who faced an AI interview said they were not told beforehand. > - Application volume is not the same as applicant quality. Robert Half found 67% of US hiring managers said AI-generated applications were slowing hiring. These are the most useful AI in hiring statistics we could verify as of 2 November 2026. Each figure keeps its source, population and fieldwork context. A survey of active job seekers cannot tell you what percentage of employers use AI, and a fraud rate from proctored coding assessments cannot be relabelled as an interview cheating rate. ## AI use by job seekers ### 1. 77% of US job seekers had used AI in their search Greenhouse surveyed more than 2,800 job seekers across the United States, United Kingdom, Ireland, Germany and Australia for a report released in October 2026. Among US respondents, **77% said they had used AI in their job search**. The same release found **60% had changed or considered changing how they write to avoid sounding like AI**. The split was 39% who had already rewritten material and 21% who had considered it. Recruiters are therefore reading applications in which genuine candidates may deliberately sand off polished language to look more human. ### 2. 81% of US job seekers had been rejected without a conversation In that October Greenhouse survey, **81% of US respondents said they had been rejected without speaking to a person**. The survey does not attribute those rejections to AI. It records the experience against which candidates judge automated screening. ### 3. 71% took extra steps to get noticed Greenhouse reported that **71% of US job seekers took an extra step to get noticed**, such as seeking a referral or paying for a resume service. Separately, **62% tried tactics to get past automated screening**. These figures describe candidate actions; they do not measure how many employers added hiring stages. ## AI interview adoption and candidate experience ### 4. 63% of US job seekers had faced an AI interview Greenhouse's release on 1 May 2026 describes a survey of **2,950 active job seekers** across the US, UK, Ireland, Germany and Australia. Its methodology says the reported findings refer to **US respondents, n=1,200**, unless otherwise noted. **Sixty-three percent of that US group said they had experienced an AI interview.** "AI interview" covered more than one format. Respondents could be describing recorded video scored by AI, a text exchange or a spoken AI interviewer. Buyers should name the actual format when comparing tools. ### 5. 38% of US candidates had left a process because it included an AI interview Among US respondents, **38% said they had withdrawn from a hiring process because it included an AI interview**, while another **12% said they would consider doing so**. That is a warning about implementation, not proof that every candidate rejects the format. Greenhouse also reported that **nine in ten job seekers were open to an AI interview** in its later October survey. The figures can coexist: people may accept a useful, disclosed conversation and reject a confusing or impersonal format. ### 6. 70% said the employer did not disclose the AI interview beforehand Among respondents who had faced an AI interview, **70% said they were not told in advance**. **Twenty-one percent said they learned only after the interview had begun.** Disclosure belongs in the invitation, before a candidate spends time preparing or starts recording. ### 7. 33% named AI-scored prerecorded video with no human present as a reason to withdraw Greenhouse asked US candidates what would make them leave an AI interview process. **Thirty-three percent selected prerecorded video scored by AI with no human present**. Other triggers included **failure to disclose how AI would be used at 27%** and **monitoring during the interview at 26%**. Those percentages describe stated triggers, not market share. They should guide the candidate experience: explain the format, the data collected and who reviews the result. ### 8. 51% said they never heard back after an AI interview Greenhouse reported that **51% of US candidates who completed an AI interview said they never heard back**. **Twenty-eight percent reached the next round and 13% received a rejection.** The missing share reflects other outcomes in the survey. Automation does not excuse silence. If a team can send an interview link at scale, it can send a clear outcome at scale too. ![Candidates named AI-scored prerecorded video without a human present at 33%, undisclosed AI use at 27%, and AI monitoring at 26% as triggers for leaving an AI interview process.](/blog/images/hiring-statistics-withdrawal.png) *Source: [Greenhouse, May 2026](https://www.greenhouse.com/newsroom/63-of-job-seekers-have-faced-an-ai-interview-most-havent-had-a-good-one-yet). US respondents, n=1,200. These figures describe selected withdrawal triggers, not a breakdown of all outcomes. The percentage axis starts at zero and runs to 100%.* ## Application quality and verification ### 9. 67% of US hiring managers said AI-generated applications were slowing hiring Robert Half surveyed more than **2,000 US hiring managers in November 2025** and released the results in March 2026. **Sixty-seven percent said AI-generated applications were slowing the hiring process.** **Twenty percent reported delays longer than two weeks.** This was a manager perception survey, not a timed audit of hiring systems. It is still useful evidence of the operational pressure teams feel. ### 10. 65% said verifying applicant skills had become harder In the same Robert Half survey, **65% said it was harder to verify whether candidates had the skills they claimed**. **Eighty-four percent said AI-generated applications increased workload.** The reported responses included reviewing applications more closely at **42%**, adding interviews at **38%**, and updating job descriptions at **32%**. Adding interviews only helps if those conversations produce evidence a reviewer can check. ![Robert Half reports that 84% of US HR leaders saw heavier workloads, 67% said AI applications slowed hiring, and 65% of hiring managers found skills harder to verify.](/blog/images/hiring-statistics-manager-workload.png) *Source: [Robert Half, March 2026 release](https://press.roberthalf.com/2026-03-10-Robert-Half-survey-67-of-HR-leaders-report-AI-generated-applications-are-slowing-hiring). Survey of more than 2,000 US hiring managers, conducted in November 2025. These are reported perceptions, not measured processing times.* ### 11. Suspicious attempts in CodeSignal assessments rose from 16% to 35% CodeSignal reported that attempts flagged by its detection systems rose from **16% in 2024 to 35% in 2025**. This is vendor data from **proctored skills assessments**. It does not establish that 35% of candidates cheat, and it is not an AI interview statistic. Use the figure narrowly: assessment providers are seeing more behaviour their systems classify as suspicious. Any integrity signal should prompt review, not an automatic verdict. ## Human oversight and public attitudes ### 12. 71% opposed AI making the final hiring decision Pew Research Center surveyed US adults in **December 2022**, before the current wave of generative AI hiring tools. **Seventy-one percent opposed employers using AI to make a final hiring decision**, 7% favoured it and 22% were unsure. The age of the fieldwork is part of the citation. It remains a useful baseline for a durable principle: tools can collect and organise evidence, while a person remains accountable for the decision. ### 13. 41% opposed AI reviewing job applications In the same Pew survey, **41% opposed using AI to review applications**, **28% favoured it**, and **30% were unsure**. Public opinion was less negative about assisting the review than making the final decision, but it was far from settled. ### 14. 38% wanted confirmation that a person reviews AI output Greenhouse's 2026 candidate research found **38% of US candidates wanted confirmation that a human reviews AI output before decisions are made**. This is an implementation requirement a team can act on immediately: say who reviews the result and where the final decision sits. ### 15. 27% would withdraw if AI use was not disclosed In the same Greenhouse research, **27% named nondisclosure as a reason to withdraw**. Disclosure is therefore not a legal footnote or a line buried in a privacy policy. It affects whether candidates continue. ## How to use AI hiring statistics without misleading readers 1. **Name the population.** "63% of US job seekers surveyed" is accurate. "63% of interviews use AI" is not. 2. **Keep the fieldwork date.** Pew's attitudes were measured in 2022, even though its report was published in 2023. 3. **Separate formats.** A prerecorded video, a chatbot and a two-way spoken interview create different experiences. 4. **Separate interviews from assessments.** CodeSignal's 35% relates to suspicious attempts in its assessments. 5. **Do not combine ratings or percentages.** Different sources asked different questions of different groups. 6. **Link to the original.** A reader should be able to inspect the method and wording. For teams choosing a format, [one-way vs two-way video interviews](/blog/one-way-vs-two-way-video-interviews?utm_source=blog&utm_medium=seo&utm_campaign=ai-in-hiring-statistics&utm_content=internal-link) explains the practical difference. [What is an AI interviewer?](/blog/what-is-an-ai-interviewer?utm_source=blog&utm_medium=seo&utm_campaign=ai-in-hiring-statistics&utm_content=internal-link) covers the category without treating every automated step as the same product. ## Common questions ### How common are AI interviews in 2026? In Greenhouse's 2026 survey, 63% of US job seekers said they had faced an AI interview. The figure describes the surveyed job seekers, not all employers or all interviews. ### Do candidates want a human to review AI interview results? Yes. Greenhouse reported that 38% of US candidates wanted confirmation that a person reviews AI output before a decision. A separate Pew survey fielded in 2022 found 71% opposed AI making a final hiring decision. ### How many job seekers use AI in their search? Greenhouse's October 2026 survey reported that 77% of US job seekers had used AI in their search. ### Is AI-generated application fraud increasing? Employers report more difficulty verifying skills, and CodeSignal reported suspicious attempts in its proctored assessments rose from 16% in 2024 to 35% in 2025. That CodeSignal figure is about assessments, not job interviews. ### Can I cite these statistics? Yes. Cite the original source, fieldwork date, sample and population shown beside each number rather than citing this summary alone. --- **[Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=ai-in-hiring-statistics&utm_content=try-free)** or **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=ai-in-hiring-statistics&utm_content=demo)** ## Sources All sources retrieved 2 Nov 2026. - Greenhouse, [2026 Candidate AI Interview Report release](https://www.greenhouse.com/newsroom/63-of-job-seekers-have-faced-an-ai-interview-most-havent-had-a-good-one-yet): 2,950 active job seekers, interview exposure, withdrawal, disclosure and outcome figures. - Greenhouse, [October 2026 job-seeker survey release](https://www.greenhouse.com/newsroom/6-in-10-job-seekers-are-rethinking-how-they-write-to-avoid-sounding-like-ai-even-as-most-use-it-in-their-search): more than 2,800 job seekers, US AI use, writing changes and hiring experience. - Robert Half, [AI-generated applications survey](https://press.roberthalf.com/2026-03-10-Robert-Half-survey-67-of-HR-leaders-report-AI-generated-applications-are-slowing-hiring): more than 2,000 US hiring managers surveyed in November 2025. - CodeSignal, [2025 assessment fraud release](https://codesignal.com/newsroom/press-releases/codesignal-detection-systems-identify-and-stop-record-high-cheating-attempts-as-assessment-fraud-more-than-doubled-in-2025): vendor detection data for proctored assessments. - Pew Research Center, [AI in hiring and evaluating workers](https://www.pewresearch.org/social-trends/2023/04/20/ai-in-hiring-and-evaluating-workers-what-americans-think/): US adult attitudes from a December 2022 survey. # AI Interview Reports: What Reviewers Should See > A good AI interview report traces every conclusion to the candidate's answer. Use this checklist for summaries, scores, quotes, recordings and flags. - URL: https://tracyinterview.ai/blog/ai-interview-reports - Published: 2026-11-03 - Updated: 2026-11-13 - Author: Tracy team ![An illustrated Tracy result report for a fictional engineering candidate: a summary, the recording with a timestamped transcript, each question with what the candidate said and quoted evidence, and fields marked Stated.](/blog/images/lib-report-overview-technical-en.png) *A recreation of the result report layout with fictional data, not a screenshot of a real candidate (illustration with sample data).* An AI interview report is useful when a reviewer can move in both directions: from the candidate's answer to the summary, and from any summary or score back to the exact answer that supports it. If the report only says "communication: 82," the reviewer has a number but no evidence. If it contains a two-hour recording with no structure, the evidence exists but is expensive to use. A good report puts the decision fields first and keeps the candidate's words one click away. ## The minimum AI interview report Every report should answer six questions before a reviewer decides who moves forward. 1. **What was this interview meant to learn?** Show the role, brief, competencies and any must-have requirements. 2. **What was asked?** Keep the opening questions and the follow-ups in order. 3. **What did the candidate say or do?** Preserve the transcript, recording and any work sample. 4. **What did the system conclude?** Show structured fields, summaries, scores and flags. 5. **Why did it reach that conclusion?** Attach a quote, timestamp, code sample or explanation. 6. **Who decided what happened next?** Record the reviewer and outcome separately from the AI output. The sixth question is not administrative decoration. Greenhouse's 2026 candidate research found 38% of US candidates wanted confirmation that a person reviews AI output before decisions. Pew's US survey, fielded in December 2022, found 71% opposed AI making a final hiring decision. ![An illustrative interview question about a migration, the candidate's answer about designing an idempotency layer, and the reviewer's check against the transcript.](/blog/images/ai-interview-reports-process.png) *A question, answer and reviewer check (illustration with sample data).* ## 1. Interview brief and version Put the job of the interview at the top. Include: - role or programme; - fields the team asked the interviewer to collect; - competencies and scoring anchors; - required questions; - interview language and format; - the brief version and date. Versioning matters. If the team changes the system-design prompt halfway through a hiring round, the reports should show which candidates received which version. Reviewers can then avoid comparing unlike evidence as if it were identical. ## 2. Question and answer trail A two-way interviewer may ask each candidate a different follow-up. The report should show the full sequence alongside the prepared questions. For example: > **Question:** Tell me about the payment migration you led. > > **Answer:** I moved our checkout from a legacy gateway to a new provider over three months. > > **Follow-up:** Which part did you own personally, and what failed during the rollout? > > **Answer:** I designed the idempotency layer and rollout flags. Our first one-percent release exposed duplicate webhook processing, so I stopped the rollout and added a unique event constraint before resuming. The first answer states a project. The follow-up supplies ownership, a failure and the response. A summary that says "owned idempotency design and corrected duplicate webhook processing during staged rollout" can now cite the second answer. ## 3. Structured fields tied to the brief Turn the conversation into the fields a hiring manager actually needs. For a backend engineer, that could be: | Field | Evidence in the report | | --- | --- | | Relevant scale | Traffic, data volume and team context quoted from the answer | | Personal contribution | What the candidate designed, built or decided | | Technical reasoning | Options considered and why one won | | Failure handling | What went wrong, how it was detected and what changed | | Availability | Notice period or start date, when appropriate to collect | Avoid generic generated paragraphs that repeat the CV. A field should either change the decision or disappear from the report. Tracy AI Interviewer's current product page describes structured results backed by the participant's words and direct quotes. This description comes from Tracy's own page. Tracy's result report follows that order. It opens with a summary of the interview, then the recording beside a transcript timed to it. Each question shows what the person said and the quotes behind it, each field is marked Stated with its evidence, and integrity signals are counted for review. It gives no overall score. The illustration above is a recreation with fictional data. ## 4. Quotes, timestamps and source labels Each material conclusion should carry a short route to its source. Useful patterns include: - a direct quote with its timestamp; - a link to the relevant transcript turn; - a marker on the recording; - the submitted code or file; - a label that separates candidate-provided facts from system interpretation. Quotes need context. A five-word fragment can reverse the meaning of a longer answer. Let the reviewer expand the surrounding exchange without hunting through the recording. The report should also distinguish three layers: 1. **Observed:** what the candidate said or did. 2. **Derived:** a calculation or structured extraction from the answer. 3. **Interpreted:** the system's judgement about strength, risk or fit. That separation prevents a generated interpretation from looking like a candidate quote. ## 5. Scores with reasons A score is useful only if the reviewer knows the scale and can inspect the reason. Instead of: > System design: 4/5 Use: > System design: 4/5. The candidate separated ingestion from delivery, used idempotency keys at the payment boundary and described a rollback. They did not address cross-region state until prompted. Then link the explanation to the relevant answers. HackerRank describes a clear "why" for its Chakra scores, and other vendors make similar claims. Buyers should test the path in a real demo: click a low score, inspect the evidence and see whether a reviewer can disagree or override it. Use anchored scales. A practical four-point version is: | Score | Meaning | | --- | --- | | 1 | No relevant evidence, or the approach would fail the stated requirement. | | 2 | Some relevant evidence, with an important gap or untested assumption. | | 3 | Sound evidence that meets the requirement and explains the trade-off. | | 4 | Strong evidence that also anticipates failure modes and adapts when challenged. | Do not hide separate competencies inside one overall number. Reviewers need to know whether the gap is technical depth, communication, domain knowledge or a must-have constraint. ## 6. Recording, transcript and work sample The summary helps the reviewer scan. The source material lets them verify. Keep the recording and transcript aligned by timestamp. Let reviewers search the transcript and jump to the matching video or audio. For technical interviews, keep the submitted code, output and task version beside the conversation. Retention should be deliberate. The candidate invitation should explain what is recorded and how it is used. The organisation should delete material according to its policy rather than keeping every interview indefinitely because storage is cheap. ## 7. Integrity signals with context An integrity section should report observations, not pronounce guilt. A tab change, long pause, second screen or unusual gaze can have an innocent explanation. A candidate may be reading the task, using an approved accommodation or dealing with a connection problem. For each flag, show: - the signal name; - time and duration; - the relevant recording or transcript moment; - whether the behaviour was allowed under the interview rules; - space for a reviewer note and outcome. Never turn one signal into an automatic rejection. Use it to direct a person's attention to a moment worth checking. The same principle applies to a suspected copied answer: ask a follow-up about the candidate's reasoning and review what follows. ## 8. Human review and audit trail Keep AI output and human action separate. The report should show: - who opened or reviewed it; - any score the reviewer changed and why; - the decision, such as progress, hold or decline; - the date and responsible person; - the candidate communication sent after the decision. An audit trail makes calibration possible. If reviewers repeatedly override one competency, the team can inspect the question or rubric instead of assuming the people are inconsistent. Greenhouse reported that 51% of surveyed US candidates who completed an AI interview never heard back. The report workflow should end in a candidate response as well as an internal status. ## A review workflow that takes five minutes For a first-round report, a hiring manager should be able to review in this order: 1. Read the brief and must-have fields. 2. Scan the structured results. 3. Open the evidence behind any very strong, very weak or surprising point. 4. Check follow-ups for ownership and reasoning. 5. Review integrity signals in context. 6. Record the decision and reason. 7. Trigger the next-step or decline message. The five-minute target is not a promise for every role. A regulated, senior or technical decision may deserve longer. The design principle is progressive detail: the report is scannable first and verifiable when needed. ## AI interview report buyer test During a vendor demo, ask to use a brief you wrote. Give one complete answer and one vague answer. Then inspect the report. - Did the interviewer ask a relevant follow-up? - Did the summary preserve the meaning of the answer? - Can you click from a conclusion to the source? - Does each score have a scale and explanation? - Are quotes clearly separated from generated text? - Can a reviewer correct the report or override a score? - Does a flag show the underlying moment? - Is the human decision recorded separately? - Can the team export or delete the data? - Does the workflow send an outcome to the candidate? If the demo cannot answer those questions with a real result, the product is asking you to trust its output before it has made that output reviewable. For question design, use the [25 software engineer interview questions](/blog/software-engineer-interview-questions?utm_source=blog&utm_medium=seo&utm_campaign=ai-interview-reports&utm_content=internal-link) or the broader [AI technical interview guide](/blog/ai-technical-interviews?utm_source=blog&utm_medium=seo&utm_campaign=ai-interview-reports&utm_content=internal-link). For format choice, compare [one-way and two-way video interviews](/blog/one-way-vs-two-way-video-interviews?utm_source=blog&utm_medium=seo&utm_campaign=ai-interview-reports&utm_content=internal-link). ![A checklist of the main actions for ai interview reports.](/blog/images/ai-interview-reports-takeaways.png) ## Common questions ### What is an AI interview report? It is the reviewer-facing record created after an AI interview, usually containing the questions, answers, transcript or recording, summary fields and any scores or flags. ### What should an AI interview report include? Include the interview brief, questions, candidate answers, traceable summary fields, direct evidence, score rationales, recording or transcript, integrity signals and the human decision. ### Should recruiters trust an AI interview score? Treat it as a review aid, not a fact. A reviewer should be able to see the reason, inspect the underlying answer and change the outcome. ### How long should an interview report be? Make the first screen scannable, then let reviewers open the evidence. Length should follow the decision, not the interview duration. ### Can an integrity flag reject a candidate automatically? It should not. A flag can be caused by benign behaviour or technical conditions. Show the relevant moment and require a person to interpret it. --- **[Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=ai-interview-reports&utm_content=try-free)** or **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=ai-interview-reports&utm_content=demo)** ## Sources All sources retrieved 2 Nov 2026. - Tracy AI Interviewer, [product page](https://tracyinterview.ai/): structured results, direct quotes, video and audio interviews, adaptive follow-ups and human decision. - Greenhouse, [2026 Candidate AI Interview Report release](https://www.greenhouse.com/newsroom/63-of-job-seekers-have-faced-an-ai-interview-most-havent-had-a-good-one-yet): candidate expectations for human review and post-interview outcomes. - Pew Research Center, [AI in hiring and evaluating workers](https://www.pewresearch.org/social-trends/2023/04/20/ai-in-hiring-and-evaluating-workers-what-americans-think/): attitudes to final AI hiring decisions from a December 2022 survey. - HireVue, [AI Interviewer](https://www.hirevue.com/platform/ai-interviewer): current vendor description of two-way interviews, criteria and recruiter review. - Alex, [AI Interviewer](https://www.alex.com/product/ai-interviewer): current vendor description of interview formats and reviewer output. - HackerRank, [Chakra](https://www.hackerrank.com/chakra): current vendor description of score explanations and follow-up questions. # One-Way vs Two-Way Video Interviews: Which Shows You More? > Compare one-way vs two-way video interviews by candidate experience, follow-up depth, reviewer time, consistency and best use case. - URL: https://tracyinterview.ai/blog/one-way-vs-two-way-video-interviews - Published: 2026-11-03 - Author: Tracy team ![Cover graphic introducing one way vs two way video interviews.](/blog/images/one-way-vs-two-way-video-interviews-cover.png) Fact-checked on 2 Nov 2026 against current vendor documentation and Greenhouse's 2026 candidate survey ([see Sources](#sources)). > **In short** > - A **one-way video interview** records answers to preset questions. It removes scheduling and keeps prompts consistent, but someone still watches the answers and no question adapts to what the candidate said. > - A **two-way AI interview** asks follow-ups during the conversation. It can reach the example, constraint or trade-off behind a polished first answer. > - Use one-way video for short, repeatable prompts where the recording itself is the evidence. Use two-way interviews when reasoning and specifics determine who moves forward. **On this page** - [What is the difference between one-way and two-way interviews?](#what-is-the-difference-between-one-way-and-two-way-interviews) - [What does a one-way interview do well?](#what-does-a-one-way-interview-do-well) - [What does a two-way interview add?](#what-does-a-two-way-interview-add) - [What should candidates be told?](#what-should-candidates-be-told) - [Which format fits each hiring job?](#which-format-fits-each-hiring-job) - [Questions](#questions) - [Sources](#sources) ## What is the difference between one-way and two-way interviews? Spark Hire defines a one-way interview as a series of questions that a candidate answers on video at their convenience. The employer can set think time, answer length and allowed retakes. Once submitted, the hiring team watches and reviews the recordings. A two-way AI interview runs as a live spoken conversation. Tracy AI Interviewer starts from the employer's questions, listens to the answer and asks for relevant missing detail. The host chooses video or audio, and the participant joins in a browser. ![A two-way interview asks a question, hears the answer and can follow up.](/blog/images/one-way-vs-two-way-video-interviews-process.png) | Question | One-way video | Two-way AI interview | |---|---|---| | Does it require scheduling? | No | No, when the AI interview is available on demand | | Can the next question use the previous answer? | No | Yes | | What does the reviewer receive? | Separate recordings, sometimes transcripts and summaries | A conversation record, transcript and structured result | | Main review burden | Watch or scan each preset answer | Check the summarized evidence and open the recording where needed | | Best use | Consistent short prompts and employer-brand video | Examples, reasoning, technical depth and role-specific follow-up | ## What does a one-way interview do well? One-way video removes calendar coordination. Candidates can record before the deadline, and every person receives the same prompts. Employers can add introduction videos, practice steps and retakes. For roles where presentation or a short prepared response matters, the recording is useful evidence. It can also be the simpler purchase. A team that wants to replace a 15-minute phone screen with three fixed questions may not need a conversational AI. The hiring manager already knows what to ask and expects to watch the replies. Choose one-way video when: - the questions should remain identical for every candidate; - a prepared response is acceptable evidence; - reviewers want to watch each recording; - follow-up depth would not change the decision; - the organization has a clear alternative for candidates who cannot record video. ## What does a two-way interview add? Consider this first answer: "I improved the API and reduced latency." A fixed recording moves to the next prompt. A two-way interviewer can ask what changed, how latency was measured, which trade-off the candidate accepted and what failed in the first attempt. That branch matters in technical, expert and behavioral interviews. The candidate's second and third answers often reveal whether they did the work, understood the constraint and can explain the consequence. Our [software engineer interview questions](/blog/software-engineer-interview-questions?utm_source=blog&utm_medium=seo&utm_campaign=one-way-vs-two-way-video-interviews&utm_content=internal-link) include follow-ups designed for that purpose. Two-way does not remove human review. Tracy says it asks questions and writes up what the participant said while the decision stays with the customer. A reviewer should check the evidence, especially for a low score or integrity flag. See [what an AI interview report should contain](/blog/ai-interview-reports?utm_source=blog&utm_medium=seo&utm_campaign=one-way-vs-two-way-video-interviews&utm_content=internal-link). ## What should candidates be told? Greenhouse surveyed 2,950 active job seekers for its 2026 Candidate AI Interview Report. Among U.S. respondents, 70% said they were not clearly told upfront that AI would evaluate them. Pre-recorded video scored by AI with no person present was the most cited trigger among candidates who walked away, at 33%. Those figures do not prove that every one-way interview causes drop-off. They show that format, scoring and disclosure affect the experience. Before the interview, tell candidates: - whether the interviewer is a person, a recorder or an AI; - whether the session is recorded and transcribed; - what the system evaluates; - whether a person reviews the result; - how to request another format or accommodation; - how long the process takes and whether retakes are allowed. ## Which format fits each hiring job? | Hiring job | Better starting format | Why | |---|---|---| | Short customer-service introduction | One-way | The same short prompts and a recording may be enough | | Software engineering first round | Two-way | Reasoning appears in follow-ups on code and trade-offs | | Audition or presentation sample | One-way | Delivery is part of the evidence | | Domain-expert vetting | Two-way | The interviewer can test specificity and recency | | High-volume eligibility screen | One-way or structured chat | Fixed questions may decide the stage | | Behavioral interview | Two-way | Follow-ups can test the detail behind a rehearsed STAR answer | Run the chosen format on a real role before changing the full process. Track completion, candidate questions, reviewer minutes and next-stage quality. The [best AI interview software guide](/blog/best-ai-interview-software?utm_source=blog&utm_medium=seo&utm_campaign=one-way-vs-two-way-video-interviews&utm_content=internal-link) compares products across both formats. ![A checklist of the main actions for one way vs two way video interviews.](/blog/images/one-way-vs-two-way-video-interviews-takeaways.png) ## Common questions ### What is a one-way video interview? A candidate records answers to preset questions on their own time. No person or AI interviewer asks a question based on the answer. ### What is a two-way AI interview? A spoken interview where the AI responds to the candidate's answer and can ask a follow-up before moving on. ### When is a one-way interview enough? Use one-way video when the team needs the same short prompts from everyone and reviewers are willing to watch the recordings. ### When should I use a two-way interview? Use a two-way interview when the role requires examples, reasoning or trade-offs that usually appear after a follow-up question. ### Do candidates prefer one format? Preferences vary. Clear disclosure, practice, accessible alternatives and human review matter in either format. --- **[Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=one-way-vs-two-way-video-interviews&utm_content=try-free)** or **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=one-way-vs-two-way-video-interviews&utm_content=demo)** ## Sources All sources retrieved 2 Nov 2026. - Spark Hire, [one-way overview](https://recruit-support.sparkhire.com/hc/en-us/articles/40045121196699-One-Way-Video-Interview-overview) and [candidate process](https://candidates.sparkhire.com/hc/en-us/articles/48167885321755-How-a-Spark-Hire-one-way-video-interview-works): format, setup, prompts, retakes and review. - HireVue, [video interviewing](https://www.hirevue.com/platform/video-interviewing): one-way and live video scope. - Greenhouse, [2026 Candidate AI Interview Report release](https://www.greenhouse.com/newsroom/63-of-job-seekers-have-faced-an-ai-interview-most-havent-had-a-good-one-yet): sample, disclosure and candidate walk-away figures. - Tracy AI Interviewer, [product page](https://tracyinterview.ai/): two-way format, follow-ups, browser access and report evidence. # What Is an AI Interviewer? How It Works and What to Check > An AI interviewer asks candidates questions and records or evaluates their answers. Learn the formats, workflow, limits and buyer checklist. - URL: https://tracyinterview.ai/blog/what-is-an-ai-interviewer - Published: 2026-11-03 - Author: Tracy team ![Cover graphic introducing what is an ai interviewer.](/blog/images/what-is-an-ai-interviewer-cover.png) An AI interviewer is software that runs part of an interview. It asks a candidate questions, receives answers by voice, video or text, and produces a recording, transcript, summary or evaluation for a person to review. That definition excludes tools that only schedule an interview or rank a CV. It also covers more than the familiar one-way recording. Some AI interviewers hold a two-way spoken conversation and ask a follow-up based on the answer they just heard. ## The three common AI interview formats ### 1. One-way recorded interview The candidate sees or hears a preset question, gets preparation time, then records an answer. The next prompt is already chosen. Spark Hire's documentation describes this flow, including optional time limits, think time and retakes. One-way video is useful when a team wants the same short prompts from every candidate. It is asynchronous, so nobody has to match calendars. The trade-off is fixed depth: if an answer is vague, the system cannot ask what the candidate personally did. ### 2. Text or chat interview The candidate answers in writing. A chat interviewer may ask a fixed sequence or generate a follow-up. Text can work for eligibility, availability and short structured questions. It is less suitable when spoken communication or explaining a decision aloud is part of the role. ### 3. Two-way spoken AI interview The candidate speaks with an AI interviewer by video or audio. The system listens to an answer and can choose a follow-up before moving on. Tracy AI Interviewer, HireVue, Alex and HeyMilo each describe a conversational interview product, though their scope, channels and reports differ. This format matters when the hiring signal sits one question deeper. "I led the migration" can become "Which part did you own?" A system design answer can become "What fails first at ten times the traffic?" ![An AI interview starts with a brief, continues with an interview and ends in a human review.](/blog/images/what-is-an-ai-interviewer-process.png) ## How does an AI interviewer work? The exact system varies, but a responsible workflow has seven visible steps. 1. **The team defines the job of the interview.** A first round might check relevant experience, a technical task, availability and motivation. The interview should not collect information nobody will use. 2. **The team reviews the questions and rubric.** The product may draft questions, but a person should check wording, relevance and the standard for a strong answer. 3. **The candidate receives a clear invitation.** It should say an AI interviewer is involved, whether the session uses video or audio, what is recorded, and who reviews the result. 4. **The candidate joins and answers.** Some tools require an account or app. Tracy's current product page says participants can join from a browser link without signing up. 5. **The interviewer follows its question plan.** In a one-way system, that means the next preset prompt. In a conversational system, the next question may depend on the answer. 6. **The system builds a result.** Depending on the product, this may include a transcript, recording, structured fields, quotes, scores, rationales and integrity signals. 7. **A person reviews and decides.** The recruiter or hiring manager checks the evidence, resolves uncertain flags and chooses the next step. Step 7 keeps the decision with a person. Greenhouse's 2026 candidate research found 38% of US candidates wanted confirmation that a person reviews AI output before a decision. In Pew's older survey, fielded in December 2022, 71% of US adults opposed AI making the final hiring decision. ## What can an AI interviewer ask? The useful range is wider than basic screening, but the questions still need to match the role. - **Experience:** "Tell me about the last billing system you migrated." - **Behaviour:** "Tell me about a decision your team disagreed with." - **Technical reasoning:** "How would you make that service tolerate a regional outage?" - **Work sample:** "Walk me through this code and identify the bug." - **Eligibility:** "When could you start?" - **Role constraints:** "Are you available for the stated shift pattern?" For a spoken interview, the follow-up is often more revealing than the opening prompt. A candidate can prepare a broad success story. It is harder to prepare every detail about the constraint, mistake, measurement and personal contribution. Our [software engineer interview questions](/blog/software-engineer-interview-questions?utm_source=blog&utm_medium=seo&utm_campaign=what-is-an-ai-interviewer&utm_content=internal-link) turn that principle into 25 question-and-follow-up pairs. ## What should the report contain? A useful report lets the reviewer move from a conclusion back to the answer that supports it. Look for: - the questions asked and the candidate's full transcript; - the recording, when the candidate agreed to video or audio; - structured fields tied to the team's brief; - direct quotes or timestamps behind each summary point; - the reason for any score, not only the number; - clear separation between observed evidence and generated interpretation; - integrity signals presented for human review; - a record of who made the hiring decision. Tracy's current product page describes structured results backed by the participant's words, with direct quotes, video or audio and adaptive follow-ups. This description comes from Tracy's own page. See [what an AI interview report should contain](/blog/ai-interview-reports?utm_source=blog&utm_medium=seo&utm_campaign=what-is-an-ai-interviewer&utm_content=internal-link) for a field-by-field checklist. ## What are AI interviewers good at? ### Giving every applicant the same first opportunity An asynchronous link can stay open across time zones, and a core question set can remain consistent. This helps when applicant volume exceeds recruiter hours. Consistency does not guarantee fairness. A consistently irrelevant question is still irrelevant. The team must review the brief, accommodations and outcomes. ### Asking for specifics at scale A conversational interviewer can challenge a vague claim with a neutral follow-up. This is useful in technical, expert and behavioural screens where the evidence appears in the second answer. ### Turning a conversation into reviewable evidence A structured report can save the hiring manager from replaying every minute. The saving is real only when the summary is traceable. Reviewers should be able to check a quote, transcript or recording when something looks strong, weak or inconsistent. ## What are the limits? ### You still have to define the job Generic questions create generic evidence. The team needs to name the decisions, skills and constraints that matter. ### Generated scores can be wrong A polished explanation is not proof that a score is correct. Keep the underlying answer, show the rationale and make uncertain or consequential decisions reviewable. ### Candidate access differs Video, speech recognition, bandwidth, disability and language can affect the experience. Offer clear technical requirements, practice, accommodations and another route when needed. ### Monitoring can damage trust Greenhouse found 26% of US candidates named monitoring during an AI interview as a reason they would withdraw. Collect only what is needed, disclose it before the session and treat a signal as a prompt to inspect, not a verdict. ### Automation can make a bad process faster Greenhouse reported 51% of surveyed US candidates who completed an AI interview never heard back. An interview tool should connect to an outcome and a response, not become a holding pen. ## AI interviewer buyer checklist Ask a vendor to show the candidate journey and the reviewer journey in the same demo. | Area | Question to ask | | --- | --- | | Format | Is it one-way, chat or a live spoken exchange? | | Follow-ups | Can it ask about the candidate's actual answer? | | Evidence | Can a reviewer trace each conclusion to a quote, transcript or recording? | | Decision | Does a person approve every progression or rejection? | | Disclosure | What does the invitation tell the candidate before they begin? | | Access | What alternatives and accommodations are available? | | Data | What is recorded, where is it stored and when is it deleted? | | Integrity | Which signals are collected, and can a person review the underlying moment? | | Cost | Is pricing per result, interview, seat, role, month or hire? | | Workflow | How does the result reach the ATS and hiring manager? | Run the same sample brief through two products. Answer once with a clear example and once vaguely. Then compare the follow-up, the report and the evidence behind any score. A feature list cannot show those differences. For the financial comparison, see [AI interview cost](/blog/ai-interview-cost?utm_source=blog&utm_medium=seo&utm_campaign=what-is-an-ai-interviewer&utm_content=internal-link). For format trade-offs, see [one-way vs two-way video interviews](/blog/one-way-vs-two-way-video-interviews?utm_source=blog&utm_medium=seo&utm_campaign=what-is-an-ai-interviewer&utm_content=internal-link). ![A checklist of the main actions for what is an ai interviewer.](/blog/images/what-is-an-ai-interviewer-takeaways.png) ## Common questions ### What is an AI interviewer? An AI interviewer runs part of an interview by asking questions, receiving spoken, video or text answers, and producing a record or structured report for a person to review. ### Is an AI interview the same as a one-way video interview? No. A one-way interview presents preset prompts and records answers. A two-way AI interviewer can ask a follow-up based on what the candidate just said. ### Does an AI interviewer make the hiring decision? It should not. A responsible workflow uses the tool to collect and organise evidence while a recruiter or hiring manager reviews it and decides. ### What does a candidate experience in an AI interview? The candidate receives an invitation, learns the format and data use, joins by link, answers questions, and may receive follow-ups. The exact experience varies by vendor. ### What should I ask an AI interview vendor? Ask about disclosure, follow-up behaviour, report evidence, accessibility, retention, security, integrations, pricing, integrity signals and where human review occurs. --- **[Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=what-is-an-ai-interviewer&utm_content=try-free)** or **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=what-is-an-ai-interviewer&utm_content=demo)** ## Sources All sources retrieved 2 Nov 2026. - Tracy AI Interviewer, [product page](https://tracyinterview.ai/): two-way video and audio, browser access, adaptive follow-ups, technical tasks, structured results and human decision. - HireVue, [AI Interviewer](https://www.hirevue.com/platform/ai-interviewer): two-way voice interview, follow-up questions and recruiter review. - Alex, [AI Interviewer](https://www.alex.com/product/ai-interviewer): video, phone and chat formats, scheduling and follow-up positioning. - HeyMilo, [product page](https://www.heymilo.ai/): voice, video, SMS and form channels, scoring and multilingual scope. - Spark Hire, [one-way video overview](https://recruit-support.sparkhire.com/hc/en-us/articles/40045121196699-One-Way-Video-Interview-overview): fixed one-way format, think time, retakes and review. - Greenhouse, [2026 Candidate AI Interview Report release](https://www.greenhouse.com/newsroom/63-of-job-seekers-have-faced-an-ai-interview-most-havent-had-a-good-one-yet): candidate disclosure, monitoring, human review and outcome findings. - Pew Research Center, [AI in hiring and evaluating workers](https://www.pewresearch.org/social-trends/2023/04/20/ai-in-hiring-and-evaluating-workers-what-americans-think/): US public attitudes measured in December 2022. # Accelerator Interview: How to Meet Every Founder Who Applies > Run the first accelerator interview with every founder who applies. Hear what they're building and where they want to take it, then meet the strongest. - URL: https://tracyinterview.ai/blog/accelerator-interview - Published: 2026-11-02 - Updated: 2026-11-11 - Author: Tracy team > **In short** > - **Give every founder a first conversation before the application cut.** Tracy AI Interviewer asks your programme's questions, follows up on each answer and gives partners a concise record in the founder's own words. > - Strong founders with plain applications get a chance to be heard, while partners spend live calls on teams worth exploring. Partners still choose whom to meet and fund. **Contents** - [How do programmes run the first accelerator interview today?](#how-do-programmes-run-the-first-accelerator-interview-today) - [What can't an accelerator application show you?](#what-cant-an-accelerator-application-show-you) - [How does Tracy give every founder a first accelerator interview?](#how-does-tracy-give-every-founder-a-first-accelerator-interview) - [What accelerator interview questions should the first round ask?](#what-accelerator-interview-questions-should-the-first-round-ask) - [When are partner first calls the better choice?](#when-are-partner-first-calls-the-better-choice) - [How do you set it up before your next intake?](#how-do-you-set-it-up-before-your-next-intake) - [Questions](#questions) - [Sources](#sources) If you run a programme, the accelerator interview is where you hear what a founder is building and how far they want to take it. Partner hours limit how many founders get that first interview, so many programmes judge most applicants on the form alone. This guide shows how to give every applicant a first conversation and let partners choose who to meet from what each founder said. (Preparing for an accelerator interview as a founder? Skip to the [Questions](#questions).) ## How do programmes run the first accelerator interview today? A typical funnel looks like this: 1. An application form, sometimes with a short founder video. 2. A first call with a partner, a venture partner or an alum. 3. A longer interview with the partners who decide. 4. An offer. The second step is the expensive one. In a 2019 guide, 500 Global described the start of its process as "a 30 to a 45-minute phone screen", followed by "two twenty-minute group interviews". Multiply a 30-minute call by every applicant and the hours run out, so a programme with thousands of applications can only call a share of them. A founder with a sharp insight and a plain application can miss that call, while partner hours go to calls that were a clear no in the first five minutes. ## What can't an accelerator application show you? A form shows how well someone writes about their company. A one-minute video shows how well they pitch to a camera. Neither can ask the second question, which is where a founder's story comes out. Ask "What are you building?" and you get the one-line pitch. Ask "Who was the first person you saw with this problem?" and you hear how they found it. Ask "Where do you want this company to be in five years, and what has to be true for that?" and you hear the ambition, and whether they've thought about what stands in the way. ## How does Tracy give every founder a first accelerator interview? ![Diagram of an accelerator first round. Today: a row of founders, of whom only a few get a partner call while the rest are judged on a form or a one-minute video. With Tracy: every founder in the row gets a first conversation, and partners meet the strongest three, two of whom got no partner call before.](/blog/images/accelerator-interview-first-round.svg) *With every founder interviewed, partners find strong founders the form-only round passed over.* Tracy AI Interviewer interviews like your best partner on a first call. It runs natural two-way video interviews, asks your questions, then "follows up on vague answers, digs into the interesting ones": the customer behind a claim, the reason behind a pivot, the plan behind an ambition. Partners walk into their calls already knowing each founder's story. For a technical co-founder, Tracy can "introduce a coding task, and probe the trade-offs behind an answer", so you learn whether the team in front of you can build the product. Our page on [AI technical interviews](/blog/ai-technical-interviews?utm_source=blog&utm_medium=seo&utm_campaign=accelerator-interview&utm_content=internal-link) shows how. ![A founder applying to a programme. Tracy's question on the left asks about the traction they just described: "Four of eleven clinics pay. What do those four have in common?" A short form with an uploaded pitch deck sits beside it and is optional.](/blog/images/lib-founder-pitch-accelerator.png) *The founder talks it through with Tracy, and the short form with a deck upload is optional (illustration with sample data).* Tracy meets every founder. Each one opens a link in their browser and joins with no sign-up, whenever it suits them, so the whole intake can get a first conversation without adding a reviewer. Our guide on how to [interview every applicant in depth](/blog/how-to-screen-candidates?utm_source=blog&utm_medium=seo&utm_campaign=accelerator-interview&utm_content=internal-link) covers the volume side. Partners read instead of scheduling. Each founder's result is a readable summary "with the participant's own words attached", with every answer organised around the fields you define and the video and transcript beside it. The AI explains each answer score, so a partner can check a strong answer and invite that founder the same day. ![A Tracy result report for a founder application. Fields for the problem, evidence so far, what they learned and why this team are marked Stated, each quoting the founder. The recording and timestamped transcript sit above them.](/blog/images/lib-result-traced-founder.png) *Partners see each founder's answers in their own words and choose who to invite (illustration with sample data).* **[Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=accelerator-interview&utm_content=try-free)** or **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=accelerator-interview&utm_content=demo)** ## What accelerator interview questions should the first round ask? Here is a first round we would set up for a pre-seed programme. Tracy adapts each follow-up to the founder's answer. For why the follow-up carries the weight, see our [AI trainer interview questions](/blog/ai-trainer-interview-questions?utm_source=blog&utm_medium=seo&utm_campaign=accelerator-interview&utm_content=internal-link), written for expert vetting but built on the same idea. | Question | Tracy follows up on | |---|---| | What are you building, and who needs it most? | The last time they spoke to one of those people, and what they heard | | How did you come to this problem? | What they saw or lived through that made them start | | What do those people use today instead? | Why they haven't switched to something better already | | What's the most important thing you've learned from users, and what did you change? | Evidence that the change worked | | Where do you want this company to be in five years? | What has to be true to get there, and what they'd do first with the programme | | Why is your team the one to build this? | The experience behind the answer | Example: an illustrative follow-up chain from the same demo intake, with test data, not a real founder. > **Tracy:** How many clinics use the pilot today, and how many pay? > > **Founder:** Eleven clinics use our pilot. Four of them pay. > > **Tracy:** Four of eleven clinics pay. What do those four have in common? > > **Founder:** They're the bigger ones, with night shifts. Night cover is where the group chat breaks down, so that's where we save managers the most time. > > **Tracy:** Where do you want this company to be in five years? > > **Founder:** Booking shifts for clinics across the region. To get there we need payroll too, because managers won't run two systems. Three questions in, a partner knows who pays, why, and what the founder plans to build next. A form would have shown "11 pilot clinics". ## When are partner first calls the better choice? Partner calls still win on relationship. Pick them for the first round instead if: - Your partners can call most of your applicants each intake. - The programme is small enough that every founder already gets a conversation. Once applications outrun partner hours, Tracy works best as the step before the partner call, so the relationship starts with the founders most worth meeting. If you use one-way video today, our list of [HireVue alternatives](/compare/hirevue-alternatives?utm_source=blog&utm_medium=seo&utm_campaign=accelerator-interview&utm_content=internal-link) shows what changes when the tool can follow up. For the wider category, see the [best AI interview software](/blog/best-ai-interview-software?utm_source=blog&utm_medium=seo&utm_campaign=accelerator-interview&utm_content=internal-link). ## How do you set it up before your next intake? 1. Write down what partners want to know after a first call. Three to five fields is enough: what they're building, how they found the problem, evidence, learning and ambition. 2. Turn those fields into questions. Bring your own, or describe who you're interviewing and let Tracy draft them; "your questions are drafted for you", and you add the questions that matter. 3. Take the interview yourself, the way a founder would, and change anything that feels off. 4. Add the link to your application confirmation, or send it to everyone who applies. 5. Read the results each week and invite the strongest founders to a partner interview. Keep the recording for the partner who meets them. 6. Tell applicants that an AI runs the first interview and that partners make every decision. You pay for results, not seats, and a session with no answers is never charged. Enterprise pricing is custom, based on your volume, so the cost follows the size of each intake. ## Common questions ### Does an AI interview replace the partner interview? No. It gives every applicant the first conversation that only a few get today. Partners still run the interviews that decide. ### Will founders find it impersonal? Tracy follows up on vague answers and digs into the interesting ones, so it plays out as a conversation, and candidates rate Tracy 4.7 out of 5 on average. A separate study involved about 70,000 job applicants; among those offered a choice, 78% chose the AI interviewer. The study did not involve founders, so watch your own completion rate in the first intake. ### How long should the first-round interview be? Five or six questions with follow-ups make a good first round. One credit covers up to 20 minutes. ### Do these work as startup incubator interview questions too? Yes. Incubators, fellowships and grant programmes use the same steps and change only the fields, for example research plan, motivation and track record. ### Can partners see why a founder was ranked high? Yes. The AI explains each answer score, and the summary has the founder's own words attached, with direct quotes, the interview video and transcript. ### I'm a founder with an accelerator interview coming up. How should I prepare? Know your story in your own words: who has the problem, how you found it, what users taught you and what you changed, and where you want the company to be in five years. Expect a follow-up on every answer, so bring the specifics behind each claim, such as the customer, the number or the date. 500 Global's own interview guide walks through its stages. --- **[Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=accelerator-interview&utm_content=try-free)** or **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=accelerator-interview&utm_content=demo)** ## Sources - 500 Global, "Startup accelerator interview: how to ace your interview with 500 Global", 5 Mar 2019, [500.co](https://500.co/content/startup-accelerator-interview-how-to-ace-your-interview-with-500-global), retrieved 29 Oct 2026: "You'll begin the startup accelerator interview process with a 30 to a 45-minute phone screen"; "There will be two twenty-minute group interviews with around three to five people". - Jabarian and Henkel, "Voice AI in Firms: A Natural Field Experiment on Automated Job Interviews", Aug 2025, abstract at [Luohan Academy](https://www.luohanacademy.com/en/webinar/1111/), retrieved 29 Oct 2026: "70,000 applicants were randomly assigned to be interviewed by human recruiters, AI voice agents"; "when offered the choice, 78% of applicants choose the AI recruiter". Full paper: [SSRN 5395709](https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5395709). - Qode, [qode.world/tracy](https://qode.world/tracy), retrieved 29 Oct 2026: "4.7/5 average candidate rating across interviews". - Qode, [qode.world/pricing](https://qode.world/pricing), retrieved 29 Oct 2026: "Video & chapter breakdown", "Transcript", "Answer summary & score", "Explain answer score by AI Agent"; Enterprise "Custom based on volume". - Qode, 28 Oct 2026: enterprise pricing is custom, based on volume. - Tracy AI Interviewer, [tracyinterview.ai](https://tracyinterview.ai), retrieved 29 Oct 2026: "Tracy AI Interviewer runs natural two-way video interviews, follows up on vague answers, digs into the interesting ones"; "Tracy AI Interviewer can ask role-specific questions, introduce a coding task, and probe the trade-offs behind an answer"; "Your questions are drafted for you." and "add the questions that matter"; "Each person opens a link in their browser and joins with no sign-up."; "People join in their browser whenever it suits them."; "Each interview becomes a readable summary with the participant's own words attached"; "Direct quotes for context"; "your own repeatable interview, with every answer organized around the fields you define." - Tracy AI Interviewer pricing, [tracyinterview.ai/pricing](https://tracyinterview.ai/pricing), retrieved 29 Oct 2026: "$1 per credit, in USD. Credits never expire."; "1 credit for each started 20 minutes, charged when the result is ready."; "You pay for results, not seats."; "A session with no answers is never charged." # AI Interview Cheating Detection: How to Tell If a Candidate Is Using AI > AI interview cheating detection in two layers: follow-ups that break AI-fed answers, and integrity signals that start a review. Steps for any interview. - URL: https://tracyinterview.ai/blog/ai-interview-cheating-detection - Published: 2026-11-02 - Updated: 2026-11-13 - Author: Tracy team > **In short** > - **Tracy AI Interviewer gives reviewers better evidence when an answer raises a concern.** It asks follow-ups based on each candidate's answer and flags integrity signals, such as a tab switch or a late answer, for a closer look. > - Reviewers can check the answers, video and transcript before deciding. A signal calls for review; it does not establish that someone cheated. **[Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=ai-interview-cheating-detection&utm_content=try-free)** or **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=ai-interview-cheating-detection&utm_content=demo)** **In this guide** - [How does AI interview cheating detection work?](#how-does-ai-interview-cheating-detection-work) - [What does a follow-up chain look like?](#what-does-a-follow-up-chain-look-like) - [What does AI interview cheating detection show the reviewer?](#what-does-ai-interview-cheating-detection-show-the-reviewer) - [How common is interview cheating with AI?](#how-common-is-interview-cheating-with-ai) - [What interview cheating tools do candidates use?](#what-interview-cheating-tools-do-candidates-use) - [How to tell if someone is using AI in an interview](#how-to-tell-if-someone-is-using-ai-in-an-interview) - [Should you bring back in-person interviews?](#should-you-bring-back-in-person-interviews) - [What should you do when a signal fires?](#what-should-you-do-when-a-signal-fires) - [Questions](#questions) - [Sources](#sources) AI interview cheating detection works best in two layers. First, ask questions that build on what the candidate just said, which no script can prepare for. Second, watch for signals such as tab changes or reading eye movements, and treat each one as a reason to review. A person makes the decision. This guide shows how Tracy does both for every applicant, then gives steps you can use in any interview. ## How does AI interview cheating detection work? Tracy runs the first round as a natural two-way video interview and handles cheating in two layers. ![Diagram of AI interview cheating detection in two layers. Layer 1, follow-ups on the candidate's own answer, makes rehearsed and AI-fed answers thin. Layer 2, integrity signals: tab switching, leaving fullscreen, an external screen, face and gaze, response timing and repeated requests for a question. Then a person reviews the evidence and decides.](/blog/images/ai-interview-cheating-detection-layers.svg) *Follow-ups do most of the work; signals tell your reviewer where to look.* **Layer 1: follow-ups.** Tracy asks role-specific questions and follows up on vague answers, "a real follow-up, not a rigid script", so you hear how deep a candidate's knowledge goes, the way your best interviewer would. For engineers it can set a coding task and probe the trade-offs behind the answer, so a copied solution still has to be explained. **Layer 2: signals.** Tracy captures integrity signals during the interview: the face and where the person is looking, response timing, tab switching, leaving fullscreen, an external screen and repeated requests for a question. The report counts them, for example how often the person left the tab, how many answers started late and how long the average answer took to start. They appear in the interview report, so your reviewer knows which interviews to look at closely instead of rewatching all of them. **A person decides.** Each interview becomes a summary "with the participant's own words attached" and direct quotes for context, and the AI explains each answer score, next to the video and transcript, so the reviewer judges from evidence and makes the call. It runs from tools you already use, such as Slack, WhatsApp and Workable, so it fits the process you run today. ## What does a follow-up chain look like? A copilot can feed a good answer to a known question. It helps much less when the next question is about the answer itself. ![The Integrity signals section of a Tracy result report. Seven signals with counts: the participant left the interview tab once and started one answer late, the average time to start an answer was 2.6 seconds, and the other four show zero.](/blog/images/lib-integrity-signals-one-flag.png) *The report counts each signal so your team knows where to look in the recording. Tracy reports them and your team decides what they mean (illustration with sample data).* Example: an illustrative first-round chain for a finance domain expert, the kind of role AI-lab data platforms vet at volume. It is written for this guide and is not a real candidate. | Step | Tracy asks | A strong answer | An AI-fed answer tends to | | --- | --- | --- | --- | | 1 | "How would you value a SaaS company with negative free cash flow?" | A method, and why it fits | Give a correct textbook summary | | 2 | "You said revenue multiple. Which comparables for 70% gross margin and 20% growth?" | Named companies or a clear rule | Repeat that comparables should be "similar" | | 3 | "Two of those were acquired last year. Keep them in the set?" | A judgment, such as acquisition premiums skewing the multiple | Pause, then a general paragraph on M&A | | 4 | "When did you last get a valuation wrong, and what did you change?" | A specific deal, the mistake and the fix | Offer no example, or one with no detail | Step 1 is easy to fake. At step 3 the candidate has to reason about their own earlier choice under a new condition, live, and that is where a real expert gets more specific. ## What does AI interview cheating detection show the reviewer? Here is part of a report for the chain above. It is an illustration with test data. | Field | What the candidate said | | --- | --- | | Choosing comparables | "I'd drop the two acquired companies, because the deal price includes a control premium the public multiple doesn't." | | Last valuation mistake | "Valuation always involves uncertainty, and you should use several methods." | | Integrity signals | Late answers; unusual eye movement | The reviewer sees two specific answers, one generic one, and two signals. They open the video at the generic answer, watch it and decide whether to invite the candidate to a live call. Screening every applicant this way costs only what you use: $1 per credit; one credit covers up to 20 minutes. You pay for results, not seats, and a session with no answers is never charged. At high volume, enterprise pricing is custom, based on your volume. ## How common is interview cheating with AI? Two in three US hiring managers in Greenhouse's survey have caught it. - **Assessments.** CodeSignal found that cheating and fraud attempt rates on its proctored assessments more than doubled, "rising from 16 percent in 2024 to 35 percent" in 2025. These are attempts its system flagged, not confirmed successes. - **Live interviews.** In Greenhouse's 2025 AI in Hiring Report, 65% of US hiring managers had caught applicants using AI deceptively, including "reading from AI-generated scripts (32%)" and "showing up as deepfakes (18%)". - **Employers.** Google, McKinsey and Cisco have brought back face-to-face steps, according to Round Zero. For platforms that supply experts to AI labs, one fake expert can cost the lab's project. For technical hiring teams, AI-polished applications already look alike, and an AI-fed interview removes the last place depth showed. ## What interview cheating tools do candidates use? The best-known interview cheating AI tools are desktop apps that listen to the call or read the screen, then show a suggested answer in an overlay only the candidate sees. Cluely says it "gives real-time answers, all while completely undetectable" and "never shows up in shared screens, recordings, or external meeting tools." Interview Coder targets technical interviews, with "Support for all kinds of interviews(leetcode, system design, etc.)" and an overlay that "remains hidden when sharing your screen." Screen sharing won't show these tools, and their answers are fluent. They are weakest against a question that depends on the candidate's earlier answer, asked straight after it. ## How to tell if someone is using AI in an interview Watch for the same signals Tracy checks: 1. **Tab switching.** The window loses focus while a question is open. 2. **Leaving fullscreen.** The interview is shrunk or left while answering. 3. **An external screen.** Signs of a second device, such as turning to one fixed side before each answer. 4. **Face and gaze.** The face leaves the camera, or the eyes move in steady lines as if reading a script. 5. **Response timing.** A pause before every answer, even easy ones, then a fluent paragraph, often opened with a stock filler line. 6. **Repeated requests for a question.** Asking for questions again and again can buy time for a lookup. None of these proves cheating. People look away to think and pause when nervous. Check the answers too. Watch for a polished first answer with a vague follow-up, correct terms with no example from their own work, or spoken answers that don't match the depth of the written application. Seven steps that work in any video interview, with or without Tracy: 1. **Say the rules at the start**, including whether AI, notes or a second screen are allowed. 2. **Ask about their own work.** "What was the worst result you got, and what did you change?" needs a memory a copilot doesn't have. 3. **Follow up on the detail they just gave**, then follow up again. 4. **Change the problem halfway**: double the traffic, cut the budget, remove the library they used. 5. **Ask them to explain one line** of their code, or the assumption they are least sure of. 6. **Keep the camera on and note timing.** One long pause is nothing; a pause before every answer is a pattern. 7. **Compare the interview with the application** and look closer when the depth doesn't match. For technical roles, our guide to [AI technical interviews](/blog/ai-technical-interviews?utm_source=blog&utm_medium=seo&utm_campaign=ai-interview-cheating-detection&utm_content=internal-link) shows a full chain on a coding task. For expert vetting, see [interview questions for AI trainers and domain experts](/blog/ai-trainer-interview-questions?utm_source=blog&utm_medium=seo&utm_campaign=ai-interview-cheating-detection&utm_content=internal-link). ## Should you bring back in-person interviews? For the final round, often yes. Round Zero reports that "Google reinstated at least one in-person round for technical hires" and that "McKinsey now asks hiring managers to meet candidates face to face before an offer goes out." An onsite round proves the person is real. It also costs travel and interviewer hours for every finalist, and it can't cover a first round of thousands of applicants across countries. A remote first round with deep follow-ups and signals, then an in-person final for a short list, puts the onsite hours on people who already showed depth. Our guide on [how to screen candidates](/blog/how-to-screen-candidates?utm_source=blog&utm_medium=seo&utm_campaign=ai-interview-cheating-detection&utm_content=internal-link) covers the first-round side. ## What should you do when a signal fires? 1. **Read the answers first.** Specific, consistent follow-ups outweigh a single pause. 2. **Watch the flagged interview.** Many flags have an innocent cause, like a noise or a candidate thinking. 3. **Look for a pattern** across questions, not one moment. 4. **Ask for a short live call** on the same topic when in doubt. 5. **Write down why** you decided, so reviewers stay consistent. 6. **Tell candidates up front** that an AI runs the interview, what it checks, and that a person decides. What you must disclose depends on where you hire, so check with your own counsel. Comparing tools? See the [best AI interview software for hiring teams](/blog/best-ai-interview-software?utm_source=blog&utm_medium=seo&utm_campaign=ai-interview-cheating-detection&utm_content=internal-link) or [Tracy vs HireVue](/compare/tracy-vs-hirevue?utm_source=blog&utm_medium=seo&utm_campaign=ai-interview-cheating-detection&utm_content=internal-link). ## Common questions ### Can AI interview cheating detection catch tools like Cluely? Their makers say the overlay is hidden from screen sharing. The candidate's behaviour still shows: reading eye movements, a pause before each answer, a second screen. Follow-ups on their own answer are harder for any copilot to keep up with. ### What are interview cheating detection tools? Software that flags signs of cheating during an interview or test, such as tab changes, a second face or a second screen. In Tracy, the signals sit in the same report as the follow-up answers, so a reviewer judges both together. ### Is interview cheating AI getting more common? Yes. CodeSignal's flagged attempts more than doubled from 2024 to 2025, and a third of US hiring managers in Greenhouse's survey had caught an applicant reading an AI-generated script. ### Does Tracy tell me a candidate cheated? It flags signals for a person to check, next to what the candidate said. Your reviewer looks at the answers and the video, and decides. ### Is it fair to nervous candidates? Yes, when signals start a review and a person decides. Nervous candidates pause too, but their follow-up answers stay specific. --- **[Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=ai-interview-cheating-detection&utm_content=try-free)** or **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=ai-interview-cheating-detection&utm_content=demo)** ## Sources - CodeSignal, "CodeSignal detection systems identify and stop record-high cheating attempts as assessment fraud more than doubled in 2025", press release, 25 Feb 2026. [codesignal.com](https://codesignal.com/newsroom/press-releases/codesignal-detection-systems-identify-and-stop-record-high-cheating-attempts-as-assessment-fraud-more-than-doubled-in-2025/). Retrieved 29 Oct 2026. Quote: "cheating and fraud attempt rates for proctored assessments more than doubled, rising from 16 percent in 2024 to 35 percent." Also: "these signals reflect fraud and cheating attempts that were detected and flagged, not successful circumvention of assessment standards." - Greenhouse, 2025 AI in Hiring Report release, 19 Nov 2025. [greenhouse.com](https://www.greenhouse.com/newsroom/an-ai-trust-crisis-70-of-hiring-managers-trust-ai-to-make-faster-and-better-hiring-decisions-only-8-of-job-seekers-call-it-fair). Retrieved 29 Oct 2026. Quote: "companies in the U.S. cite an integrity crisis. ... Two in every three (65%) hiring managers have caught applicants using AI deceptively, like reading from AI-generated scripts (32%), hiding prompt injections in resumes (22%), or showing up as deepfakes (18%)." - Round Zero, "In-person interviews are back", Danish Siddique, 25 Aug 2026. [joinroundzero.com](https://joinroundzero.com/blog/in-person-interviews-are-back). Retrieved 29 Oct 2026. Quote: "Google reinstated at least one in-person round for technical hires. McKinsey now asks hiring managers to meet candidates face to face before an offer goes out. Cisco and others have made similar moves." - Cluely, home page. [cluely.com](https://cluely.com/). Retrieved 29 Oct 2026. Quote: "Cluely takes perfect meeting notes and gives real-time answers, all while completely undetectable." Also: "Cluely never shows up in shared screens, recordings, or external meeting tools." - Interview Coder, home page. [interviewcoder.co](https://www.interviewcoder.co/). Retrieved 29 Oct 2026. Quote: "Support for all kinds of interviews(leetcode, system design, etc.)" Also: "Overlay remains hidden when sharing your screen, ensuring interviewers can't detect the assistance tool visually." - Tracy AI Interviewer, home page. [tracyinterview.ai](https://tracyinterview.ai). Retrieved 29 Oct 2026. Quote: "Tracy AI Interviewer runs natural two-way video interviews, follows up on vague answers, digs into the interesting ones, and turns every conversation into structured results." Also: "Tracy AI Interviewer can ask role-specific questions, introduce a coding task, and probe the trade-offs behind an answer." "A real follow-up, not a rigid script." "Each interview becomes a readable summary with the participant's own words attached." "Direct quotes for context." - Tracy AI Interviewer, pricing. [tracyinterview.ai/pricing](https://tracyinterview.ai/pricing). Retrieved 29 Oct 2026. Quote: "$1 per credit, in USD. Credits never expire." Also: "You pay for results, not seats."; "1 credit for each started 20 minutes"; "A session with no answers is never charged." - Tracy AI Interviewer, product page: Integrity. [tracyinterview.ai/product](https://tracyinterview.ai/product#integrity). Retrieved 13 Nov 2026. Quote: "Tracy captures signals during the interview: the face and where the person is looking, response timing, tab switching, leaving fullscreen, an external screen and repeated requests for a question." - Qode, 28 Oct 2026: enterprise pricing is custom, based on volume. - Qode, confirmed product facts (integrations: ChatGPT, Claude, Gemini, Mistral, Grok, Slack, WhatsApp, Zapier, Workable, HubSpot), 29 Oct 2026. # AI Interviews for Accelerator, Fellowship and Program Applications > How accelerators and fellowships can give every applicant a first interview, what partners read afterwards, and how to set it up before your next intake. - URL: https://tracyinterview.ai/blog/ai-interviews-for-applications - Published: 2026-11-02 - Updated: 2026-11-11 - Author: Tracy team Most accelerators read every application but talk to only a few of the people behind them. An AI interview gives every applicant a short first conversation, so partners can choose who to meet from what each founder said. It works the same way for fellowships, grant programs and any program that picks people from an open application. ## How programs choose today Most programs run some version of the same funnel: 1. An application form, sometimes with a short founder video. 2. A first call with a partner, a venture partner or an alum. 3. A longer interview with the partners who decide. 4. An offer. The second step is the expensive one. In a 2019 guide, 500 Global described its first step as "a 30 to a 45-minute phone screen with one of our venture partners or entrepreneurs-in-residence." Multiply that by every applicant and the hours run out, so most programs give first calls to a small share of applicants and judge the rest on the form. ## What a form can't tell you A form shows how well someone writes. It can't ask a follow-up question. Founders can now polish an application with AI in an afternoon, and in our view that makes the strong forms and the weak ones harder to tell apart. The founder with a sharp insight and a rough application rarely gets the call, while partner hours go to calls that are a clear no early on. ## What an AI interview changes ![Two application funnels. Today: application form, a first call for a few applicants, partner interview, decision. With an AI interview: application form, an AI interview for every applicant, partner interview picked from the answers, decision.](/blog/images/first-round-today-vs-ai-interview.svg) Today only a few applicants get a first call. With an AI interview every applicant gets one, and partners pick who to meet from the answers. - **Every applicant gets a first conversation.** Add one link to your application confirmation. Founders join from their browser, with no account to create. - **It's a two-way conversation.** Tracy at Qode runs a video interview that asks your questions and follows up on each answer, the way a partner would. - **Partners read instead of scheduling.** Each founder gets a summary organized by the fields your partners care about, backed by the founder's own words, with the recording and transcript side by side. See AI interview summaries for how a summary is laid out. - **People still decide.** Scoring is optional. Partners choose who to meet from what the founders said. The partner call stays where it is. Founders now reach it on the strength of what they said in the first round, and the partner who meets them already has their answers on file. A founder who writes a plain application but explains the business well on video now gets the same look as one with a polished form. ## A sample first-round interview Here is a first round we would set up for a pre-seed program. Each question has a follow-up that Tracy adapts to the founder's answer. | Question | Tracy follows up on | |---|---| | What problem are you solving, and who has it worst? | The last time they spoke to one of those people, and what they heard | | What do those people use today instead? | Why they haven't switched to something better already | | What do you know about this market that others don't? | The specific claim, and how they found it out | | What's the most important thing you've learned from users, and what did you change? | Evidence that the change worked | | Why is your team the one to build this? | The experience behind the answer | This is what a partner reads afterwards. It's an illustration with test data, not a real founder. | Field | What the founder said | |---|---| | Problem | "Clinic managers still book nurses through group chats. One told me she spends two hours a day on it." | | Evidence so far | "Eleven clinics use our pilot. Four of them pay." | | Learning | "We dropped the patient app. Managers only cared about filling shifts." | ![A founder in a Tracy first-round interview in the browser. Tracy asks a follow-up about the traction they just described.](/blog/images/founder-first-round-interview.png) *A founder's first-round interview. Tracy follows up on the answer they just gave.* ![A Tracy result report for a founder application. Fields for the problem, evidence so far, what they learned and why this team are marked Stated, each quoting the founder. The recording and timestamped transcript sit above them.](/blog/images/founder-interview-result.png) *What the partner reads afterwards: every answer quoted and linked to its moment in the recording. Partners decide who to meet.* ## First-round options compared | Option | What partners learn | Partner time | Founder experience | |---|---|---|---| | Form only | What the founder wrote, or what AI helped them write | Reading time | No conversation | | One-way video | A recorded pitch, with no follow-up | Watching each video | Talking to a camera | | Partner first calls | The most, and the start of a relationship | A call per founder, so only a few get one | The best, for the few who get one | | Two-way AI interview | Answers to your questions, with follow-ups, in the founder's own words | Reading a summary | A real conversation, for every applicant | Partner calls still win on relationship. If your partners can call most of your applicants, keep doing it. An AI interview helps once applications outrun partner hours, and it works best as the step before the partner call. AI interview vs human interview covers what each does better. ## How to set it up before your next intake 1. **Write down what partners want to know after a first call.** Three to five fields is enough: problem, evidence, insight, learning, team. 2. **Turn those fields into questions.** Describe who you're interviewing and what you want to learn; Tracy drafts the questions and you edit them. 3. **Take the interview yourself,** the way a founder would, and change anything that feels off. 4. **Add the link** to your application confirmation, or send it to everyone who applies. 5. **Read the summaries each week,** field by field, and invite the strongest founders to a partner interview. Keep the recording for the partner who meets them. 6. **Tell applicants** that an AI runs the first interview and that partners make every decision. Are AI interviews legal? covers what to disclose. Pricing is per result, with no seats to buy: $1 per credit, and one credit covers up to 20 minutes. The cost follows the size of each intake. See [Tracy pricing](https://tracyinterview.ai/pricing?utm_source=blog&utm_medium=seo&utm_campaign=ai-interviews-for-applications&utm_content=internal-link). ## Common questions ### Does an AI interview replace the partner interview? No. It gives every applicant the first conversation that only a few get today. Partners still run the interviews that decide. ### Will founders find it impersonal? It responds to what each founder says and follows up, so it plays out as a conversation. In a field study of about 70,000 job applicants, 78% chose the AI interviewer when they were offered the choice. Those were applicants for entry-level customer service jobs, not founders, so watch your own completion rate in the first intake. ### Can it tell whether AI wrote the application? It doesn't try to judge that. It asks founders to explain their own answers out loud, with follow-ups, which tells partners more than a guess about the form. How to detect AI cheating in interviews covers the signs to watch for. ### How long should the interview be? You choose the questions, so you choose the length. A short first round of about five questions is easy to finish. ### Does it work for fellowships and grant programs? Yes. The steps are the same and only the fields change, for example research plan, motivation and track record. How do AI interviews work? explains the format from the applicant's side. --- **[Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=ai-interviews-for-applications&utm_content=try-free)** or **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=ai-interviews-for-applications&utm_content=demo)** ## Sources - 500 Global, "Startup accelerator interview: how to ace your interview with 500 Global", 5 Mar 2019: [500.co](https://500.co/content/startup-accelerator-interview-how-to-ace-your-interview-with-500-global). Retrieved 26 Oct 2026. - Jabarian and Henkel, "Voice AI in Firms: A Natural Field Experiment on Automated Job Interviews", Aug 2025: [SSRN 5395709](https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5395709). Retrieved 26 Oct 2026. - Tracy at Qode product facts: [tracyinterview.ai](https://tracyinterview.ai). Retrieved 26 Oct 2026. # AI Technical Interviews: Coding, System Design and Debugging Screens > How an AI technical interview runs coding, system design and debugging screens, with a sample follow-up chain and when a test library fits better. - URL: https://tracyinterview.ai/blog/ai-technical-interviews - Published: 2026-11-02 - Updated: 2026-11-11 - Author: Tracy team > **In short** > - **Tracy AI Interviewer gives every engineering applicant a coding or system-design screen with adaptive follow-ups.** Reviewers can see the candidate's reasoning and trade-offs, then check each score against their own words and the recording. > - Choose a test library for a standardized scored task. Choose Tracy when how someone explains and adapts their own solution is the hiring signal. **[Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=ai-technical-interviews&utm_content=try-free)** or **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=ai-technical-interviews&utm_content=demo)** **In this guide** - [Why use Tracy as your AI technical interview platform?](#why-use-tracy-as-your-ai-technical-interview-platform) - [What can an AI technical interview cover?](#what-can-an-ai-technical-interview-cover) - [What does an AI coding interview follow-up chain look like?](#what-does-an-ai-coding-interview-follow-up-chain-look-like) - [What does the reviewer get after an AI technical interview?](#what-does-the-reviewer-get-after-an-ai-technical-interview) - [How does Tracy compare with CodeSignal and HackerRank?](#how-does-tracy-compare-with-codesignal-and-hackerrank) - [How do you set up an AI technical interview tool?](#how-do-you-set-up-an-ai-technical-interview-tool) - [Questions](#questions) - [Sources](#sources) An AI technical interview gives every applicant a technical first round with follow-up questions, like the ones a senior engineer asks. A test tells you whether the code passes; an AI technical interview also asks why it was written that way and what changes when the requirements move. This guide shows what that looks like on a real coding task, and how Tracy compares with CodeSignal and HackerRank. ## Why use Tracy as your AI technical interview platform? **It interviews like your best engineer.** Tracy runs natural two-way video interviews, follows up on vague answers, and can introduce a coding task and probe the trade-offs behind an answer. So you hear how a candidate reasons about complexity, scale and testing in their own solution, which is where a strong engineer and someone who memorised the pattern separate. Because the follow-ups come from the candidate's own answer, a copied solution still has to be explained. **Every point is traced to the answer.** Each result is a readable summary "with the participant's own words attached", with every answer organised around the fields you define and the video and transcript alongside. The AI explains each answer score, and direct quotes give the context. Your engineers check the evidence for the strongest candidates instead of running another phone screen. Tracy's interview integrity detection flags tab changes, abnormal eye tracking, live face detection, average delay before an answer, external screens and other unusual patterns, so a copied answer gets a second look (see [how AI interview cheating detection works](/blog/ai-interview-cheating-detection?utm_source=blog&utm_medium=seo&utm_campaign=ai-technical-interviews&utm_content=internal-link)). Candidates rate Tracy 4.7 out of 5 on average. ## What can an AI technical interview cover? Tracy's technical interviews cover four task types: coding task, system design, debugging and code walk-through. Pick the ones that match the role. - **Coding task.** A short problem solved live, then questions on the trade-offs, so you learn whether they understand their own solution. - **System design.** For example, "Design a rate limiter," then follow-ups on concurrency, failure modes and testing, so you see how senior engineers handle open problems. - **Debugging.** A bug and a failing symptom, so you see how they would find the cause in a real codebase. - **Code walk-through.** The candidate explains code they wrote or a sample you provide, so you hear the choices behind it. For AI-lab data platforms, the same format vets coding experts who write and review training data for labs. For other expert fields, see [interview questions for AI trainers and domain experts](/blog/ai-trainer-interview-questions?utm_source=blog&utm_medium=seo&utm_campaign=ai-technical-interviews&utm_content=internal-link). ## What does an AI coding interview follow-up chain look like? Example: an illustrative chain for a backend engineer, written for this guide. It is not a real candidate. **The task.** "Write a function that reads a stream of log events and returns the 10 most frequent error codes in the last hour." The candidate uses a hash map of counts, a min-heap of size 10, and a queue that drops events older than an hour. ![Staircase diagram of five follow-up questions on the top-10 error codes task. Step 1, why a heap and not a sort, tests the cost of each approach. Step 2, what happens at 50,000 events a second, tests memory under load. Step 3, how codes leave the heap, tests spotting a flaw in their own design. Step 4, whether an approximate top 10 is fine, tests trading memory for accuracy. Step 5, how they would test it given time, tests testing code that depends on a clock.](/blog/images/ai-technical-interview-follow-up-chain.svg) *Each follow-up climbs from the last answer, and step 3 is where a memorised pattern runs out.* 1. **"You used a heap of size 10. Why that over sorting the counts each time?"** A strong answer compares O(n log k) with O(n log n) and says when the gap matters. 2. **"Your queue keeps every event from the last hour. What happens at 50,000 events a second?"** A strong answer moves to per-second buckets, so memory depends on the window length instead of the traffic. 3. **"With buckets, how do you remove an error code from the heap when its count drops?"** A strong answer spots that a plain heap can't update counts cheaply and proposes a fix, such as rebuilding it each second. 4. **"Say an approximate top 10 is fine. Would you change anything?"** A strong answer names a sketch such as count-min and the trade: less memory for some error. 5. **"How would you test this, given that it depends on time?"** A strong answer injects a clock, tests the window edge and handles events that arrive out of order. At step 1, someone who memorised "use a heap for top-k" gives the right words. By step 3 they have to reason about their own design under a new condition, live. ## What does the reviewer get after an AI technical interview? ![A candidate's view of a Tracy coding interview. The problem and test cases sit above the code editor and output, and Tracy's conversation and the video tiles sit on the right.](/blog/images/lib-coding-interview-remove-duplicates-en.png) *The candidate codes in the same room where Tracy asks the follow-ups (illustration with sample data).* For the coding chain above, the report would read like this illustration with test data: | Field | What the candidate said | | --- | --- | | Complexity trade-off | "The heap keeps it at n log 10, so basically linear. Sorting re-sorts every event." | | Scaling the window | "I'd keep 3,600 one-second buckets, so memory is fixed whatever the traffic." | | Heap updates | "A plain heap can't decrease a key cheaply. I'd rebuild it once a second from the buckets." | | Testing | "I'd pass the clock in, so a test can jump 59 minutes and check the edge." | Your team reads the fields, opens the video where it wants more, and decides who reaches the panel. ## How does Tracy compare with CodeSignal and HackerRank? CodeSignal and HackerRank built their names on test libraries and have since added AI interviewers. | | Tracy | CodeSignal | HackerRank | | --- | --- | --- | --- | | **Interviews like your best expert** | Follows up on vague answers and probes the trade-offs behind an answer on a coding task, with role-specific questions | Assessments "validated by IO psychologists"; its AI Interviewer "probes and asks follow up questions" when answers are thin. Plans list "Sales, CS, & Marketing AI Interviewers" and "Product, Design, & Engineering AI Interviewers" | Chakra "asks follow-up questions on the candidate's work to see how well they're able to defend their decisions" | | **Thousands of interviews, zero new hires** | Applicants join from a link in their browser with no sign-up, whenever it suits them | Build plan: "60 Annual credits" for "Assessments & Interviews" | Starter: "60 candidate attempts / year"; Pro: 360 | | **Work with your stack** | Workable, HubSpot, Zapier, Slack, WhatsApp, and AI assistants such as ChatGPT and Claude; bring your own questions | "Basic ATS Integrations" on the Grow plan | "ATS and calendar integrations" on the Pro plan | | **Every score, traced to the answer** | A summary by the fields you define, with the candidate's own words attached and each answer score explained by the AI, plus video and transcript | "an in-depth skills report + full transcript with rationale" | "a clear 'why' for every score" | | **Price** | $1 per credit; one credit covers up to 20 minutes. You pay for results, not seats. Enterprise: custom, based on your volume | Build from $79 a month billed annually; extra credits $20 each | Starter from $79 a month, $948 billed annually, Chakra included; extra attempts $20 each | Tracy starts with a conversation and can put a task inside it, so reviewers hear how each applicant reasons about the solution. CodeSignal and HackerRank also offer AI interviewers alongside their test libraries. The same Tracy interviewer covers non-engineering experts, which matters if you vet both. **Pick CodeSignal or HackerRank instead if:** - You screen thousands of similar applicants, such as new graduates, and want one standard auto-scored test. Tracy then fits as the next step for those who pass. - You need validated, comparable test scores over years. CodeSignal says each assessment is "backed by 2,800 hours of research and testing"; Tracy can sit after that test to hear how the top scorers reason. - You want a large rotated question bank. CodeSignal's Build plan includes "Access to 2,000 Questions", and HackerRank's Starter plan lists "2000+ library questions". If your roles need your own questions, Tracy lets you bring them. Our guide on [how to screen candidates](/blog/how-to-screen-candidates?utm_source=blog&utm_medium=seo&utm_campaign=ai-technical-interviews&utm_content=internal-link) shows the test-then-interview set-up, and [AI interview tools for staffing agencies](/blog/ai-interview-tools-for-staffing-agencies?utm_source=blog&utm_medium=seo&utm_campaign=ai-technical-interviews&utm_content=internal-link) covers agencies that submit engineers to many clients. For the wider field, see the [best AI interview software for hiring teams](/blog/best-ai-interview-software?utm_source=blog&utm_medium=seo&utm_campaign=ai-technical-interviews&utm_content=internal-link). ## How do you set up an AI technical interview tool? 1. **List what the panel wants to know**, in three to five fields: problem solving, trade-offs, scaling, testing, communication. 2. **Pick the task types**: coding for most roles, system design for senior roles, debugging for large codebases. 3. **Tell Tracy who you are interviewing and what you need to learn.** Your questions are drafted for you, and you add the questions that matter. 4. **Decide whether AI help is allowed** for the task and tell candidates at the start. 5. **Open the link yourself first** and answer as a candidate would, then change anything that feels off. 6. **Send personal invites or a public link.** Candidates join in the browser with no sign-up. 7. **Read reports field by field** and invite the strongest to the panel. 8. **Tell candidates an AI runs the first round** and that a person makes every decision. ## Common questions ### What is an AI technical interview? A first-round interview where an AI interviewer sets a coding, system design or debugging task and asks follow-up questions about the candidate's answer. Your team reviews the report and decides who goes to the panel. ### Can an AI coding interview run live? Yes. Tracy can introduce a coding task during the interview and then probe the trade-offs in the candidate's own solution. ### Is an AI technical interview better than a coding test? They answer different questions. A coding test checks whether the solution works on a standard scale. An AI technical interview shows how the candidate reasons about it. Running a short test first, then an interview, gets you both. ### Do CodeSignal and HackerRank have AI interviewers? Yes. CodeSignal lists AI Interviewers by role family in its plans, and HackerRank includes its Chakra AI Interviewer from the Starter plan. Check their pages for current details. ### Is Tracy SOC 2 compliant? Yes. Tracy is SOC 2 Type 2 compliant, with the report on request at tracy@qode.world. --- **[Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=ai-technical-interviews&utm_content=try-free)** or **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=ai-technical-interviews&utm_content=demo)** ## Sources - CodeSignal, Technical Assessments. [codesignal.com/technical-assessments](https://codesignal.com/technical-assessments). Retrieved 29 Oct 2026. Quote: "assessments written by subject matter experts and validated by IO psychologists, each backed by 2,800 hours of research and testing." - CodeSignal, AI Interviewer. [codesignal.com/ai-interviewer](https://codesignal.com/ai-interviewer). Retrieved 29 Oct 2026. Quote: "If answers are thin, it probes and asks follow up questions based on your criteria." Also: "produces an in-depth skills report + full transcript with rationale." - CodeSignal, Pricing and Plans. [codesignal.com/pricing](https://codesignal.com/pricing/). Retrieved 29 Oct 2026. Quote: "Build $79 /mo Billed annually ... 60 Annual credits Use for Assessments & Interviews ... Product, Design, & Engineering AI Interviewers Access to 2,000 Questions." Also: "Sales, CS, & Marketing AI Interviewers Basic ATS Integrations" (Grow) and "CodeSignal charges on-demand credits at $20 each." - HackerRank, Screen. [hackerrank.com/products/screen](https://www.hackerrank.com/products/screen/). Retrieved 29 Oct 2026. Quote: "HackerRank has the largest library of developer hiring content in the world." - HackerRank, Chakra. [hackerrank.com/chakra](https://www.hackerrank.com/chakra). Retrieved 29 Oct 2026. Quote: "Then it asks follow-up questions on the candidate's work to see how well they're able to defend their decisions and approach." Also: "a clear 'why' for every score." - HackerRank, Pricing. [hackerrank.com/work/pricing](https://www.hackerrank.com/work/pricing) (loaded in a browser; plain fetches return a cookie notice). Retrieved 29 Oct 2026. Quote: "Starter For teams making their first technical hires. $79 /mo $948 billed annually ... 60 candidate attempts / year Chakra - AI Interviewer ... 2000+ library questions." Also: "Pro ... 360 candidate attempts / year ATS and calendar integrations" and "Additional attempts $20 each". - Tracy AI Interviewer, home page. [tracyinterview.ai](https://tracyinterview.ai). Retrieved 29 Oct 2026. Quote: "Tracy AI Interviewer can ask role-specific questions, introduce a coding task, and probe the trade-offs behind an answer. CODING TASK SYSTEM DESIGN DEBUGGING CODE WALK-THROUGH Example brief Ask a backend engineer to design a rate limiter, then follow up on concurrency, failure modes, and testing." Also: "Tracy AI Interviewer runs natural two-way video interviews, follows up on vague answers, digs into the interesting ones"; "Your questions are drafted for you." and "add the questions that matter"; "People join in their browser whenever it suits them."; "Each person opens a link in their browser and joins with no sign-up."; "Each interview becomes a readable summary with the participant's own words attached"; "Direct quotes for context"; "your own repeatable interview, with every answer organized around the fields you define." - Tracy AI Interviewer, pricing. [tracyinterview.ai/pricing](https://tracyinterview.ai/pricing). Retrieved 29 Oct 2026. Quote: "$1 per credit, in USD. Credits never expire." Also: "You pay for results, not seats." and "1 credit for each started 20 minutes". - Qode, Tracy product page. [qode.world/tracy](https://qode.world/tracy). Retrieved 29 Oct 2026. Quote: "4.7/5 average candidate rating across interviews." - Qode, pricing. [qode.world/pricing](https://qode.world/pricing). Retrieved 29 Oct 2026. Quote: "Interview integrity detection: Tab changes, Abnormal eye tracking, Live face detection, Average delay time before answer, External screens, Other unusual patterns." Also: "Video & chapter breakdown", "Transcript", "Answer summary & score", "Explain answer score by AI Agent" and, for Enterprise, "Custom based on volume". - Qode, 28 Oct 2026: enterprise pricing is custom, based on volume. - Qode, confirmed product facts (integrations: ChatGPT, Claude, Gemini, Mistral, Grok, Slack, WhatsApp, Zapier, Workable, HubSpot; SOC 2 Type 2, report on request), 29 Oct 2026. # How to Screen Candidates When There Are Too Many Applicants > How to screen candidates when AI-written CVs look alike: interview every applicant, ask follow-ups and shortlist from answers you can check. - URL: https://tracyinterview.ai/blog/how-to-screen-candidates - Published: 2026-11-02 - Updated: 2026-11-11 - Author: Tracy team > **In short** > - **Interview before you filter.** Tracy AI Interviewer gives every applicant a two-way first interview, follows up on their answers and returns a summary tied to what they actually said. > - This moves the shortlist from AI-polished CVs toward evidence of how people did the work. Recruiters and hiring managers check the answers, video and transcript, then make the decision. **[Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=how-to-screen-candidates&utm_content=try-free)** or **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=how-to-screen-candidates&utm_content=demo)** **On this page** - [Why are there too many applicants for every job now?](#why-are-there-too-many-applicants-for-every-job-now) - [Why doesn't screening applicants for a job faster fix it?](#why-doesnt-screening-applicants-for-a-job-faster-fix-it) - [How to screen candidates so you see who they really are](#how-to-screen-candidates-so-you-see-who-they-really-are) - [How do you screen candidates effectively when every CV looks good?](#how-do-you-screen-candidates-effectively-when-every-cv-looks-good) - [Won't candidates drop out of an AI interview?](#wont-candidates-drop-out-of-an-ai-interview) - [How to screen candidates for one role, step by step](#how-to-screen-candidates-for-one-role-step-by-step) - [Questions](#questions) - [Sources](#sources) ## Why are there too many applicants for every job now? If you're working out how to screen candidates for a skilled role today, the pile is bigger and harder to read than it was. A candidate can tailor a CV and cover letter to your job description in minutes with an AI tool, then do the same for dozens of other jobs. Robert Half's March 2026 release, based on a survey of more than 2,000 US hiring managers in November 2025, reports: - 67% of US HR leaders say reviewing AI-generated applications has slowed hiring, and 20% report delays of more than two weeks. - 84% of HR teams report heavier workloads as AI-tailored applications increase. - 65% of hiring managers say AI-enhanced resumes make skills harder to verify. The last number matters most. A big pile would be manageable if the CVs told you who could do the job. When every CV reads well, you can't sort the pile however hard you cut it. ## Why doesn't screening applicants for a job faster fix it? The usual answer is a faster filter: stricter knock-out questions, a keyword score, a one-way video, a quicker skim. Each one shortens the list without telling you more about the person. Among the steps employers took to validate candidates, the same release lists spending more time reviewing applications (42%) and increasing the number of interviews per candidate (38%). They added conversations because a candidate in a conversation has to explain their own work. They can't afford that conversation with everyone. ![Two funnels side by side. Filter first: all applicants, a CV filter cuts the pile, a few are interviewed. Interview everyone: all applicants, everyone interviewed, a shortlist from their answers.](/blog/images/how-to-screen-candidates-funnel.svg) *Filter first and you meet whoever wrote the best CV. Interview everyone first and you meet whoever gave the best answers.* ## How to screen candidates so you see who they really are Keep the conversation and change who runs the first one. Each applicant opens Tracy's link in the browser and joins with no sign-up, whenever it suits them, so you can give the whole pile the conversation you used to save for the top ten. Three things make that first round worth reading. **1. Interview everyone, before you cut.** If you filter first, the filter still decides who you meet, and the filter reads CVs. Send the interview link with the application confirmation, so the candidate with a plain CV and real experience gets the same look as the one whose CV was polished by AI. **2. Ask the follow-up on their own answer.** A fixed question gets a prepared answer; the second and third questions can't be rehearsed because they're built on what the candidate just said. Tracy asks role-specific questions and follows up on vague answers, "a real follow-up, not a rigid script", so a thin answer shows up in the first round instead of the final one. For engineers it can introduce a coding task and probe the trade-offs behind an answer, as [AI technical interviews](/blog/ai-technical-interviews?utm_source=blog&utm_medium=seo&utm_campaign=how-to-screen-candidates&utm_content=internal-link) shows. For specialist roles, see our [interview questions for AI trainers and domain experts](/blog/ai-trainer-interview-questions?utm_source=blog&utm_medium=seo&utm_campaign=how-to-screen-candidates&utm_content=internal-link). ![Tracy's interviews list for a hiring team. Each row is a candidate with a status, the answers given, the date their link runs out and a View result link.](/blog/images/lib-interviews-list-hiring-en.png) *When everyone is interviewed, the pile becomes one list you can open row by row (illustration with sample data).* **3. Read a summary you can check.** Each interview becomes a readable summary "with the participant's own words attached", with every answer organised around the fields you define and the video and transcript beside it. Direct quotes give context and the AI explains each answer score, so a hiring manager can confirm a strong answer, or catch a thin one, without rewatching the whole interview. Your team makes the decision. Tracy's interview integrity detection flags tab changes, abnormal eye tracking, live face detection, average delay before an answer, external screens and other unusual patterns, so a candidate getting outside help doesn't reach your shortlist unseen. See [how to detect AI cheating in interviews](/blog/ai-interview-cheating-detection?utm_source=blog&utm_medium=seo&utm_campaign=how-to-screen-candidates&utm_content=internal-link). ## How do you screen candidates effectively when every CV looks good? Two applicants for an analytics engineer role can both write "owned the revenue dashboard". The follow-up chain separates them. Example: a first-round chain Tracy could run. Illustration, not a real candidate. > **Tracy:** Tell me about a metric you built that other teams relied on. > > **Candidate:** Weekly active accounts. Sales and finance both used it. > > **Tracy:** Sales and finance often count "active" differently. Did they agree on your definition? > > **Candidate:** Not at first. Finance wanted paying accounts only, so I split it into two metrics and documented which dashboard used which. > > **Tracy:** What happened to the old numbers when you split it? > > **Candidate:** The historic chart dropped by about 30% overnight. I backfilled both series and warned the two leads before they saw it. The first answer is what every CV says. The second shows the candidate knew a metric definition is a decision between teams, and the third shows they thought about the people reading the numbers. Example: part of the Tracy summary for that candidate. Illustration with test data. | Field | What the candidate said | | --- | --- | | Metric design | "I split it into two metrics and documented which dashboard used which." | | Handling change | "I backfilled both series and warned the two leads before they saw it." | Because the summary quotes the answer, two reviewers can argue about what the candidate said instead of about a number. ## Won't candidates drop out of an AI interview? Some will, and the reasons are known. Greenhouse's 2026 survey covered 2,950 job seekers in five countries. Among US respondents, 63% had already had an AI interview and 38% had withdrawn from a hiring process because it included one. The top reasons US candidates gave were pre-recorded video interviews scored by AI with no human present (33%) and companies failing to disclose how AI would be used (27%). Another 38% want confirmation that a person reviews AI output before decisions are made. Each of those is a setup choice. Run a two-way interview that responds to what the candidate says. Tell candidates up front that an AI runs the first round, what it looks at, including the integrity signals, and that a person reads every summary and decides. Tracy reports a 72% candidate completion rate and a 4.7 out of 5 average candidate rating. Measure completion in your own hiring process, too. ## How to screen candidates for one role, step by step Pick the role with the most applicants and the shortlist you trust least. 1. Write down what your best interviewer for that role wants to know after a first call. Four to six fields is enough. 2. Turn the fields into questions, or tell Tracy who you want to speak with and what you need to learn, and the questions are drafted for you. 3. Take the interview yourself, the way a candidate would, and change anything that feels off. 4. Add the link to your application confirmation, so every applicant gets it. 5. Read the summaries field by field and invite the strongest candidates to your next round. 6. Run it next to your current screen for two weeks and compare who each method would have shortlisted. Tracy works inside tools your team already uses, such as Slack, Workable and HubSpot, so the trial needs no new system. Pricing is per result: $1 per credit; one credit covers up to 20 minutes, which fits a first round of five or six questions. A session with no answers is never charged, so a two-week trial costs only the interviews candidates finish. At high volume, enterprise pricing is custom, based on your volume. ## Common questions ### What should I do with too many applicants for a job? Give every applicant the same short first interview, then read the summaries instead of the CVs. Applicants take Tracy's interview whenever it suits them, so the bottleneck moves from scheduling calls to reading answers. ### Is it fair to interview every applicant with AI? Everyone gets the same questions and the same chance to explain their work, which is fairer to the candidates a CV filter would have cut unseen. Tell candidates an AI runs the first round and that a person makes every decision. This isn't legal advice; check the rules where you hire. ### Can candidates use AI to answer an AI interview? Some try. Follow-ups on their own answers make it harder, because a prepared script doesn't cover the third question, and Tracy flags tab changes, abnormal eye tracking, live face detection, average delay before an answer, external screens and other unusual patterns to your reviewer. ### We already use a one-way video tool. Why change? One-way video records a prepared answer and asks no follow-up, and Greenhouse found pre-recorded video scored by AI with no human present was the top reason US candidates walked away (33%). A two-way interview asks about what the candidate just said, and a person reviews the result. ### Does this work for staffing agencies? Yes, though agencies care most about speed to submit. Our guide to AI interview tools for staffing agencies compares the options. --- **[Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=how-to-screen-candidates&utm_content=try-free)** or **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=how-to-screen-candidates&utm_content=demo)** ## Sources All retrieved 29 Oct 2026. 1. Robert Half, "Robert Half survey: 67% of HR leaders report AI-generated applications are slowing hiring", press release, 10 Mar 2026: [press.roberthalf.com](https://press.roberthalf.com/2026-03-10-Robert-Half-survey-67-of-HR-leaders-report-AI-generated-applications-are-slowing-hiring). Quotes: "67% of U.S. HR leaders say reviewing AI-generated applications has slowed the hiring process" / "20% reporting delays of more than 2 weeks" / "Increasing the number of interviews per candidate (38%)" / "Spending more time reviewing applications (42%)" / "84% of HR teams report heavier workloads as AI-tailored applications increase." / "65% of hiring managers say AI-enhanced resumes make skills harder to verify." / "The survey contains responses from more than 2,000 hiring managers in the United States." / "conducted by an independent research firm in November 2025." 2. Greenhouse, 2026 candidate AI interview report, 29 Apr 2026: [greenhouse.com](https://www.greenhouse.com/blog/2026-candidate-ai-interview-report). Quotes: "nearly two-thirds of US job seekers (63%) say they've already experienced one" / "38% of US candidates have already withdrawn from a hiring process because it included an AI interview" / "based on a survey of 2,950 job seekers across the US, UK, Ireland, Germany and Australia" / "pre-recorded video interviews scored by AI with no human present (33%)" / "companies failing to disclose how AI would be used (27%)" / "38% of US candidates want confirmation that a person reviews AI output before decisions are made" 3. Tracy AI Interviewer, [pricing](https://tracyinterview.ai/pricing): "$1 per credit, in USD. Credits never expire." / "1 credit for each started 20 minutes, charged when the result is ready" / "A session with no answers is never charged." 4. Tracy AI Interviewer, [tracyinterview.ai](https://tracyinterview.ai): "Tracy AI Interviewer runs natural two-way video interviews, follows up on vague answers, digs into the interesting ones" / "A real follow-up, not a rigid script." / "Each person opens a link in their browser and joins with no sign-up." / "People join in their browser whenever it suits them." / "Tracy AI Interviewer can ask role-specific questions, introduce a coding task, and probe the trade-offs behind an answer" / "Each interview becomes a readable summary with the participant's own words attached" / "Direct quotes for context" / "your own repeatable interview, with every answer organized around the fields you define." / "Your questions are drafted for you." 5. Qode, [qode.world/tracy](https://qode.world/tracy): "4.7/5 average candidate rating across interviews" / "72% candidate completion rate for Tracy interviews" 6. Qode, [qode.world/pricing](https://qode.world/pricing): "Interview integrity detection: Tab changes, Abnormal eye tracking, Live face detection, Average delay time before answer, External screens, Other unusual patterns" / "Video & chapter breakdown", "Transcript", "Answer summary & score", "Explain answer score by AI Agent" / Enterprise: "Custom based on volume" 7. Qode, 28 Oct 2026: enterprise pricing is custom, based on volume. 8. Qode, 29 Oct 2026: Tracy works inside Slack, Workable and HubSpot, among other named tools. # Candidate Comparison Template: Compare Finalists Fairly > A free candidate comparison template: score finalists on the same criteria and evidence, apply a close-call rule, and see a filled three-finalist example. - URL: https://tracyinterview.ai/blog/candidate-comparison-template - Published: 2026-11-12 - Author: Tracy team > **In short** Tracy AI Interviewer interviews each applicant on the skills you set, so finalists are compared on the same criteria and the same kind of evidence: a recording, a timestamped transcript and a written assessment per question. > > Below: a candidate comparison template with a close-call rule, and a filled three-finalist example. > > A person reviews the evidence and chooses. > > Next step: copy the template, then run one role as a Tracy interview. Three finalists, three interviewers, three sets of notes. One finalist was interviewed last and is the one everyone remembers. One answered different questions on a Monday morning. The third has two pages of notes because the interviewer likes writing. A candidate comparison template makes the finalists comparable: the same criteria, the same scale, one line of evidence for each score and a rule for close calls, agreed before anyone is rated. Copy it below, see it filled in for three fictional finalists, then see how [Tracy AI Interviewer](/product) interviews every finalist on the same skills. ## Why finalists get compared unfairly Finalists get compared unfairly for four reasons. They answered different questions, so the scores measure the questions. They have different evidence: one has a recording, another a paragraph written from memory. The last interview is the freshest, so the last finalist gains. And in an open discussion the first strong opinion anchors everyone else. The template answers each one: shared criteria, evidence beside every score, one criterion compared at a time, and scores written before the discussion. ## The candidate comparison template Copy the tables into a sheet. Put the weights in column B and the finalists' scores in columns C to E, rows 2 to 6. For the finalist in column C the weighted score is `=SUMPRODUCT($B$2:$B$6,C2:C6)/SUM($B$2:$B$6)`. Fill it right for the others. **Table 1: setup** | Field | Entry | | --- | --- | | Role | | | Decision owner | | | Decision date | | | Finalists | A, B, C | | Must-haves (pass or fail, checked before scoring) | | **Table 2: weighted scores sheet** | Criterion | Weight (adds to 100) | Finalist A | Finalist B | Finalist C | | --- | --- | --- | --- | --- | | Criterion 1 | 30 | | | | | Criterion 2 | 20 | | | | | Criterion 3 | 20 | | | | | Criterion 4 | 20 | | | | | Criterion 5 | 10 | | | | | Weighted score | | | | | **Table 3: evidence per criterion** | Criterion | Finalist A | Finalist B | Finalist C | | --- | --- | --- | --- | | Criterion 1 | Quote or moment | | | | Criterion 2 | | | | | Criterion 3 | | | | | Criterion 4 | | | | | Criterion 5 | | | | Scale: 1 is no evidence, 3 meets what the role needs, 5 is exceptional. Write what 2 and 4 mean for your role before the first interview. **Table 4: decision record** | Field | Entry | | --- | --- | | Weighted scores | | | Gap between first and second | | | Close-call rule | If the top two are within 0.3, choose on the heaviest criterion; if still level, on the smaller risk. | | Biggest risk for each finalist | | | Reference checks | Done, pending or not needed | | Decision | | | Why, in two sentences | | | Plan for the risk you accept | | Use four or five criteria. Set the weights, the must-haves and the close-call rule before the first finalist is interviewed. ## How to compare finalists fairly - **Treat must-haves as pass or fail.** A finalist who fails one is not rescued by a high score elsewhere. - **Compare one criterion at a time.** Read all three finalists on criterion 1, then criterion 2. The evidence for the same question sits side by side. - **Score alone first.** Write scores and evidence before the group meets. - **Compare like with like.** Same round, same criteria, same kind of evidence. If one finalist had a different interview, say so on the sheet. - **Use the rule, and overrule it only in writing.** If the group chooses against the numbers, record why. - **Write the decision in two sentences** and name the risk you accept. ![Sample scoresheet entry a team filled in for a fictional finalist: five criteria scored 3 to 5 out of 5, weighted 4.0.](/blog/images/candidate-comparison-template-1.png) *The team scores each finalist on the same criteria and weights, from the same kind of evidence. Illustration with sample data on a sample 1 to 5 scale. Illustration with sample data* Each finalist gets the same criteria, a score and a line of evidence per criterion. ## A filled example: three finalists for a support team lead Fictional finalists. The evidence is written for this page, and the scale is the one in the template. **Table 5: filled example, scores (fictional finalists)** | Criterion | Weight | Finalist A | Finalist B | Finalist C | | --- | --- | --- | --- | --- | | Coaching | 30 | 4 | 5 | 3 | | Escalation handling | 20 | 5 | 3 | 4 | | Process improvement | 20 | 3 | 4 | 3 | | Communication | 20 | 4 | 4 | 5 | | Schedule fit | 10 | 5 | 3 | 4 | | Weighted score | 100 | 4.1 | 4.0 | 3.7 | **Table 6: filled example, evidence (fictional finalists)** | Criterion | Finalist A | Finalist B | Finalist C | | --- | --- | --- | --- | | Coaching | Weekly one to ones for two new hires; changed the plan after one missed targets. | 90-day plan for a struggling agent, with a measure at day 30. | Talked about morale, no plan described. | | Escalation handling | Walked through a P1 outage and who was told when. | Handed escalations to a manager without a rule for when. | Named a clear trigger, less detail on follow-through. | | Process improvement | One macro change, no result given. | Rewrote the refund checklist and cut repeat tickets. | General ideas, no example. | | Communication | Clear, a little long. | Clear and brief. | Checked understanding after each point. | | Schedule fit | Any shift. | Mornings only. | Can cover most evenings. | Finalist A leads by 0.1, inside the 0.3 close-call band, so the rule applies. The heaviest criterion is coaching, where B scored 5 against A's 4. Decision: B. B's risk is escalation handling at 3, so the plan is to pair B with an escalation lead for the first 60 days and to ask B's referee about escalations. The record says why the number alone did not decide it. ## Compare finalists on the same criteria A comparison is fair when every finalist is scored on the same criteria and the evidence has the same shape. [Tracy AI Interviewer](/product) interviews each applicant for the role on the interview you set up, so the criteria match. - **Same skills.** You pick the skills for each role, and every applicant is interviewed on them. - **Same evidence.** A recording of both sides, a timestamped transcript and a written assessment per question with its reasoning, for every finalist and not only those whose interviewer took good notes. - **Checkable.** Every point in an assessment links to the moment in the recording, so a reviewer can check it without watching the whole interview. - **Shareable.** The summary, the recording, the transcript and an assessment per question sit in one record, shareable by link or PDF. A Director at a top-tier early-stage VC: "For borderline candidates who are difficult to evaluate based on quantitative scores alone, Tracy makes it much easier to share their responses and assessment results with other team members for further discussion." - **A person decides.** A recruiter can move a candidate forward or invite someone to interview again, and the full history stays on the record. See [how an AI interview report reads](/blog/ai-interview-reports). The first 3 results are free with no card. After that it costs $1 per credit, one credit per started 20 minutes of an interview that produced a result, with no seats and no upfront cost. See [pricing](/pricing). To try it, run one role with three real interviews and put the results in the template. ![Four steps: set the skills once, interview every applicant, read the evidence, a person chooses.](/blog/images/candidate-comparison-template-2.png) *Tracy interviews every applicant on one set of skills and leaves the choice to a person. Illustration with sample data* The same criteria and the same kind of evidence make the finalists comparable. ## Which path fits you - Many applicants for one role, and finalists to pick: [volume hiring](/use-cases/volume-hiring). - Engineering finalists: [technical interviews](/use-cases/technical-interviews). - Choosing how many to interview before you compare: [AI screening interviews](/blog/ai-screening-interviews). ## Common questions ### How do you compare candidates fairly? Use the same criteria, scale and questions for every finalist, write the score and one line of evidence before the group meets, and compare one criterion at a time across all finalists. ### What should a candidate comparison template include? The role, the must-haves, four or five weighted criteria, a defined scale, one line of evidence for each score, each finalist's biggest risk and a decision with the reason in two sentences. ### How do you compare candidates in Excel? Put weights in one column and each finalist's scores in the next columns, then use SUMPRODUCT divided by the sum of the weights for each finalist. The template above gives the formula. ### What if two finalists score almost the same? Agree a close-call rule in advance. The template uses a 0.3 band: inside it, choose on the heaviest criterion, then on the smaller risk, and write down the reason. ### Should the numbers decide, or the group? The numbers organise the discussion and the group decides. If the group overrules the numbers, the reason belongs in the record. --- Copy the template, then run it as a Tracy interview. Send us one role and we will build the interview plan in days, with a recording and report you can check. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=candidate-comparison-template&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=candidate-comparison-template&utm_content=demo) ## Sources - [Tracy AI Interviewer home page, customer quotes](https://tracyinterview.ai/), retrieved 9 Nov 2026. - [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer Connect page](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. # Interview Feedback Form Template (Free) > A free interview feedback form template to copy: ratings with evidence, strengths and concerns, a recommendation and a filled example. - URL: https://tracyinterview.ai/blog/interview-feedback-form-template - Published: 2026-11-12 - Author: Tracy team > **In short** Tracy AI Interviewer runs the first-round interview for every applicant and gives you a written assessment per question, with a recording and timestamped transcript behind each result. > > Below: a copy-ready interview feedback form, a filled example and rules for feedback that holds up in a debrief. > > A person reads the evidence and decides. > > Next step: copy the form, then run one role as a Tracy interview. A hiring manager opens three pieces of feedback on one candidate. The first says "great communicator". The second says "not sure". The third says "thumbs up". None of them can be compared with the others. An interview feedback form template fixes that by asking every interviewer the same things in the same order: what they asked, what they heard, how it rates and what they recommend. Copy the form below, read the filled example, then see how [Tracy AI Interviewer](/product) runs the first round and reports on it. ## What an interview feedback form is for An interview feedback form is the note an interviewer files right after one interview. It records what was asked, what the candidate said and did, how that rates against the role, and what the interviewer recommends. A scorecard adds scores across criteria so you can rank candidates; the feedback form keeps the reasoning in words, which a debrief needs when two interviewers disagree. Three rules shape the template. Every interviewer answers the same fields. Every rating sits beside a quote or a moment. The recommendation comes last, after the evidence. ## The interview feedback form template Copy each table into a document, a form builder or the feedback field of your applicant tracking system. Each interviewer files one form per candidate. **Table 1: header fields** | Field | Entry | | --- | --- | | Candidate | | | Role | | | Stage | First round, second round or final | | Interviewer | | | Date and length | | | Recording or notes link | | **Table 2: questions covered** | Question asked | Answered fully? (yes, partly, no) | Notes | | --- | --- | --- | | Question 1 | | | | Question 2 | | | | Question 3 | | | | Anything asked outside the plan | | | **Table 3: ratings sheet** | Competency | Rating (1 to 5) | What I heard or saw (quote or moment) | | --- | --- | --- | | Competency 1 | | | | Competency 2 | | | | Competency 3 | | | | Competency 4 | | | Rate on this scale, and write what it means for your role before the first interview. **Table 4: the 1 to 5 scale** | Rating | Meaning | | --- | --- | | 1 | Evidence against the skill, or none. | | 2 | Thin once pushed for detail. | | 3 | Does what the role needs. | | 4 | Above that, with specifics. | | 5 | Exceptional and consistent under follow-up. | **Table 5: summary and recommendation** | Field | Entry | | --- | --- | | Strengths (up to three) | | | Concerns (up to three) | | | What I could not assess | | | Candidate's questions and interest | | | Recommendation | Strong yes, yes, no or strong no | | What would change my mind | | | Probe in the next round | | | Written before reading other feedback | Yes or no | Use three to five competencies, the ones the role needs. The recommendation has no middle option on purpose: "maybe" hands the decision to the debrief without adding information. ## How to write feedback that holds up in a debrief - **File it the same day.** Do it while the details that justify a rating are still fresh. - **Evidence before verdict.** Write the quote or the moment, then the rating. If you cannot point to something said or done, the rating is a guess. - **Rate against the scale, not the last candidate.** A strong answer after a weak one is still a 3 if it meets the bar. - **Say what you could not assess.** A blank is not a pass. The gap tells the next interviewer what to ask. - **Replace trait words with behaviour.** "Polished", "energy" and "culture fit" say nothing. "Explained the error in two sentences and checked I had followed" does. - **Leave out what is not about the job.** Age, family, accent and where someone went to school are not evidence. - **Write it before you read anyone else's.** Independent feedback lets the debrief compare evidence. ![Four steps: set your criteria, Tracy interviews, read the evidence, a person decides.](/blog/images/interview-feedback-form-template-1.png) *Tracy runs the first-round interview and gives the reviewer a written assessment per question, the recording and the transcript, set against your criteria. Illustration with sample data* Tracy gives the reviewer your criteria, an assessment per question, the recording and transcript, and a person decides. ## A filled example: operations analyst, first round A fictional first-round interview. The candidate and the evidence are written for this page. **Table 6: filled example, ratings (fictional)** | Competency | Rating | What I heard or saw | | --- | --- | --- | | Data analysis | 4 | Described removing duplicates from a monthly export before building a pivot table. Asked what could still be wrong, named date formats and checked them. | | Process thinking | 3 | Mapped a refund process step by step but did not say where it usually breaks. | | Communication | 5 | Explained a reconciliation error in two sentences without jargon and checked I had followed. | | Ownership | 3 | Said "the team missed it" twice about a late report. No example of fixing the cause. | **Table 7: filled example, summary (fictional)** | Field | Entry | | --- | --- | | Strengths | Clear explanations. Checks own work before sharing it. | | Concerns | Few examples of owning a miss. Process answers stop at the happy path. | | What I could not assess | Experience with the reporting tool named in the job description. | | Candidate's questions and interest | Asked about reporting cadence and team size. Keen. | | Recommendation | Yes | | What would change my mind | A second example of a missed deadline where the candidate changed the process. | | Probe in the next round | A process they designed that failed, and what they did next. | | Written before reading other feedback | Yes | Each 3 and 4 points at something said, and the concerns become the next round's questions. ## Run the first round as a Tracy interview The form works for interviews a person has time to run and write up. For the first round of a role with many applicants, [Tracy AI Interviewer](/product) runs the interview, and her report covers much of the same ground. - **Questions covered.** Tracy generates the questions from the job description. Edit them or use your own, from quick to deep, technical to behavioural. - **Ratings on your criteria.** You choose the skills to evaluate for each role or role family. Your competency rows become the skills. - **What I heard.** Each result has a written assessment per question with its reasoning, a recording of both sides and a timestamped transcript, so a reviewer can check a rating against what was said. - **Strengths and gaps.** The report holds strengths and gaps, the recording and the fit rationale in one record, shareable by link or PDF. - **Follow-ups.** She asks in the candidate's own field, follows up on what they said and waits for the answer. A Director at a top-tier early-stage VC says: "Tracy sometimes brings up exactly those questions I might have forgotten to ask, which has been a pleasant surprise." - **The recommendation stays with a person.** A recruiter decides: they can move a candidate forward or invite someone to interview again, and the full history stays on the record. See [how an AI interview report reads](/blog/ai-interview-reports). The first 3 results are free with no card. After that it costs $1 per credit, one credit per started 20 minutes of an interview that produced a result, with no seats and no upfront cost. See [pricing](/pricing). To try it, paste your competencies into one role and run three real interviews. ![Sample interview: Tracy asks how errors are prioritised, the candidate answers, Tracy asks about a past mistake.](/blog/images/interview-feedback-form-template-2.png) *Tracy follows up on what the candidate said, so each written assessment rests on more than a first answer. Illustration with sample data* A follow-up gives the evidence line something to quote. ## Which path fits you - Screening many applicants for the same role: [volume hiring](/use-cases/volume-hiring). - Engineering and other technical roles: [technical interviews](/use-cases/technical-interviews). - Deciding how to screen before you write the form: [how to screen candidates](/blog/how-to-screen-candidates). ## Common questions ### What should an interview feedback form include? The candidate, role, stage and interviewer; the questions asked; a rating for each competency with a quote or moment beside it; strengths, concerns and what you could not assess; and a recommendation with what would change your mind. ### How long should interview feedback be? One page. A line or two of evidence per competency, up to three strengths, up to three concerns and a recommendation. Anything longer is hard to compare. ### What is the difference between an interview feedback form and a scorecard? A feedback form records the reasoning in words after one interview. A scorecard weights scores across criteria so you can rank candidates. Use the form for the why and the scorecard for the ranking. ### When should interviewers submit feedback? The same day, before they read anyone else's feedback. Independent notes are what let a debrief compare evidence. ### What does good interview feedback look like? It names what the candidate said or did, rates it against the scale and says what could not be assessed. The filled example above rates data analysis a 4 because the candidate named and checked date formats, not because they seemed sharp. --- Copy the form, then run it as a Tracy interview. Send us one role and we will build the interview plan in days, with a recording and report you can check. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=interview-feedback-form-template&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=interview-feedback-form-template&utm_content=demo) ## Sources - [Tracy AI Interviewer home page, customer quotes](https://tracyinterview.ai/), retrieved 9 Nov 2026. - [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer Connect page](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. # Interview Guide Template for Hiring Managers > A free interview guide template for hiring managers: timed run order, core questions with follow-ups, a decision block and a filled example. - URL: https://tracyinterview.ai/blog/interview-guide-template - Published: 2026-11-12 - Author: Tracy team > **In short** Tracy AI Interviewer asks your core questions of each candidate: she follows up on what each person says and leaves a recording and a written assessment for each question. > > Below: a copyable guide with a timed run order, core questions with follow-ups and a decision block, plus a filled example. > > Structured interviews outperform unstructured ones in the research; the guide below is the asking half. > > The first 3 results are free. Paste the guide into one role and run it. You have the job description, a free hour and an interview tomorrow. An interview guide template turns that into a plan: what to cover, in what order, which questions every candidate gets and what a good answer sounds like. Without one, the conversation follows whoever talks most, and the next candidate gets a different interview. Below is an interview guide template for hiring managers, a filled example, and how [Tracy AI Interviewer](/product) asks your core questions of each candidate, built from the job description. ## Why use a guide A guide is the asking half of a structured interview. A 2022 re-analysis in the Journal of Applied Psychology by Sackett and colleagues, covering earlier published meta-analyses, put the mean validity of structured employment interviews at .42 and of unstructured interviews at .19. That is a correlation between the interview rating and later job performance ratings: 105 studies and 7,864 people for structured interviews, 32 studies and 3,125 people for unstructured ones. It is a stronger link, not a guarantee. The guide fixes the questions and their order; a scorecard fixes how you rate the answers. For screening before the interview, see [how to screen candidates](/blog/how-to-screen-candidates). ## The interview guide template Copy each table into a document or sheet. Fill the header first, then write the questions, then fix the minutes. A 45-minute guide works for most first and second rounds. **Table 1: header** | Field | Entry | | --- | --- | | Role and level | | | Hiring manager | | | Interviewer | | | Length | 45 minutes | | Three must-haves | 1. 2. 3. | | What would end the process | | **Table 2: timed run order, 45 minutes** | Block | Minutes | What you do | | --- | --- | --- | | Opening | 3 | Say who you are, how long this takes, that you will take notes, and what happens after. | | Context | 7 | Ask about their current role and why this one. Listen for fit, not polish. | | Core questions | 25 | Ask the same four or five questions of every candidate. Use the follow-ups. | | Their questions | 7 | Let them ask. Note what they ask. | | Close | 3 | Say when and how they will hear back, and keep to it. | **Table 3: core questions** | # | Skill | Ask | Follow up with | Listen for | Score (1 to 5) | | --- | --- | --- | --- | --- | --- | | 1 | | | | | | | 2 | | | | | | | 3 | | | | | | | 4 | | | | | | | 5 | | | | | | **Table 4: notes and decision** | Field | Entry | | --- | --- | | Best moment (a quote) | | | Biggest concern | | | Recommendation | Advance, hold or do not advance | | Send to the next interviewer | The one thing to probe next | Write each question from a must-have, so that every row of table 3 traces back to something in table 1. If a question does not test a must-have, cut it. ## How to use the guide - **Start from the job description.** Pull three must-haves from it, and write the questions from those. - **Ask the same core questions of every candidate.** Add or skip the context questions as you like. - **Keep the follow-up for after the answer.** Ask it only once you have heard the first one. - **Write the quote, not the verdict.** "Strong communicator" tells the next interviewer nothing; the sentence the candidate said does. - **Hold the minutes.** If the core block runs long, drop a question for everyone, not for the candidate you like. - **Close the loop.** Say when they will hear back, and do it. ## A filled example: support team lead A fictional guide for a support team lead role: the header, then the core questions. **Table 5: filled example, header (fictional role)** | Field | Entry | | --- | --- | | Role and level | Support team lead, 8 agents | | Must-haves | Coaches agents. Handles an angry escalation. Reads queue data. | | What would end the process | Cannot give one example of coaching someone else | **Table 6: filled example, core questions (fictional role)** | # | Skill | Ask | Follow up with | Listen for | | --- | --- | --- | --- | --- | | 1 | Coaching | Tell me about an agent whose work you helped turn around. | What did you say in the first conversation? | A specific change in the agent's behaviour, not a general approach | | 2 | Escalations | Describe the angriest customer you handled and how it ended. | What would you do differently now? | Calm steps, an owner for the fix, a follow-up call | | 3 | Queue data | Which numbers do you look at first on a Monday? | Which one did you act on last month? | A decision made from a number, not a list of dashboards | | 4 | Planning | A teammate is out sick and the queue doubles. What happens in the next hour? | Who do you tell, and what do you drop? | A clear order of actions and who hears about it | After the interview the manager adds the score for each row, a best-moment quote and a recommendation. Fictional content, written for this page. ## Run the guide with Tracy A guide still needs someone to ask it of every applicant. Tracy can take the asking: - **Set up from the job description.** Tracy generates the questions from the role. Edit them or paste in your own core questions, from a quick screen to a deep one, technical to behavioural. - **Follow-ups on what was said.** She asks in the candidate's own field, follows up on the answer and waits for it. A Director at a top-tier early-stage VC says: "Tracy sometimes brings up exactly those questions I might have forgotten to ask, which has been a pleasant surprise." - **Your criteria.** You pick the skills for each role, so the rows of your guide become the skills each assessment covers. - **Evidence per result.** A recording of both sides, a timestamped transcript and a written assessment per question with its reasoning, so a reviewer can check each result. - **Any hour.** Each candidate starts on a click, with no scheduling and no one from your team present. A person reviews the evidence and decides. A recruiter can move a candidate forward or invite someone to interview again, and the history stays on the record. See [how an AI interview report reads](/blog/ai-interview-reports) and the [volume hiring](/use-cases/volume-hiring) page. To try it, paste the core questions from your guide into one role and run it. ![Four steps: copy the guide, set up Tracy from the job description, candidates join by link, read the result.](/blog/images/interview-guide-template-1.png) *Tracy asks your core questions and leaves a recording, a transcript and a written assessment per question for your team. Illustration with sample data* Your questions stay yours; Tracy asks them and leaves the evidence. ![Transcript card: Tracy asks about a changed plan, the candidate mentions a lost supplier, Tracy asks who was told first.](/blog/images/interview-guide-template-2.png) *Tracy asks the guide's question, then a follow-up on what the candidate said, so the next interviewer reads the detail. Illustration with sample data* The follow-up asks for the detail behind the answer, so the next reviewer reads it. ## Which path fits you - One role, many applicants: [volume hiring](/use-cases/volume-hiring). - Technical roles with a coding task in the interview: [technical interviews](/use-cases/technical-interviews). - Cost before you start: [pricing](/pricing). ## Common questions ### What is an interview guide? A written plan for one interview: the order of the conversation, the questions every candidate gets, the follow-ups and what a good answer sounds like. It is also called an interview plan. ### What should an interview guide template include? A header with the role and the three must-haves, a timed run order, four or five core questions with follow-ups and things to listen for, and a notes and decision block. ### How long should an interview guide be? One page is enough. A 45-minute guide holds four or five core questions; more than that and the later ones get rushed. ### What is the difference between an interview guide and a scorecard? The guide says what to ask and in what order. The scorecard says how to rate the answers. Use both for the same role. ### Can Tracy use my interview guide? Yes. Paste in your own questions or let Tracy generate them from the job description, then edit them. Tracy asks the questions you set and follows up on what each candidate says. --- Copy the template, then run it as a Tracy interview. Send us one role and we will build the interview plan in days, with a recording and report you can check. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=interview-guide-template&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=interview-guide-template&utm_content=demo) ## Sources - [Sackett, Zhang, Berry and Lievens (2022), Revisiting meta-analytic estimates of validity in personnel selection, Journal of Applied Psychology 107(11)](https://doi.org/10.1037/apl0000994), retrieved 9 Nov 2026. - [Tracy AI Interviewer home page, customer quotes](https://tracyinterview.ai/), retrieved 9 Nov 2026. - [Tracy AI Interviewer Connect page](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. # Interview Rating Scale: 1 to 5 With Anchors > An interview rating scale from 1 to 5 with written anchors, a scoring sheet to copy, the rating errors to catch and a filled example. - URL: https://tracyinterview.ai/blog/interview-rating-scale - Published: 2026-11-12 - Author: Tracy team > **In short** Tracy AI Interviewer interviews each applicant on the skills you set and writes an assessment per question, with the recording behind it, which a reviewer can read against your anchors. > > Below: a 1 to 5 interview rating scale with written anchors, a scoring sheet to copy, the rating errors to catch and a filled example. > > Define what a 3 means before the first interview. > > Next step: copy the scale, then run one role as a Tracy interview. Two interviewers hear the same answer. One writes a 4, the other a 3, and neither is wrong, because nobody told them what a 4 is. An interview rating scale fixes that by defining each point in terms of what the candidate said or did. This page gives you a 1 to 5 interview rating scale with anchors, an interview scoring sheet to copy and a filled example. Then it shows how [Tracy AI Interviewer](/product) interviews each applicant on your skills and leaves the recording and an assessment per question for your reviewers. ## The interview rating scale: 1 to 5 with anchors An anchor describes what you would see or hear at that score, in behaviour. "Good communicator" is not an anchor. "Explained the trade-off in two sentences and checked I had followed" is. These generic anchors work for any criterion. A 2 sits between the 1 and 3 descriptions, and a 4 between the 3 and 5. **Table 1: generic anchors, 1 to 5** | Score | Label | Anchor | In the answer | | --- | --- | --- | --- | | 1 | Absent | The skill is missing, or the answer shows the opposite. | No example, or an example that went wrong and was not recognised. | | 2 | Thin | Some evidence, thin under follow-up. | An example with no detail on the candidate's own part. | | 3 | Role-ready | A real example and sensible answers to follow-ups. | What they did, why, and the result. | | 4 | Strong | Specific detail, a result and what they would change. | The trade-off named, with an outcome. | | 5 | Outstanding | Deep and consistent under follow-up. | Handles a harder version of the question unprompted. | A 3 means the person can do the job at the level you hire for. ## Anchors for a real criterion Generic anchors need a version for each criterion. Write the 1, 3 and 5, and let interviewers place a 2 or a 4 between them. Three examples: **Table 2: anchors for three criteria** | Criterion | 1 | 3 | 5 | | --- | --- | --- | --- | | Prioritising work | Does tasks in the order they arrive. | Ranks by deadline and names one trade-off. | Ranks by impact and deadline, says what was dropped and who was told. | | Explaining a decision | States the outcome with no reason. | Gives the reason. | Gives the reason, the option rejected and what would change the call. | | Learning from a mistake | Cannot name one, or blames others. | Names the error and what they fixed. | Names the error, the change to how they work and shows the change held. | ## The interview scoring sheet Copy the tables into a sheet. With the scores in column B, rows 2 to 6, the average for equal weights is `=AVERAGE(B2:B6)`. **Table 3: scoring sheet header** | Field | Entry | | --- | --- | | Candidate | | | Role | | | Interviewer | | | Date | | | Recording or notes link | | **Table 4: scoring sheet** | Criterion | Score (1 to 5) | Evidence: quote or time | Why not one higher | | --- | --- | --- | --- | | Criterion 1 | | | | | Criterion 2 | | | | | Criterion 3 | | | | | Criterion 4 | | | | | Criterion 5 | | | | | Average | | | | The last column does the work. To justify a 3, the interviewer must say what the 4 anchor needed and the candidate did not give. ## Choose a 1 to 3, 1 to 4 or 1 to 5 scale **Table 5: choosing a scale length** | Scale | Use it when | Watch for | | --- | --- | --- | | 1 to 3 | A fast screen of many applicants where the decision is pass, maybe or stop. | Too coarse to separate strong candidates. | | 1 to 4 | Interviewers who all drift to the middle. With no midpoint, they have to lean. | A 2 and a 3 can feel like fail and pass, so define both. | | 1 to 5 | Most roles. A clear bar at 3 and two steps either side. | Without anchors, 3 becomes the safe answer for everyone. | Skip 1 to 10. In our view nobody can say what separates a 6 from a 7, and the numbers look more precise than the judgement behind them. ## Rating errors to catch Central tendency, halo, leniency, contrast and recency are the common rating errors. The "why not one higher" column in the scoring sheet catches most of them, because it forces the interviewer to name the evidence the next anchor needed. ![Sample scoring sheet a reviewer filled in for a fictional project coordinator: five criteria scored 2 to 4 out of 5, equal-weight average 3.2.](/blog/images/interview-rating-scale-1.png) *A reviewer reads the assessment per question beside the recording and rates each criterion against your anchors. Illustration with sample data on a sample 1 to 5 scale. Illustration with sample data* Each result carries a written assessment per question, so a reviewer can compare it with your anchors. ## A filled example: project coordinator A fictional candidate. The evidence is written for this page, and the anchors are the ones in the section above. **Table 6: filled example, fictional candidate** | Criterion | Score | Evidence | Why not one higher | | --- | --- | --- | --- | | Prioritising work | 4 | "I rank by deadline, then by who is blocked." Said the reporting task would slip and told the client. | Did not say who else was told. | | Explaining decisions | 3 | Gave the reason for choosing a vendor. | Did not name the option rejected or what would change the call. | | Learning from mistakes | 2 | Named a missed deadline and said "it was the other team". | Named no change to how they work. A 3 needs the fix. | | Stakeholder communication | 4 | Updates the sponsor on Fridays and before a risk lands. | Example had one stakeholder, not conflicting ones. | | Planning | 3 | Plan had milestones and owners. | No fallback if a milestone slipped. | | Average | 3.2 | Equal weights: (4 + 3 + 2 + 4 + 3) / 5. | | The 2 follows the anchors: the mistake is named, but the 3 anchor needs the fix. A line like that is one a debrief can argue with. ## Calibrate before you score - Pick two recorded or written answers and have every interviewer score them alone. - Compare. Any gap of two points or more means an anchor is vague. Rewrite it. - Revisit the anchors after the first ten interviews for a role. Keep the ones that separated candidates and rewrite the ones everybody scored the same. - Give every candidate for the role the same core questions. A scale cannot compare answers to different questions. ## Use the scale in Tracy A scale only helps for the candidates who are scored on it. [Tracy AI Interviewer](/product) interviews each applicant who takes it, on the skills you set. - **Your criteria, your anchors.** Set the skills for each role. Keep your 1, 3 and 5 anchors in your own scoring sheet so reviewers read each assessment against them. A Director at a top-tier early-stage VC advises other teams to "spend time designing the depth and structure of your interview questions during the initial setup. The better you define what you want to evaluate, the more meaningful and actionable the results will be." - **A follow-up that separates a 3 from a 4.** A first answer rarely shows the difference. She follows up on what the candidate said and waits for the answer, in the candidate's own field. - **Evidence for each rating.** A written assessment per question with its reasoning, a recording of both sides and a timestamped transcript, so a reviewer can read the rating against your anchors. - **A person decides.** A recruiter can move a candidate forward or invite someone to interview again, and the full history stays on the record. See [how an AI interview report reads](/blog/ai-interview-reports). The first 3 results are free with no card. After that it costs $1 per credit, one credit per started 20 minutes of an interview that produced a result, with no seats and no upfront cost. See [pricing](/pricing). To try it, write your anchors into one role and run three real interviews. ![Sample interview: Tracy asks how a request was prioritised, the candidate answers, Tracy asks about two clashing deadlines.](/blog/images/interview-rating-scale-2.png) *Tracy follows up on what the candidate said, so the difference between a 3 and a 4 shows up in the answer. Illustration with sample data* The follow-up is where a 3 and a 4 come apart. ## Which path fits you - Scoring many applicants for one role: [volume hiring](/use-cases/volume-hiring). - Technical roles: [technical interviews](/use-cases/technical-interviews). - Choosing the questions the scale will score: [how to screen candidates](/blog/how-to-screen-candidates). ## Common questions ### What is a good interview rating scale? A 1 to 5 scale with a written anchor for each point, in behaviour a candidate shows. The anchor for 3 is the bar for the role, so every interviewer knows what meets it. ### What does a 3 mean on an interview rating scale? The candidate meets what the role needs: a real example and sensible answers to follow-ups. It is the hiring bar, not an average. ### Should I use a 1 to 5 or a 1 to 10 scale? 1 to 5, in our view. Five levels can each be anchored in words. Ten rarely can, so a 1 to 10 scale adds numbers without adding information. ### Should an interview rating scale have an odd or even number of points? Odd gives a clear midpoint to anchor the bar. Use an even scale such as 1 to 4 if your interviewers all settle on the middle and you want them to lean. ### How do you stop every interviewer scoring a 3? Anchor the scale, ask for the evidence and the reason it was not one higher, and calibrate on a shared recording until the gaps close. --- Copy the scale, then run it as a Tracy interview. Send us one role and we will build the interview plan in days, with a recording and report you can check. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=interview-rating-scale&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=interview-rating-scale&utm_content=demo) ## Sources - [Tracy AI Interviewer home page, customer quotes](https://tracyinterview.ai/), retrieved 9 Nov 2026. - [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer Connect page](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. # Phone Screening Script for Recruiters > A free phone screening script for recruiters: opening, knock-out checks, core questions with follow-ups and a call record, plus a filled example. - URL: https://tracyinterview.ai/blog/phone-screening-script - Published: 2026-11-12 - Author: Tracy team > **In short** Tracy AI Interviewer runs your screening script as an interview every applicant starts on a click, with no call to book, and leaves a recording and a written assessment for each question. > > Below: a word-for-word phone screening script with knock-out checks, core questions and a call record, plus a filled example. > > One Tracy customer cut its average time to hire from 44 days to 7. > > Run the script by hand while volume is low. Move it into Tracy when calls pile up. A recruiter's phone screen is short, and the same ten minutes get repeated for every applicant. Skip the script and each call asks different things, the notes are hard to compare, and the hiring manager asks what the candidate actually said. A phone screening script fixes the order, the words for the knock-out questions and the shape of the notes. Below is a phone screening script for recruiters you can copy, a filled example, and how [Tracy AI Interviewer](/product) runs your screening script when each applicant starts from a link and there are too many to call. ## What a phone screen has to settle Three things, in this order. Anything else can wait for the interview with the hiring manager. - **Can they take the job?** Location, right to work, pay, notice. These are yes or no, so ask them in the same words every time. - **Do they have the core skill?** One recent example, with a result. - **Do they want this role?** A reason that fits what the job really is. For the wider funnel around the call, see [how to screen candidates](/blog/how-to-screen-candidates). ## The phone screening script Copy the script into a document and keep it open during the call. Say the quoted lines as written; the rest is yours. A first call of 15 to 20 minutes fits this layout. **1. Opening (about 2 minutes)** > "Hi [name], this is [your name] from [company]. Is now still a good time? I would like about 15 minutes to ask a few questions about your background and the [role] position, then tell you what happens next. I will take notes as we talk. Is that fine?" **2. Knock-out checks (about 3 minutes)** Five checks, each with the question to ask and the answer you need. | Check | Ask | Needed answer | Pass? | | --- | --- | --- | --- | | Location or work arrangement | Where are you based, and does [on-site, hybrid or remote] work for you? | | | | Right to work | Do you have the right to work in [country] without sponsorship? | | | | Pay | What range are you looking for? | | | | Notice and start | What is your notice period, and when could you start? | | | | Other processes | Are you talking to other companies right now? | | | Ask the knock-outs of every candidate in the same words. Write what they said, not only pass or fail. **3. Core questions (about 8 minutes)** Four questions, each with a follow-up and what to listen for. | # | Ask | Follow up with | Listen for | | --- | --- | --- | --- | | 1 | What are you doing in your current role, day to day? | What part is yours alone? | Ownership, a real scope | | 2 | Why are you looking, and why this role? | What would make you say no? | A reason that fits the role | | 3 | [Skill from the job description]: describe a recent time you used it. | What was the result? | A specific example with an outcome | | 4 | [Second skill]: how would you handle [a situation from the job]? | What did you do the last time? | Steps, not a slogan | **4. Their questions and close (about 2 minutes)** > "What questions do you have for me? ... Here is what happens next: [step], and you will hear from me by [date]." **5. Call record** Fill this in straight after the call. | Field | Entry | | --- | --- | | Knock-outs | All passed, or which one failed | | Best quote | | | Concern | | | Recommendation | Submit, hold or decline | | Next step and date | | ## How to use the script - **Say the knock-outs the same way to everyone.** It keeps the call fair and the notes comparable. - **Follow up once, then move on.** One good follow-up beats three questions on the same point. - **Write quotes.** The hiring manager reads what the candidate said, not your summary. - **Stop when a knock-out fails.** Thank them, say what happens next and end the call kindly. - **Send the record the same day.** The hiring manager then reads it while the call is fresh. ## A filled example: IT support specialist A fictional call for an IT support specialist role at a client company. The candidate and answers are invented for this page. | Check or question | What the candidate said | Note | | --- | --- | --- | | Work arrangement | Hybrid, two days on-site in Columbus | Pass | | Notice and start | Two weeks, free from the first of next month | Pass | | Pay | 62 to 68 thousand dollars a year | Within the client's band | | Day to day | Tier 2 desk for about 400 users, owns the onboarding checklist | Real scope | | Why looking | The team is shrinking and nobody mentors new hires | Fits a mentoring role | | Skill: Active Directory | Rebuilt the group structure after a merger, cut access tickets by about a third | Specific, ask for the method | | Situation: a VPN outage | Checked the gateway, then the certificate, told the floor lead in ten minutes | Clear order | Recommendation: submit. Best quote: "I told the floor lead before I knew the cause, so nobody was guessing." Next step: send the profile to the client the same day. ## When calls pile up: run the script in Tracy When one role draws more applicants than you can call in a day, calling everyone is the bottleneck. That is the point to move the first screen into Tracy and keep your own calls for the shortlist. - **Your script becomes the interview.** Paste in the knock-outs and core questions, or let Tracy generate questions from the job description and edit them. - **Every applicant, any hour.** Each applicant starts on a click, with no scheduling and no recruiter present. A reminder goes out after a day, and failed interviews are chased automatically. - **Follow-ups on what was said.** She asks in the applicant's own field, follows up and waits for the answer. - **Candidates enjoy it.** Steven Perry, Account Executive at IT Resources, an IT staffing firm in Tampa, US, says: "By the time they finish with this platform, they're even laughing and joking." - **Evidence for each person.** Each candidate gets a written assessment per question, with the reasoning, and every point links to the moment in the recording. - **A package you can send on.** Resume match, assessment per question, strengths and gaps, recording and fit rationale in one record, shareable by link or PDF. A person decides: you can move a candidate forward or invite someone to interview again. See [staffing agencies](/blog/ai-interview-tools-for-staffing-agencies) and [how an AI interview report reads](/blog/ai-interview-reports). To try it, paste your knock-outs and core questions into one role and run it. ![Four steps: write the script, add it to Tracy, applicants click to start, review the evidence.](/blog/images/phone-screening-script-1.png) *Tracy runs your screening script when each applicant clicks, with a recording and written assessment behind each result. Illustration with sample data* You write the script once; Tracy runs it from a link and leaves the evidence to review. ![Transcript card: Tracy asks notice period, the candidate mentions a Friday second round, Tracy asks how long to decide.](/blog/images/phone-screening-script-2.png) *Tracy asks the script's question, then the follow-up a recruiter would ask next, so the notes hold the detail. Illustration with sample data* The follow-up asks what a recruiter would ask next, on every screen. ## Which path fits you - Agencies and RPO teams sending candidates to clients: [volume hiring](/use-cases/volume-hiring). - Screening applicants who arrive through an apply page: [applications](/use-cases/applications). - What candidates see on their side: [participant experience](/participant-experience). ## Common questions ### What do recruiters ask in a phone screen? Location or work arrangement, right to work, pay range, notice period and start date, then a current-role question, why they are looking and one or two examples of the core skill. Ask the first group in the same words every time. ### How long should a recruiter phone screen be? Fifteen to twenty minutes. The script above fits in that: two minutes to open, three for knock-outs, eight on the core questions and two to close. ### What is a recruiter screening call script? A fixed outline for the first call: the opening words, the knock-out questions, the core questions with follow-ups and a record for the notes. It is the phone version of an interview guide. ### What should you write down on a phone screen? The answer to each knock-out, one quote that shows the core skill, one concern and your recommendation with the next step and date. ### Can Tracy replace the phone screen? She runs the first screen as an interview each applicant starts from a link, with a recording and a written assessment per question. A person still decides who goes forward and talks to the shortlist. --- Copy the script, then run it as a Tracy interview. Send us one role and we will build the interview plan in days, with a recording and report you can check. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=phone-screening-script&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=phone-screening-script&utm_content=demo) ## Sources - [Tracy AI Interviewer home page, customer quotes](https://tracyinterview.ai/), retrieved 9 Nov 2026. - [Tracy AI Interviewer Connect page](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. # Due Diligence Interview Questions for Early-Stage Founders > Due diligence interview questions for early-stage founders: 15 questions with follow-ups, the evidence to request, a scoring sheet and a filled example. - URL: https://tracyinterview.ai/blog/due-diligence-interview-questions - Published: 2026-11-11 - Author: Tracy team > **In short** Tracy AI Interviewer runs this 15-question template (five areas, a follow-up and evidence to request) as the first interview for every founder, with a recording, transcript and assessment per question. > > A Director at a top-tier early-stage VC says Tracy "asks meaningful, investment-relevant questions." > > An interview collects claims, and documents still verify them. Next: set the template up as a Tracy interview. Due diligence interview questions for an early-stage founder have a different job from a pitch Q&A. Ask what the founder knows, how they know it and what they could show you. A seed-stage company has little to audit, so the interview carries more weight than it does later, and a vague one leaves the decision to instinct. Below is a working template: 15 questions in five areas, with a follow-up, the evidence to request and columns for a score and notes. Use it as it stands or edit it. If you run it on more founders than your partners can meet, Tracy AI Interviewer can ask the same questions as a first call and leave a record for each. Our [applications page](/use-cases/applications) shows that flow, and the guide to the [accelerator interview](/blog/accelerator-interview) covers the programme side. ## What are the due diligence interview questions to ask a founder? [Make a copy of the template in Google Sheets](https://docs.google.com/spreadsheets/d/1ynPkMGs8ZbQupqpfGUUnR33ETp7FmCYgCy_-7mJauHA/copy), or copy this table into your own sheet. Ask the question, wait for the answer, then ask the follow-up. The last three columns are yours: ask for the evidence after the call, score from 1 to 5, and write what the founder said in their words. | Area | Ask | Follow up with | Evidence to request | Score (1 to 5) | Notes | | --- | --- | --- | --- | --- | --- | | Customer and problem | 1. Who is the customer, and what does the problem cost them today? | Who paid you last, and what were they using before? | Customer list with start dates | | | | Customer and problem | 2. Tell me about a customer you lost. | How many lost customers gave the same reason? | Churn list with reasons | | | | Customer and problem | 3. What do customers do today when your product is missing? | Which workaround do they like most? | Two customer call notes | | | | Traction | 4. Which number best shows people want this? | How is it counted, and over what period? | The export behind it | | | | Traction | 5. What moved that number last quarter? | What would it have been if you had changed nothing? | Month-by-month figures | | | | Traction | 6. What would you show me if I asked to see the data? | Who else can confirm it? | A reference customer | | | | Market | 7. How big is the market, built from the bottom up? | Which single assumption would change the answer most? | The sheet with the assumptions | | | | Market | 8. Who is the next customer after the first group? | What stops them buying today? | A list of target accounts | | | | Market | 9. Who else solves this, including doing nothing? | What would a well-funded competitor do first? | Competitor notes | | | | Team and execution | 10. What have you shipped in the last 90 days? | What slipped, and why? | Release log or roadmap | | | | Team and execution | 11. Who decides what, and who is missing? | What have you not hired for yet, and why? | Org chart and equity split | | | | Team and execution | 12. Tell me about a disagreement between the founders. | How was it settled, and would you settle it that way again? | None | | | | Why now and risks | 13. What changed that makes this possible now? | What happens if that change reverses? | A source for the change | | | | Why now and risks | 14. What is the biggest risk to the plan? | What have you done to test it? | The test and its result | | | | Why now and risks | 15. What would you do with the next 12 months of funding? | What would tell you at six months that it was going wrong? | Budget by quarter | | | ## How do you score the answers? A score is only useful if two partners reading the same answer reach the same one. Use anchors. | Score | What the answer sounds like | | --- | --- | | 1 | No specifics. Adjectives where a number should be. | | 2 | Some detail, but it cannot say where it came from. | | 3 | Specific on the point asked, with no way yet to confirm it. | | 4 | Specific, consistent with other answers, and the founder offers the source. | | 5 | All of 4, and the founder volunteers a weakness before you ask. | Weight the areas to the round. For a pre-seed programme you might weight team and customer more heavily. For a later round, traction and market. In Tracy you choose the skills to evaluate as tags when you set up the interview, so the same five areas steer the questions and the assessment for every founder. A founder interview is also the place to compare founders with each other, so use the same opening questions for everyone. See our guide to [how to screen candidates](/blog/how-to-screen-candidates) for the general method. ## What does a filled-in scoring sheet look like? This is a filled example with sample data. The company and the founder are fictional. | Ask | What the founder said (sample) | Follow-up and answer | Score | Notes | | --- | --- | --- | --- | --- | | 2. Tell me about a customer you lost. | A bakery left in March. Setup took a week and the owner gave up. | Four of seven lost customers left over setup, three over price. | 4 | Offered the churn list. Has cut setup to a day. | | 5. What moved that number last quarter? | Trial sign-ups doubled after the setup change. | Would have grown about 10 percent without it, from the old trend. | 3 | Estimate only. Ask for monthly figures. | | 7. How big is the market? | Roughly 40,000 small bakeries in the markets served. | Assumes 30 percent would pay at $80 a month. The 30 percent is a guess. | 3 | Named the weak assumption without being asked. | The scores sit beside the founder's words, so a second partner can read the same line and disagree with it. ![Partner scorecard with five diligence areas scored out of 5, average 3.7, marked as sample data.](/blog/images/due-diligence-interview-questions-2.png) *A partner's own scorecard, filled in from the call with the anchors above. Sample data, fictional founder. Sample data, fictional founder.* A partner's scorecard for one founder: five areas, each scored out of 5 from the call using the anchors above. ## What can an interview not tell you? An interview collects claims. It does not confirm them. Revenue, contracts, the cap table, intellectual property and legal standing need documents, and the third column of the table is there so you ask for them. Tracy records what the founder says; checking documents, customers and credentials is a step your team does. Treat the interview as the step that decides which founders earn the document review. It also does not replace a conversation with a person. The second round, and the references, stay with your partners. ## How do you run these questions as a Tracy interview? Copy the table, then run it as a Tracy interview in four steps. 1. **Load the questions.** Paste the 15 questions into a new interview, or describe what you want to learn and let Tracy draft them. Edit freely, and choose a style from quick to deep. 2. **Set the criteria.** Enter the five areas as the skills to evaluate. 3. **Share one link.** Put it in the application form or an email. A founder opens it in a browser with no account, at any hour, and is told that the conversation is recorded and transcribed. 4. **Read the results.** Tracy asks each question, follows up on the founder's answer and waits for them to finish. Each founder gets a recording of both sides, a timestamped transcript, a written assessment per question with its reasoning. A partner reads them and decides. The diagram below shows the path. A result can be shared by link or PDF with an investment committee. The [product page](/product) lists the settings, and [AI interview reports](/blog/ai-interview-reports) shows how a report reads. Request the documents in the third column yourself after the call. ![Four steps: load 15 questions, share one link, Tracy interviews, a partner decides.](/blog/images/due-diligence-interview-questions-1.png) *Tracy turns the template into one link, an interview for every founder and a report a partner decides from. Illustration of the workflow. Sample data, fictional founder.* The template becomes a first call for every founder, and a partner decides from the report. ## When is a call with a partner enough? If you diligence ten founders a quarter, a partner on a video call is likely cheaper and better than any tool. The case for an automated first call starts when applications outnumber partner hours. Our guide to [AI interview versus human interview](/blog/ai-interview-vs-human-interview) compares the two. Tracy costs $1 per credit, one credit per started 20 minutes of an interview that produced a result. The first 3 results are free with no card, there are no seats and a session with no answers is never charged. 100 interviews of 20 minutes cost $100. Try one on the [Connect page](/connect), and see [pricing](/pricing) for the current plan. [Make a copy of the template in Google Sheets](https://docs.google.com/spreadsheets/d/1ynPkMGs8ZbQupqpfGUUnR33ETp7FmCYgCy_-7mJauHA/copy) and fill it in on your next call. Which path fits you: - Programme, fellowship or scholarship applications: [applications](/use-cases/applications). - Experts to vet for a project: [expert vetting](/use-cases/expert-vetting). - Customer discovery or win-loss calls: [customer research](/use-cases/customer-research). ## Common questions ### What are due diligence interview questions? Questions that test what a founder knows and can prove about the customer, the traction, the market, the team and the risks. Each should have a follow-up that asks for a count, an example or a document. ### What are some due diligence interview questions and answers for a founder? Ask who paid last and what they used before. A strong answer names the customer and the thing it replaced. A weak one describes a market. The table above gives 15 questions with follow-ups and what to listen for. ### How do you do due diligence on an early-stage startup? Interview the founders, ask for the documents behind each claim, speak to two or three customers, and score what you learn the same way for every company. ### Can due diligence interviews be automated? The first interview can. Tracy asks your questions, follows up and leaves a recording, a transcript and an assessment per question. Document review and references stay with your team. ### How long should a founder due diligence interview be? Thirty to forty-five minutes covers 15 questions with follow-ups. That is our rule of thumb. --- Send us your questions and we will set up the interview and report for you in days. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=due-diligence-interview-questions&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=due-diligence-interview-questions&utm_content=demo) ## Sources - [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer Connect page](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. - [Tracy AI Interviewer product page](https://tracyinterview.ai/product), retrieved 9 Nov 2026. - [Tracy AI Interviewer home page, customer testimonials](https://tracyinterview.ai/), retrieved 9 Nov 2026. # First-Round Interview Questions to Ask Candidates > First round interview questions to ask candidates: 16 questions in four groups, each with a follow-up, what to listen for, and a six-question template. - URL: https://tracyinterview.ai/blog/first-round-interview-questions - Published: 2026-11-10 - Author: Tracy team > **In short** Tracy AI Interviewer runs this first round for every applicant: questions from your job description, follow-ups in the candidate's own field, and a recording and assessment per question. > > Below: sixteen questions in four groups with follow-ups and what to listen for, and a six-question 20-minute template. > > Next step: send us one role and we build the interview plan. The first round has a short job: decide who deserves more of your team's time. Most first rounds do it badly. They ask "tell me about yourself", hear a rehearsed answer and move on, and the follow-up that would have shown whether the person did the work never gets asked. The first round interview questions below are built around the follow-up. Each one comes with the question to ask after the first answer and a note on what a good answer contains. Use them for any role by swapping the role-specific questions in the first group. Further down, we show how [Tracy AI Interviewer](/product) runs the same first round for every applicant and hands your team a recording and a written assessment to read. ## What a first round must settle Four things, in this order. Everything else can wait for the second round. - **Can they do the work?** Skills for this role, in their own field. - **How do they think?** Reasoning they can explain, not answers they recited. - **How do they work?** Ownership, feedback and working with other people. - **Can they start?** Hours, time zone and start date. Ask every candidate for the role the same core questions. A 2022 re-analysis by Sackett and colleagues in the Journal of Applied Psychology, covering earlier published meta-analyses, put the mean validity of structured interviews at .42 and of unstructured interviews at .19. That is a correlation between the interview rating and later job performance ratings, based on 105 and 32 studies (7,864 and 3,125 people) ([source](https://doi.org/10.1037/apl0000994)). Same questions give you answers you can compare. Follow-ups let you check them. ## Sixteen first round interview questions, grouped by what they settle Pick one or two from each group and keep the follow-up beside the question. Ask it only after you have heard the first answer. ### Group 1: can they do the work | Ask | Follow up with | Listen for | | --- | --- | --- | | 1. Walk me through a piece of work you finished recently and are proud of. What was your part? | What would you do differently now? | Their own decisions, separate from the team's. A result with a number or a consequence. Owns a miss. | | 2. Which tools or methods do you use every day for this job, and which one limits you? | Why did you choose it over the alternative? | Reasons from experience. A list lifted from the job advert sounds the same for every candidate. | | 3. What is the hardest problem you solved in the last year? | What did you try first that did not work? | Detail about the dead end. Honest account of the time it took. | | 4. A role-specific question you could only answer if you do the job. For support: a customer reports an error you have not seen. What do you ask first? | What do you do if they cannot send you logs? | A sensible first question, then a second path when the first closes. | ### Group 2: how they think | Ask | Follow up with | Listen for | | --- | --- | --- | | 5. You have two ways to do a task and little time. How do you choose? | What did that choice cost you? | Names the criteria and the risk, and admits the cost. | | 6. Tell me about a time you were wrong about something at work. | What did you check afterwards? | Ownership without blaming someone else. A habit they changed. | | 7. How do you decide that a piece of work is done? | Tell me about a time it was not done when you thought it was. | A standard they apply, such as checks, tests or a review by someone else. | | 8. Explain something from your field to someone who does not work in it. | What did they ask that you did not expect? | Plain words. Checks that the other person understood. | ### Group 3: how they work | Ask | Follow up with | Listen for | | --- | --- | --- | | 9. Tell me about feedback that changed how you work. | What did you stop doing? | A specific change, not "I take feedback well". | | 10. Describe a time you disagreed with a colleague about how to do something. | What happened afterwards? | Listened first, used evidence, and the relationship held. | | 11. Tell me about a task you finished later than planned. | When did you tell the person waiting, and what did you change next time? | Early warning. A concrete change in how they plan. | | 12. What do you do when you are blocked and nobody is around? | How long do you try before you ask for help? | Initiative with a limit. | ### Group 4: can they start | Ask | Follow up with | Listen for | | --- | --- | --- | | 13. Which hours and time zone can you work, and from when? | Is anything likely to change that in the next three months? | Clear, consistent answers. | | 14. What made you apply for this role? | What do you want to be doing in a year? | Interest in the work itself and some preparation. | | 15. What would you need in your first month to do good work? | Which of those can you arrange yourself? | Realistic needs, and some independence. | | 16. What would you like to ask us about the role? | What would make you say no to an offer? | Questions about the work. A clear idea of what matters to them. | Skip questions about age, family plans, health, religion or where someone is from. They do not predict the work and can create legal exposure in many places. If you are unsure about a question, ask your counsel before you use it. ## A first round template: six questions in 20 minutes Copy this into your own document and swap in questions from the groups above. The minutes are a guide. | Block | Minutes | Question | | --- | --- | --- | | Opening | 1 | Say how the interview works and that every candidate gets the same questions. | | Can they do the work | 5 | Question 1 and the role-specific question 4, each with its follow-up. | | How they think | 4 | Question 5 with its follow-up. | | How they work | 5 | Question 11 with its follow-up. | | Can they start | 3 | Question 13, then question 14. | | Close | 2 | Question 16, then what happens next and when. | Write what a strong answer contains for each question before the first candidate joins. For a technical role, see our [software engineer interview questions](/blog/software-engineer-interview-questions). Our guide to [AI behavioral interviews](/blog/ai-behavioral-interviews) covers the situation-and-result questions in groups 2 and 3. ## Run the first round with Tracy Writing good questions takes an afternoon. Asking them of two hundred applicants takes a team, and this is where Tracy takes over. - **Set up from the job description.** Tracy generates the questions from the role. Edit them, or use the sixteen above as your own. Choose a style from quick to deep, technical to behavioural. - **Follow-ups on what the candidate said.** She asks in the candidate's own field, follows up on what the candidate said, using the follow-up column above as a guide, and waits for the answer. A Director at a top-tier early-stage VC said: "Tracy sometimes brings up exactly those questions I might have forgotten to ask, which has been a pleasant surprise." - **The CV, when you want it.** With the CV check on, Tracy reads the CV. - **Same first round, any hour.** Each applicant starts on a click with no scheduling and no recruiter present. The person interviewed needs only a browser link and no account, and is told on the join screen that the conversation is recorded and transcribed. See the [participant experience](/participant-experience). - **Evidence per result.** Your team reads a recording of both sides, a timestamped transcript and a written assessment per question with its reasoning. Group 4 answers, such as hours and start date, appear in the written assessment. Every point links to the moment in the recording. A person decides who goes forward. Read more on [how AI interview reports work](/blog/ai-interview-reports) and [AI screening interviews](/blog/ai-screening-interviews). A 20-minute first round is one credit, which is $1. One hundred of them cost $100. The first 3 results are free with no card, there are no seats and no upfront cost. See [pricing](/pricing). To try it, create an interview for this exact role and run it with three real applicants. ![Four steps in Tracy: your questions, Tracy follows up, an assessment per question, evidence to read.](/blog/images/first-round-interview-questions-2.png) *Tracy asks your questions, follows up on what each candidate said, writes an assessment per question and leaves the evidence for your team. Illustration with sample data* Tracy asks your questions, follows up, writes an assessment per question and leaves the evidence to read. ![A three-row transcript card. Tracy asks about a late task, the candidate mentions duplicate accounts, and Tracy asks when they told the account owner.](/blog/images/first-round-interview-questions-1.png) *Tracy asks the follow-up a rehearsed answer leaves out, so the team sees the detail. Illustration with sample data* The follow-up asks for the detail a rehearsed answer leaves out. ## Which path fits you - Screening applicants for open roles: [applications](/use-cases/applications). - Many applicants for the same role: [volume hiring](/use-cases/volume-hiring). - Wondering how to screen before you interview: [how to screen candidates](/blog/how-to-screen-candidates). ## Common questions ### What are good first round interview questions to ask candidates? Questions that settle four things: whether they can do the work, how they think, how they work and whether they can start. Pick one or two from each group above, and keep the follow-up beside each question. ### How long should a first round interview be? Twenty minutes covers six questions with their follow-ups. That is a working rule, not a researched figure. Extend it only if the role needs a skills check that cannot wait for the second round. ### What are the best first interview questions for employers to avoid? Skip questions about age, family plans, health, religion and origin. They do not predict the work and can create legal exposure in many places. Check unfamiliar questions with your counsel. ### Should every candidate get the same first round questions? Yes. The same core questions let you compare answers. Follow-ups can differ, because they depend on what each candidate said. ### Can an AI interviewer run the first round? Yes. Tracy asks the questions, follows up on each answer and gives your team a recording, a timestamped transcript and a written assessment per question. A person decides who goes forward. ### How do I score the answers? Write down what a strong answer contains for each question before the first candidate joins, then score each answer against it with one quote of evidence. In Tracy you choose the skills to evaluate as tags for each role, and each question gets a written assessment with its reasoning. --- Copy the template, then run it as a Tracy interview. Send us one role and we will build the interview plan from it in days, with a recording and report you can check. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=first-round-interview-questions&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=first-round-interview-questions&utm_content=demo) ## Sources - [Sackett, Zhang, Berry and Lievens (2022), Revisiting meta-analytic estimates of validity in personnel selection, Journal of Applied Psychology 107(11)](https://doi.org/10.1037/apl0000994), retrieved 9 Nov 2026. - [Tracy AI Interviewer home page, customer quotes](https://tracyinterview.ai/), retrieved 9 Nov 2026. - [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 5 Nov 2026. - [Tracy AI Interviewer Connect page](https://tracyinterview.ai/connect), retrieved 5 Nov 2026. # Interview Scorecard Template > A free interview scorecard template you can paste into Excel or Google Sheets: weighted criteria, a 1 to 5 scale, evidence column and a filled example. - URL: https://tracyinterview.ai/blog/interview-scorecard-template - Published: 2026-11-10 - Author: Tracy team > **In short** Tracy AI Interviewer interviews every applicant on the skills you pick, with a recording and a written assessment per question behind each result. > > Below: five weighted criteria, a 1 to 5 scale, an evidence column, a decision block, the Excel formula and a filled example. > > A person decides who goes forward. > > Next step: send us one role and we build the interview plan from your scorecard. The debrief starts well enough. Then three interviewers say "good energy", "a bit thin" and "strong" about the same candidate, and the loudest voice wins. An interview scorecard template ends that. Every interviewer scores the same criteria on the same scale and writes down what they heard, so the debrief compares evidence and not impressions. Below is a template you can paste into a sheet, a filled example, and how [Tracy AI Interviewer](/product) interviews every applicant on the skills you pick and writes an assessment for each question. ## Why a scorecard matters Structure in the interview is what a scorecard enforces. A 2022 re-analysis in the Journal of Applied Psychology by Sackett and colleagues, covering earlier published meta-analyses, put the mean validity of structured employment interviews at .42 and of unstructured interviews at .19. That is a correlation between the interview rating and later job performance ratings: 105 studies and 7,864 people for structured interviews, 32 studies and 3,125 people for unstructured ones ([source](https://doi.org/10.1037/apl0000994)). That is a stronger link and no guarantee. A scorecard is the scoring half of structure. The other half is asking every candidate for the role the same questions. ## The interview scorecard template [Make a copy of the scorecard in Google Sheets](https://docs.google.com/spreadsheets/d/1NUagAWAumI1RtioL_t4s-BjqD026bGLCxWXHk07JA2U/copy). The sheet has the same fields, weights and scale as the tables below, and the formulas are for you to add. Or copy each table into your own sheet. With the weights in column C and the scores in column D, rows 2 to 6, work out the weighted score in Excel or Google Sheets with `=SUMPRODUCT(C2:C6,D2:D6)/SUM(C2:C6)`. **1. Header** | Field | Entry | | --- | --- | | Candidate | | | Role | | | Round and date | | | Interviewer | | | Recording or notes link | | **2. Scoring** | Criterion | What it covers | Weight | Score (1 to 5) | Evidence: quote or time | | --- | --- | --- | --- | --- | | Criterion 1 | The skill or behaviour, in one line | 25 | | | | Criterion 2 | | 25 | | | | Criterion 3 | | 20 | | | | Criterion 4 | | 15 | | | | Criterion 5 | | 15 | | | **3. The scale** | Score | Meaning | | --- | --- | | 1 | No relevant example, or the answer shows the skill is missing. | | 2 | A partial example. Cannot go deeper when asked. | | 3 | Meets the bar. A real example and sensible answers to the follow-up. | | 4 | Above the bar. Specific detail, results and what they would change. | | 5 | Strong. Deep, consistent under follow-up, and could teach it. | **4. Decision** | Field | Entry | | --- | --- | | Weighted score | | | Recommendation | Advance, hold or do not advance | | Biggest risk | | | What to probe next | | Pick four or five criteria that the role really needs, and make the weights add up to 100. More than five and the scores blur, because nobody remembers six things from a twenty-minute conversation. ## How to use the scorecard - **Set the criteria and weights before the first interview.** Choosing weights after you have met the candidates defeats the point. - **Each interviewer scores alone.** Write the scores and evidence before the debrief, and do not look at anyone else's first. - **One quote or moment per score.** If you cannot point to something the candidate said or did, the score is a guess. - **Score the answer, not the person.** Accent, confidence and a shared background are not criteria. - **Discuss the gaps.** A difference of two points or more on one row is where the debrief should spend its time. - **Keep the candidates comparable.** Same criteria, same scale and the same core questions for every candidate for the role. ![Four steps from template to result: copy the template, set up Tracy, candidates interview, read the evidence.](/blog/images/interview-scorecard-template-1.png) *Tracy turns the scorecard into an interview and an evidence-linked result for every applicant. Illustration with sample data* The scorecard works because the order is fixed: criteria first, scores alone, evidence written, then the discussion. ## A filled example: support engineer A fictional candidate for a support engineer role. The evidence is written for this page. | Criterion | Weight | Score | Evidence | | --- | --- | --- | --- | | Troubleshooting | 25 | 4 | "I ask for the exact error and the time it started, then check the logs." Followed with two causes ranked by likelihood. | | Customer communication | 25 | 5 | "I describe what they will see on screen before I name the setting." Kept it plain without being asked. | | Ownership | 20 | 4 | Told a customer a fix would take two days and called back at the end of day one. | | Tool knowledge | 15 | 3 | Named the ticket tool's macros but could not say how they would automate a repeated case. | | Availability and fit | 15 | 4 | Can work the European afternoon shift from the first of next month. | The weighted score is (25 x 4 + 25 x 5 + 20 x 4 + 15 x 3 + 15 x 4) / 100 = 4.1. Recommendation: advance. Next round: probe the tool knowledge row. ## Build the Tracy interview from your scorecard A scorecard only helps for the candidates you actually interview against it. Tracy interviews every applicant on the skills you pick, with a recording and a written assessment per question behind each result. - **Set up from the job description.** Tracy generates the questions from the role. Edit them or use your own, from a quick screen to a deep one, technical to behavioural. - **Your criteria.** You pick the skills for each role. The five rows of your scorecard become the skills. - **Follow-ups.** She asks in the candidate's own field, follows up on what they said and waits for the answer. - **Evidence per result.** Each result has a recording of both sides, a timestamped transcript and a written assessment for each question with its reasoning. A reviewer can read the recording and the transcript behind each result. - **Every applicant, any hour.** Each applicant starts on a click with no scheduling, so every applicant gets an interview and a result. One Tracy customer cut its average time to hire from 44 days to 7. A person reviews the evidence and decides. A recruiter can move a candidate forward or invite someone to interview again, and the full history stays on the record. See [how an AI interview report reads](/blog/ai-interview-reports), how [AI screening interviews](/blog/ai-screening-interviews) fit a funnel, and the [volume hiring](/use-cases/volume-hiring) page for the full flow. Cost is $1 per credit, one credit per started 20 minutes of an interview that produced a result. One hundred 20-minute interviews cost $100. The first 3 results are free with no card, and there are no seats and no upfront cost. See [pricing](/pricing). To try it, [make a copy of the scorecard in Google Sheets](https://docs.google.com/spreadsheets/d/1NUagAWAumI1RtioL_t4s-BjqD026bGLCxWXHk07JA2U/copy), paste your five criteria into one role and run three real interviews. ![Scorecard a reviewer fills in for a fictional support engineer: five criteria scored out of 5 and a weighted total of 4.1.](/blog/images/interview-scorecard-template-2.png) *A reviewer fills in the team's scorecard from the recording, the transcript and the assessment per question. Illustration with sample data* A reviewer reads the evidence and fills in the team's own scorecard. ## Which path fits you - Screening applicants for one role in large numbers: [volume hiring](/use-cases/volume-hiring). - Technical roles: [technical interviews](/use-cases/technical-interviews). - Checking a candidate funnel before you set criteria: [how to screen candidates](/blog/how-to-screen-candidates). ## Common questions ### What should an interview scorecard include? The candidate, role and interviewer, four or five weighted criteria, a defined 1 to 5 scale, a line of evidence for each score, and a decision block with the weighted score and a recommendation. ### How many criteria should a scorecard have? Four or five. Each one should be something the role really needs and something an interview can show. More than five and interviewers cannot hold them in mind. ### Is there an interview scorecard template for Excel? Yes. Paste the tables above into Excel or Google Sheets and use SUMPRODUCT to weight the scores. The same layout works as a printed form. ### Should interviewers see each other's scores? Not before they have written their own. Collect scores and evidence first, then compare in the debrief. ### How do you weight the criteria? Give more weight to what is hardest to teach for this role, and make the weights add up to 100. Fix them before the first interview and do not change them after you have met candidates. ### Can Tracy use my scorecard? Yes, as the skills to cover. You pick them per role from your scorecard rows, and each result has an assessment per question with the evidence behind it. Your reviewers fill in the scorecard from that. --- Copy the template, then run it as a Tracy interview. Send us one role and we will build the interview plan from it in days, with a recording and report you can check. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=interview-scorecard-template&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=interview-scorecard-template&utm_content=demo) ## Sources - [Sackett, Zhang, Berry and Lievens (2022), Revisiting meta-analytic estimates of validity in personnel selection, Journal of Applied Psychology 107(11)](https://doi.org/10.1037/apl0000994), retrieved 9 Nov 2026. - [Tracy AI Interviewer home page, customer quotes](https://tracyinterview.ai/), retrieved 9 Nov 2026. - [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 5 Nov 2026. - [Tracy AI Interviewer Connect page](https://tracyinterview.ai/connect), retrieved 5 Nov 2026. # System Design Interview Rubric (for Interviewers) > A system design interview rubric you can copy: five weighted criteria, score anchors, a filled example and a follow-up question for each criterion. - URL: https://tracyinterview.ai/blog/system-design-interview-rubric - Published: 2026-11-10 - Author: Tracy team > **In short** Tracy AI Interviewer runs a first-round design conversation on these criteria: she asks the design questions, follows up on each answer and writes an assessment for each question. > > Every result keeps a recording, a timestamped transcript and a written assessment per question, and a person decides. > > Below: five weighted criteria with 1, 3 and 5 anchors, a follow-up per row and a filled example. > > Next step: send us one role and we build the interview plan. Two interviewers watch the same recorded design session and score it a 2 and a 4. One liked the diagram. The other noticed the candidate never asked how many users the system would have. A system design interview rubric stops that argument before it starts, because it writes down what each score means before anyone is interviewed. This one has five criteria, weights you can change, and anchors for a weak, solid and strong answer. Use it as it is for a mid-level or senior backend role, or change the wording to fit your stack. It works for a live round with an engineer and as the scoring sheet behind a recorded interview. Further down, we show how [Tracy AI Interviewer](/product) runs the first round from the same rubric, so your engineers read the evidence and spend their time on the candidates worth a deep round. ## The system design interview rubric: five criteria, scored 1 to 5 Score each row on its own, then combine the five with the weights. A 2 sits between the 1 and 3 descriptions and a 4 between the 3 and 5. The follow-up in the last column is what you ask when the first answer is thin. | Criterion (weight) | 1: weak | 3: solid | 5: strong | Follow-up to ask | | --- | --- | --- | --- | --- | | Requirements and scope (15%) | Starts drawing at once and assumes the problem. | Asks about users, scale and the main use, and says what is out of scope. | Turns vague answers into numbers and says which requirement drives the design. | Which of those numbers would change your design the most? | | Core design (20%) | Boxes with no flow. Cannot say what each part does. | Clear components and a clear request path, with sensible interfaces between them. | The simplest design that meets the requirements. Every part earns its place, and one request can be walked end to end. | Walk one request from the client to the data and back. | | Data and storage (20%) | Names a database with no reason. | Picks a store for the access pattern and knows the main key or table. | Models the data around reads and writes, and covers consistency, retention and growth. | What is the key, and which read does it make slow? | | Scale and failure (25%) | Adds caches and replicas by reflex. | Finds the likely bottleneck and says what breaks first. | Estimates the load, tests the design against a failure and says how they would know it failed. | Take the busiest part down. What does the user see? | | Trade-offs and communication (20%) | Offers one option with no costs, and is hard to follow. | Compares two options, picks one with a reason and stays organised. | Names what the choice costs and what they would revisit with more time, and adjusts when you push back. | What would make you change this choice? | The weights are a default. For a platform role, raise scale and failure. For a role that works mostly with data, raise data and storage. Fix the weights before the first interview and do not move them after you have seen candidates. ## How to score without letting the diagram decide A design interview invites a vivid impression. These habits keep the score tied to what was said. - **Score each criterion before you total.** Do the five rows first and the overall score last. - **Write one piece of evidence per score.** A short quote or the moment it happened. A score with no evidence is a feeling. - **Give everyone the same opening prompt** for the role, and let the follow-ups differ with the answers. - **Do not score handwriting, tidiness or the tool.** Score the reasoning. - **Move the bar with the level, not the rubric.** A mid-level candidate who scores 3 on every row is a reasonable pass. For a senior role, ask for 4 or more on scale and failure and on trade-offs. That is a starting point, so set your own from your last few hires. - **Talk after everyone has scored.** If two scores on the same row differ by two points or more, replay the evidence and decide which reading the evidence supports. The same discipline applies to any role. Our [how to screen candidates](/blog/how-to-screen-candidates) guide covers the first-round decisions around it. ## A filled example: a short-link service The prompt: design a service that creates short links and redirects people who open them. The candidate is fictional and the evidence is written for this page. | Criterion | Weight | Score | Evidence | | --- | --- | --- | --- | | Requirements and scope | 15% | 4 | Asked for the read-to-write ratio and how long links live before drawing anything. Did not ask about abuse. | | Core design | 20% | 4 | Separate API, key service and store. Walked a redirect from the browser to the store and back. | | Data and storage | 20% | 3 | Chose a key-value store because "it scales". Could not say what a lookup costs when the table grows. | | Scale and failure | 25% | 4 | Estimated 10,000 redirects a second, named the cache as the first bottleneck and said what happens on a burst of cache misses. | | Trade-offs and communication | 20% | 3 | Compared random and sequential keys and chose random. Named the cost only when asked, and rushed the last ten minutes. | The weighted score is 0.15 x 4 + 0.20 x 4 + 0.20 x 3 + 0.25 x 4 + 0.20 x 3 = 3.6. For a mid-level role that clears the bar. The two 3s say where to spend the next round: ask about the cost of a lookup at scale, and ask the candidate to argue for the option they did not choose. ![A scorecard a reviewer filled in, titled System design interview, with five criteria scored out of 5 and a weighted total of 3.6.](/blog/images/system-design-interview-rubric-1.png) *A reviewer scores the recorded interview on the rubric's criteria, using the recording, the transcript and the assessment per question. The scale shown is this rubric's own. Illustration with sample data* Five scores on your criteria become one weighted result a hiring lead can compare across candidates. ## Five prompts to use with the rubric Pick one prompt per interview and give every candidate for the role the same one. Each has a follow-up that checks the data and failure rows, the two where thin answers hide longest. 1. **Design a service that shortens links and redirects visitors.** Follow-up: what happens when one link gets a million visits in a minute? 2. **Design a rate limiter for a public API.** Follow-up: where do the counters live, and what happens when that store is slow? 3. **Design a notification system for email and mobile push.** Follow-up: what if the same event fires twice? 4. **Design a feed of the latest items for ten million users.** Follow-up: when does one write turn into many? 5. **Design file upload and processing for large files.** Follow-up: what happens when the processing step fails halfway? For coding and debugging questions that sit beside this round, see our [software engineer interview questions](/blog/software-engineer-interview-questions). ## Run the first round with Tracy from this rubric A person writes the rubric once. Tracy interviews every applicant on its criteria. - **Set up from the job description.** Paste the description for the role and Tracy generates the questions from it. Edit them or bring your own, and choose a style from quick to deep, technical to behavioural. - **Your criteria.** You pick the skills for the role. Put the five rows above in as the skills. Your engineers keep the weights and do the scoring on the rubric. - **Follow-ups that depend on the answer.** Tracy asks in the candidate's own field, follows up on what they just said and waits for the end of the thought. That is what the rubric's last column does for a human interviewer. A Director at a top-tier early-stage VC said: "Tracy sometimes brings up exactly those questions I might have forgotten to ask, which has been a pleasant surprise." - **Evidence per result.** Your team gets a recording of both sides, a timestamped transcript and a written assessment for each question with its reasoning, so a second engineer can check an assessment without sitting through the whole interview. A person decides. Tracy runs a first-round design conversation, a two-way video interview on one prompt, on your criteria. The deep design round stays with your engineers, and it goes to the candidates who showed their reasoning. Our guide to [AI technical interviews](/blog/ai-technical-interviews) covers where the line sits, and [how AI interview reports work](/blog/ai-interview-reports) shows what a reviewer reads. Cost is $1 per credit, one credit for each started 20 minutes of an interview that produced a result. A 34-minute design interview takes two credits, which is $2. The first 3 results are free with no card, there are no seats and no upfront cost. See [pricing](/pricing). To see the rubric applied, set up one design role and run it as a Tracy interview. ![Transcript card: Tracy asks what happens when a queue backs up, the candidate says they cap it, Tracy asks who sets priority.](/blog/images/system-design-interview-rubric-2.png) *Tracy follows up on the candidate's own answer, so the reviewer sees whether the trade-off was understood. Illustration with sample data* A follow-up built from the candidate's own answer shows whether they understood the trade-off or only named it. ## Which path fits you - Hiring engineers for a company: start with [technical interviews](/use-cases/technical-interviews). - Many applicants for the same engineering role: see [volume hiring](/use-cases/volume-hiring). - Approving engineers as experts for AI projects: see [expert vetting](/use-cases/expert-vetting). ## Common questions ### What is a system design interview rubric? A written scoring guide for a system design interview. It lists the criteria you judge, the weight of each and what a weak, solid and strong answer sounds like, so two interviewers reach the same score from the same evidence. ### How do you score a system design interview? Score each criterion from 1 to 5 against its anchors, write one piece of evidence for each score, then multiply by the weights and add them up. Compare scores after everyone has scored, not before. ### What should system design interview evaluation criteria include? Requirements and scope, core design, data and storage, scale and failure, and trade-offs with communication. Add a sixth for the specifics of your stack if the role needs one, and keep the weights at 100. ### Which system design interview questions work with this rubric? Any open design prompt where the candidate has to ask about requirements before drawing. The five prompts above are a starting set, each with a follow-up for the data and failure rows. ### Can an AI interviewer run a system design first round? Yes, as a first-round conversation on your criteria. Tracy runs it as a two-way video interview, and your engineers run the deeper design round afterwards. ### Can I change the weights? Yes. Change them for the role and fix them before the first interview. In Tracy you pick the skills for each role, and the weights stay in your own scoring sheet. --- Copy the template, then run it as a Tracy interview. Send us one role and we will build the interview plan from it in days, with a recording and report you can check. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=system-design-interview-rubric&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=system-design-interview-rubric&utm_content=demo) ## Sources - [Tracy AI Interviewer home page, customer quotes](https://tracyinterview.ai/), retrieved 9 Nov 2026. - [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 5 Nov 2026. - [Tracy AI Interviewer Connect page](https://tracyinterview.ai/connect), retrieved 5 Nov 2026. # Interview Questions for Data Scientists (With the Follow-Ups) > 20 data scientist interview questions on experiments, modelling, SQL, metrics and communication, each with the follow-up that separates real skill from a polished answer. - URL: https://tracyinterview.ai/blog/data-scientist-interview-questions - Published: 2026-11-06 - Author: Tracy team Fact-checked on 3 Nov 2026 against the public job posts and research listed in Sources. > **In short** > - **Tracy AI Interviewer can read each data scientist's CV, add 3 to 5 questions specific to it and put your question bank to every applicant,** with follow-ups that depend on the previous answer and a written assessment per question. > - Use the 20 questions below as a bank: pick four or five, one per area, and always ask the follow-up. Most lists of data scientist interview questions are written for candidates: a question and a model answer to memorise. A candidate who has read the same list gives you the model answer, and you learn nothing about whether they did the work behind it. So this guide is written for the interviewer. Each question comes with the follow-up to ask once you have heard the first answer, and with what a strong answer contains. The follow-up is built on what the candidate just said, so it cannot be rehearsed. That is the part Tracy AI Interviewer takes over when many people apply: it asks the opening question, listens, and asks the follow-up that answer calls for, for each applicant, at any hour. ## What does the role need, according to two public job posts? Job posts differ, but two current ones show the common core. Reddit's [Senior Data Scientist, Ads posting](https://job-boards.greenhouse.io/reddit/jobs/8104403) asks for a strong understanding of "statistical modeling, machine learning algorithms, causal inference and experimental design", experience with SQL, and lists communicating findings to stakeholders as a responsibility. Airbnb's [Senior Data Scientist, Guest & Host posting](https://careers.airbnb.com/positions/7662244) asks for "experimentation, causal observational analysis, and machine learning techniques" and for "an ability to communicate complex technical concepts to a non-technical audience." Both are senior roles at product companies, so treat them as a sample, not a standard. They still point to the same five things to test: - Statistics and experiments: can they reason about evidence and uncertainty? - Modelling: can they choose, check and defend a model? - Data handling and SQL: do they trust data too early? - Metrics and product sense: do they know what a number is for? - Communication and ownership: can others act on what they say? Data scientists are also hired as experts for AI training work. DataAnnotation, for one, lists a [Data Scientist role](https://www.dataannotation.tech/) in Data & ML, shown on its home page as "Up to $150+ / hr". For that kind of role the same five areas apply, with extra weight on spotting confident wrong answers. Our guide to [AI trainer interview questions](/blog/ai-trainer-interview-questions) covers that side, and [how to vet domain experts at scale](/blog/vet-domain-experts-at-scale) covers the process around it. ## Why does the follow-up matter more than the question? Structured interviews, where every candidate gets the same questions and you score against the same criteria, do well in research. A 2022 re-analysis in the Journal of Applied Psychology (Sackett and colleagues) found structured interviews to be the top-ranked selection procedure for predicting job performance. The authors also report that most of the procedures that ranked high in earlier summaries remain high in rank, with mean validity estimates reduced by .10 to .20 points. That research is about interviews run by people, not AI interviewers, and it says nothing about any one question. The practical point is the structure: same core questions, same criteria. The follow-up is where you adapt to the answer. Keep the opening question fixed so candidates can be compared, then let the follow-up go where the answer leads. ## How to use these questions 1. Pick four competencies for the role from the five above. 2. Choose one opening question for each. 3. Keep the follow-up beside it and ask it only after you hear the first answer. 4. Write down what a strong answer contains before the first candidate joins. 5. Score the evidence after the interview, with the notes or transcript open. For a junior role, look for sound reasoning and the ability to learn. For a senior role, expect named trade-offs, past mistakes and examples of changing a decision. Do not turn trivia into a harder interview. The questions and follow-ups below are our own, written for this guide. ## Statistics and experiments | Opening question | Follow-up | A strong answer | | --- | --- | --- | | 1. An A/B test shows a 2% lift in conversion with a p-value of 0.04. Do you ship? | How many other metrics did you look at, and does that change your answer? | Asks about the planned hypothesis, sample size and run time. Raises multiple comparisons, novelty effects and the gap between statistical and practical significance. | | 2. Explain a p-value to a product manager. | What would a confidence interval tell them that the p-value does not? | Uses plain words, avoids "the probability the hypothesis is true", and says an interval shows size and uncertainty. | | 3. A feature rolled out by region, so you cannot randomise. How do you estimate its effect? | What would make you distrust that estimate? | Names a method such as difference-in-differences or matching, states the assumption it needs and says how to check it. | | 4. Your two experiment groups already look different before treatment. What now? | How would you tell a randomisation bug from bad luck? | Checks the assignment logic and the split ratio, runs balance checks and does not quietly re-weight and carry on. | A weak answer to this group is a rule recited with no judgement: "p below 0.05 means ship." Ask what they would do if the lift were real but tiny, and listen for whether the cost of acting enters the answer. ## Modelling | Opening question | Follow-up | A strong answer | | --- | --- | --- | | 5. Walk me through a model you took to production. | Which choice would you change, and what did it cost? | Separates their own work from the team's, names a constraint and a real regret, and says how it was monitored. | | 6. A fraud classifier has 98% accuracy. Is it good? | What would you measure instead, and who sets the threshold? | Notices class imbalance, asks about precision and recall and the cost of each error, and says the threshold is a business decision. | | 7. The model scored well offline and badly after launch. Where do you look? | What would you have logged beforehand to find this faster? | Considers leakage, differences between training and serving data, drift and changed user behaviour, in a sensible order. | | 8. When would you pick a simpler model than the best validation score? | Tell me about a time you did. | Weighs interpretability, latency, upkeep and stability against a small gain, and gives a real case. | Listen for ownership. "We built a gradient boosting model" tells you nothing. "I chose it over the logistic baseline because the lift held up on the last two months of data" tells you what they decided. ## Data handling and SQL | Opening question | Follow-up | A strong answer | | --- | --- | --- | | 9. Revenue per user dropped last week on the dashboard. What do you check first? | The raw table looks fine. What next? | Checks definitions, pipeline changes, duplicates and late-arriving data before touching a model. | | 10. Write, or talk through, a query for each user's second purchase date. | What changes if a user has two purchases in the same second? | Uses a window function or ranking, handles ties and NULLs, and states the table's grain. | | 11. A join doubled your row count. Why? | How would you catch that automatically next time? | Spots a many-to-many key, then adds a uniqueness or row-count check. | | 12. 30% of values are missing in a key feature. What do you do? | When is dropping those rows defensible? | Asks why the values are missing before imputing, and runs a sensitivity check. | For this group, a short exercise beats talk. Ask for the query on a shared screen or as a small task, then use the follow-up to discuss it. A candidate who wrote working SQL but cannot say what happens on a tie has memorised a pattern. ## Metrics and product sense | Opening question | Follow-up | A strong answer | | --- | --- | --- | | 13. How would you tell whether a new search ranking is better? | Which metric could improve while the experience gets worse? | Pairs a main metric with guardrails, and separates a proxy from the outcome it stands for. | | 14. Pick one metric for the health of a subscription app. | How could a team game it? | Names a leading and a lagging measure and the trade-off between them. | | 15. A stakeholder wants a forecast by Friday and you have six weeks of data. What do you say? | What do you deliver instead? | Says what the data can and cannot support and offers a range with stated assumptions. | | 16. Tell me about an analysis that changed a decision. | What did the decision-maker do differently? | Links the work to a specific action and says how they knew it mattered. | ## Communication and ownership | Opening question | Follow-up | A strong answer | | --- | --- | --- | | 17. Explain the main result of your last project to someone who has not seen the data. | What would you leave out, and why? | Gives the answer in one sentence first, then the evidence, with no jargon. | | 18. Tell me about a result you reported that turned out to be wrong. | How did you find out, and who did you tell? | Owns the error, says when and how they told people, and names what changed in their process. | | 19. A senior colleague insists on a method you think is wrong. What do you do? | What evidence would change your mind, and theirs? | Proposes a test or a comparison, represents the other view fairly and is open to being wrong. | | 20. How do you check your own work before you share it? | Which check has caught the most mistakes? | Names concrete checks: a baseline, re-running from raw data, a second pair of eyes, a sanity range. | Question 18 is worth protecting when time is short. Everyone has a mistake to report, and how they tell the story shows whether they check their work. ## What does a full follow-up chain look like? This is an example we wrote to show the shape of a chain. It is not a transcript of a real interview. It starts from question 1. > **Interviewer:** An A/B test shows a 2% lift in conversion, p equals 0.04. Do you ship? > > **Candidate:** Yes, it's significant, so I'd ship it. > > **Interviewer:** How many metrics did the team look at in that test? > > **Candidate:** Probably around ten. Conversion was the one that moved. > > **Interviewer:** So what does a p-value of 0.04 mean when you looked at ten? > > **Candidate:** One of them being under 0.05 by chance is fairly likely. I'd want conversion to have been the planned primary metric, and I'd check whether the lift holds in a rerun or on a holdout before shipping. > > **Interviewer:** And if it was planned, and holds, what else could make it a bad ship? > > **Candidate:** Novelty. A new layout can lift clicks for a week. I'd look at the lift by day, and check retention and refunds as guardrails. The first answer is the one a list gives you. The third and fourth answers show who has run experiments. The three follow-ups take a minute or so. Follow-ups on a candidate's own answer are also a quiet check that the answer is theirs; our guide to [AI interview cheating detection](/blog/ai-interview-cheating-detection) goes further. ## Strong and weak answers side by side Another example written by us, for question 6, the 98% accuracy question. | | Weak answer | Strong answer | | --- | --- | --- | | First answer | "98% is very high, so it's a good model." | "It depends on the base rate. If 98% of transactions are legitimate, predicting 'not fraud' every time scores 98%." | | After "What would you measure instead?" | "F1 score." | "Precision and recall at a threshold set by cost. Missing fraud and blocking a good customer cost different amounts, so I'd ask the business for both." | | After "Who sets the threshold?" | "The data science team." | "We propose one with the trade-off curve; the fraud and customer teams own the decision." | The weak candidate is not wrong about F1. They stop when the question moves from a term to a decision. ## Can you ask these questions of every applicant? A live screen for every applicant is a real cost in interviewer time when a role draws many applications. Our guide to [screening when there are too many applicants](/blog/how-to-screen-candidates) covers the options. One option is to let an AI interviewer run the first conversation (see [what an AI interviewer is](/blog/what-is-an-ai-interviewer) if the term is new, or [AI interview software compared](/blog/best-ai-interview-software)). Tracy AI Interviewer runs [two-way video interviews](/blog/one-way-vs-two-way-video-interviews). Each applicant starts their own interview when they click the link, at any hour, in a browser with nothing to install, so a pool of hundreds is interviewed without a calendar. Three things matter for this role: - **It reads the CV.** With CV check on, Tracy asks 3 to 5 questions tailored to the candidate's CV and the role's must-have requirements, on top of your questions. A claimed production model or experiment becomes a question about that model or experiment. - **It follows up and waits.** Follow-ups build on what the candidate said, and Tracy waits for the end of a thought, with longer pauses allowed on hard technical questions, so a candidate working through a trade-off is not cut off. - **It can use your questions.** Questions are generated from the job description and can be edited or replaced, so the 20 above fit as written. Interview style runs from quick to deep and from technical to behavioural. What your team gets back is the recording of both sides with a transcript anchored to it by timestamp, a written assessment per question with the reasoning behind it, and integrity signals (where the candidate looked, response timing, tab switches, leaving fullscreen, an external screen) with the evidence, for a person to review. A signal is not a verdict. People are told about the recording before it starts, and Tracy does not reject anyone: a person reviews the evidence and decides. Our guide to [AI interview reports](/blog/ai-interview-reports) shows what a reviewer should expect to see. Tracy also connects through an API and an MCP server, so interviews can be created and read from the tools a team already works in, and enterprise teams can ask for a customised setup. For published prices across tools, see [AI interview cost](/blog/ai-interview-cost). ## When to pick a test or a person instead An AI interviewer is not the right tool for every part of a data science process. - **To see someone write and run code against real data**, use a take-home or live exercise. The interview is for the conversation about it. For software roles, see [AI technical interviews](/blog/ai-technical-interviews) and the [software engineer question bank](/blog/software-engineer-interview-questions) for how to pair a task with follow-ups. - **For a final round on a senior role**, a person should be in the room. Judgement on an ambiguous business problem, and the fit with the team, are worth an hour of a senior colleague's time once the shortlist is small. - **For a candidate who has not agreed to be recorded**, use a live call. Tracy is not for talking to someone who has not agreed to be recorded. Tracy AI Interviewer fits the stage before those: the first conversation with everyone who applies, so the shortlist your seniors meet has already been asked the follow-ups. ## Common questions ### What should you ask in a data scientist interview? Cover five areas: statistics and experiments, modelling, data handling and SQL, metrics and product sense, and communication. Pick one opening question per area and keep the follow-up beside it, because the follow-up shows whether the candidate did the work. ### How many questions fit in a 30-minute data science screen? Four or five core questions with their follow-ups. That is our rule of thumb, not a researched figure. Depth is more useful than a long list, and a candidate who has to defend one answer shows more than one who recites five. ### What is the difference between a strong and a weak answer? A strong answer states its assumptions, names what could make it wrong and ties claims to work the candidate did. A weak answer is correct in general and goes quiet when asked about a specific project, number or trade-off. ### Should a data scientist interview include a coding test? Usually yes for the SQL and data-handling part, as a short exercise you can discuss afterwards. Use the interview itself to ask why the candidate chose an approach and what would change it. ### Can an AI interviewer ask data scientist interview questions? Yes. Tracy AI Interviewer can ask your questions, read the CV and tailor questions to it, follow up on what the candidate said and give your team the recording, a timestamped transcript and a written assessment per question. A person reviews the evidence and decides. --- **[Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=data-scientist-interview-questions&utm_content=try-free)** and run these questions on your own applicants, or [book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=data-scientist-interview-questions&utm_content=demo). ## Sources All sources retrieved 3 Nov 2026. - Reddit, [Senior Data Scientist, Ads](https://job-boards.greenhouse.io/reddit/jobs/8104403), posting updated 2 Nov 2026: skills, SQL, communication responsibilities. - Airbnb, [Senior Data Scientist, Guest & Host](https://careers.airbnb.com/positions/7662244), posting updated 28 Oct 2026: experimentation, causal analysis, communication. - DataAnnotation, [home page role listings](https://www.dataannotation.tech/): Data Scientist role and hourly rate as listed by the platform. - Sackett, Zhang, Berry and Lievens, Journal of Applied Psychology, [Revisiting meta-analytic estimates of validity in personnel selection](https://doi.org/10.1037/apl0000994), 2022: structured interviews as the top-ranked selection procedure. - Tracy AI Interviewer, [product page](https://tracyinterview.ai/product): two-way video interview, questions from the job description or your own, CV check, follow-ups, recording notice, recording and transcript timed to it, an assessment per question, integrity signals, API and MCP. - Tracy AI Interviewer, [home page](https://tracyinterview.ai/): a person reviews the evidence and decides; integrity signals with their evidence. - Tracy AI Interviewer, [pricing](https://tracyinterview.ai/pricing): API and MCP connections and a customised setup for Enterprise. - Tracy AI Interviewer, [Connect page](https://tracyinterview.ai/connect): MCP server and API connection to the tools a team already uses. # Interview Questions for Founders Applying to Your Programme > 16 questions to ask founders in an accelerator or fund first call, with the follow-up a partner would ask, drawn from the public Y Combinator and Techstars interview guides. - URL: https://tracyinterview.ai/blog/founder-interview-questions - Published: 2026-11-06 - Author: Tracy team Fact-checked on 3 Nov 2026 against the public Y Combinator and Techstars interview guides listed in Sources. > **In short** > - **Tracy AI Interviewer can give every founder a first call the moment they click the link,** with follow-ups that depend on each answer, so partners read the recording and transcript before choosing whom to meet. > - The 16 questions come with the follow-up a partner would ask. Pick five or six (our rule of thumb, not a published rule) and keep the follow-up beside each. A form tells you how well a founder writes about the company. It cannot ask the second question, and the second question is where the founder's story and ambition come out. This guide gives you questions for a first call with a founder, the follow-up to ask after each, and what a strong answer sounds like. It leans on two public sources: Y Combinator's [interview guide](https://www.ycombinator.com/interviews) and two Techstars articles, [one on the first interview](https://www.techstars.com/blog/founder-advice/how-to-prepare-for-your-first-techstars-interview) and [one on the application and interview process](https://www.techstars.com/blog/advice/inside-a-techstars-accelerator-the-application-and-interview-process). Where a question or a follow-up comes from one of them, the table says so, and "based on" marks a Techstars question we reworded. Follow-ups without a label and the strong-answer notes are ours. ## What must a first call with a founder find out? A first call has one job: to help you decide whom to spend a longer conversation on. A Techstars managing director describes the first interview as "not a gotcha interview", and says it is trying to understand the company "well enough to write a strong investment memo". That is a useful test for your own first call. When it ends, could you write four honest lines about this team? The two public guides agree on the ground to cover, even though the formats differ. | | Y Combinator | Techstars | | --- | --- | --- | | Format, as the guide describes it | 10-minute conversations over Zoom, with 2 to 3 partners who have read the application | First-interview advice, with answers that should mostly take under 90 seconds | | Opening question | "What is your company working on?" | Four areas to be ready for: customer and pain, speed and execution, market size, differentiation | | What it listens for | Users and metrics, honest discussion of obstacles, knowledge of existing products | A concrete customer, what was built and skipped, a bottoms-up market, what is hard to copy | Both ask the founder to be specific. YC says there is no time for formal presentations, that interview preparation beyond the basics is not useful and is often counter-productive, and that the best edge is progress between the application and the interview. Techstars asks for clear, concrete answers, and its application-process article tells founders to practise their introduction and rehearse answers, so the two differ on preparation. ## How to use these questions 1. Pick five or six (our rule of thumb), covering the four groups below. 2. Ask the opening question and wait for the full answer. 3. Ask the follow-up that fits what they said, not the one on the page if the answer has already covered it. 4. Write what you learned in the founder's words. Quotes beat adjectives. 5. Score nothing until the call is over. Keep the opening questions the same for every founder so you can compare. A 2022 re-analysis in the Journal of Applied Psychology (Sackett and colleagues) found structured interviews to be the top-ranked selection procedure for predicting job performance. The authors also report that most of the procedures that ranked high in earlier summaries remain high in rank, with mean validity estimates reduced by .10 to .20 points. That research is about hiring and interviews run by people, so read it as a case for structure, not as a result for founder selection. Let the follow-up be where the conversation adapts. ## Questions on the problem and the customer | Opening question | Follow-up | A strong answer sounds like | | --- | --- | --- | | 1. What is your company working on? (YC) | Say that again as if I were outside your field. | Two or three plain sentences with one detail only someone in the field would know. | | 2. Who is the customer, specifically, and what hurts? (based on Techstars) | Who did you talk to, and what did one of them say that surprised you? (based on Techstars) | Names a real person or segment, quotes them and says what they do today. | | 3. How are people solving this today, and what is wrong with that? (YC, Techstars) | Why has the best of those not fixed it? | A specific limit of the current tools. "Ours is easier to use" is not an answer; YC says founders should be able to explain how. | | 4. What have you learned since you started? (based on Techstars) | What did you change because of it? | A belief they held and dropped, with what made them drop it. | ## Questions on the team | Opening question | Follow-up | A strong answer sounds like | | --- | --- | --- | | 5. Tell me about each member of your team and how they contribute. (based on Techstars) | How long have you worked together before this? (based on Techstars) | Concrete shared work, not just a shared job title. | | 6. Tell us about a conflict within the team that you have resolved. (based on Techstars) | What would you do differently now? | A real disagreement, how it was settled and what each person owned in it. | | 7. If I gave you $200K in cash to spend on hiring, who would you hire? (based on Techstars) | Why that role before the others? | Names a gap in the team honestly and has thought about the order. | | 8. How did you decide who is the CEO? (based on Techstars) | What happens if that turns out to be the wrong call? | Calm, specific, with a decision rule and no defensiveness. | Techstars says it wants to see "evidence of the team working well together". YC says it prefers that each founder answers at least one question, so you get to know all of them. On a call with two founders, direct a question at the quieter one. ## Questions on traction and speed | Opening question | Follow-up | A strong answer sounds like | | --- | --- | --- | | 9. What did you build first, and what did you skip? (based on Techstars) | What changed because you shipped it fast? (based on Techstars) | A deliberate cut, and a change in what they believed after real use. | | 10. Where do new users come from, and how fast is growth? (YC) | Which number here would you least like us to check? | Has metrics written down, gives the weak one without being pushed. YC says it may ask for verification of numbers stated in the interview. | | 11. What is your most important KPI this week? (based on Techstars) | What will you do if it does not move? | One metric, a reason it is the right one, and a next step. | | 12. Where are you seeing real pull? (based on Techstars) | How do you know it is pull and not your own effort? | Customers returning or asking, with an example. | ## Questions on obstacles and ambition | Opening question | Follow-up | A strong answer sounds like | | --- | --- | --- | | 13. What is the hardest part of the business you are building? (based on Techstars) | If this were to fail, what is the most likely reason? (based on Techstars) | A candid, specific risk. YC says it is "more convinced by candid discussion of difficulties than glib dismissal of them". | | 14. How big can this get, bottoms up? (based on Techstars) | Which assumption in that estimate worries you most? | Starts from customers and prices, not "1% of a huge market". | | 15. What is hard to copy? (based on Techstars) | What would a better-funded team need to catch up? (based on Techstars) | A real advantage such as data, access or speed, and honesty about its limit. | | 16. What do you want from us? (Techstars asks about the ideal outcome) | What have you done already to find out whether we can give it? | A specific ask tied to the programme, with signs they researched it. | Question 13 is the one to protect when time is short. A founder who can describe how the company might fail, in detail, has usually thought hard about how to make it work. ## What does a follow-up chain look like? This is an example we wrote, not a transcript of a real call. It starts from question 2. > **Interviewer:** Who is the customer, specifically, and what hurts? > > **Founder:** Mid-size logistics companies. Their dispatch process is manual and slow. > > **Interviewer:** Who did you talk to at one of them? > > **Founder:** A dispatch manager at a regional carrier. She told me they plan routes in a spreadsheet each morning and rebuild it when a driver calls in sick. > > **Interviewer:** What does a bad morning cost them? > > **Founder:** She guessed about two hours of the manager's time and a late delivery or two. I haven't measured it myself yet. > > **Interviewer:** What would you need to see to believe that number? > > **Founder:** The same morning logged for a week at three carriers. We have one pilot starting that does it. The first answer is general. The second names a person. The fourth is the one to write down: the founder says what they have not measured and what would settle it. ## Strong and weak answers side by side Another example written by us, for question 13. | | Weak answer | Strong answer | | --- | --- | --- | | First answer | "Honestly, nothing is that hard. We have a great team." | "Getting carriers to change a habit. They have used spreadsheets for ten years." | | After "What is the most likely reason this fails?" | "I don't see it failing." | "That the first five pilots like it and none pay. We are testing price in the second month for that reason." | The weak founder may be just as capable. The call has not shown it, and a form never would. ## How do you run this call with every founder? Partner time limits how many founders get this call. Techstars says thousands of companies apply to it, and with an intake that large a programme cannot call everyone and has to judge the rest on the form. Tracy AI Interviewer can run the first conversation with everyone (see [what an AI interviewer is](/blog/what-is-an-ai-interviewer) if the term is new, or [AI interview software compared](/blog/best-ai-interview-software)). It runs [two-way video interviews](/blog/one-way-vs-two-way-video-interviews), and each founder starts their own when they click the link: any hour, a browser with nothing to install, no calendar. - **Every applicant, not a shortlist.** The first call stops being a privilege of the founders who looked best on the form. - **Follow-ups that wait.** Tracy asks follow-ups on what the founder actually said and waits for the end of a thought, so a founder thinking through a hard question is not cut off. - **Your questions.** Questions are generated from a description and can be edited or replaced, so the 16 above fit as written. If founders attach a CV, Tracy can read it and tailor questions to it when CV check is on. Your partners get the recording of both sides with a transcript anchored to it by timestamp, a written assessment per question with the reasoning behind it, and integrity signals with the evidence, for a person to review. A signal is not a verdict. People are told about the recording before it starts. Tracy does not reject anyone: a person reviews the evidence and decides. That fits the first call's job: pick whom to meet. See our [accelerator interview guide](/blog/accelerator-interview) for the programme side, [AI interviews for applications](/blog/ai-interviews-for-applications) for how a programme gives every applicant a first interview, and [AI interview reports](/blog/ai-interview-reports) for what a reviewer should expect to see. Tracy also connects through an API and an MCP server, so a programme can create and read interviews from the tools it already uses, and enterprise teams can ask for a customised setup. [AI interview cost](/blog/ai-interview-cost) compares published prices across tools. ## When to pick a partner call instead - **When the relationship is the product.** YC points out that the people in the interview "are likely the same people you'll work closely with should you be accepted". Chemistry and trust are judged across a table, by the people who will sit there. - **When the intake is small.** If you see thirty founders, call all thirty yourself. The time saved by an AI first round matters when the intake is in the thousands. Our guide to [screening when there are too many applicants](/blog/how-to-screen-candidates) covers the options when volume is high. - **For a founder who has not agreed to be recorded.** Tracy is not for talking to someone who has not agreed to be recorded. Tracy fits the stage before those calls: the first conversation with every applicant, so partner hours go to the founders who were worth hearing. It does not replace the call that decides. ## Common questions ### What should you ask founders in a first call? Ask about the customer and the problem, the team, what has been built and learned, and the ambition. Keep one follow-up ready for each, such as who they spoke to or what would make the company fail. ### How long should a first founder interview be? Short. Y Combinator describes its interviews as 10-minute conversations, and a Techstars managing director advises that most answers should take under 90 seconds. Five or six questions with follow-ups is a sensible size for a first call, though that is our judgement, not a published rule. ### How do you tell a strong founder answer from a rehearsed one? A strong answer is specific: a named customer, a number the founder can source, a decision and what it cost. A rehearsed answer stays smooth and general, and slows down when you ask for the detail. ### Should every founder who applies get a first call? Ideally, because a form cannot ask a second question. When partner hours are limited, an AI interviewer can run the first conversation with everyone, and partners choose whom to meet. ### Can an AI interviewer ask founder interview questions? Yes. Tracy AI Interviewer can give every founder who applies a first call, ask your own questions, follow up on what the founder said and give your team the recording, a timestamped transcript and a written assessment per question. Partners still decide whom to meet and fund. --- **[Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=founder-interview-questions&utm_content=try-free)** with the questions above, or [book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=founder-interview-questions&utm_content=demo). ## Sources All sources retrieved 3 Nov 2026. - Y Combinator, [YC Interview Guide](https://www.ycombinator.com/interviews): interview format, first question, metrics, obstacles, existing products, verification of numbers. - Techstars, [How To Prepare for Your First Techstars Interview](https://www.techstars.com/blog/founder-advice/how-to-prepare-for-your-first-techstars-interview), 22 Jan 2026: four areas, answer length, not a gotcha interview. - Techstars, [Inside a Techstars Accelerator: The Application & Interview Process](https://www.techstars.com/blog/advice/inside-a-techstars-accelerator-the-application-and-interview-process), 12 Jan 2026: sample questions on team, hiring, failure, market and KPIs. - Sackett, Zhang, Berry and Lievens, Journal of Applied Psychology, [Revisiting meta-analytic estimates of validity in personnel selection](https://doi.org/10.1037/apl0000994), 2022: structured interviews as the top-ranked selection procedure. - Tracy AI Interviewer, [product page](https://tracyinterview.ai/product): two-way video interview, questions from a description or your own, CV check, follow-ups, recording notice, recording and transcript timed to it, an assessment per question, integrity signals, API and MCP. - Tracy AI Interviewer, [home page](https://tracyinterview.ai/): a person reviews the evidence and decides; integrity signals with their evidence. - Tracy AI Interviewer, [pricing](https://tracyinterview.ai/pricing): API and MCP connections and a customised setup for Enterprise. - Tracy AI Interviewer, [Connect page](https://tracyinterview.ai/connect): MCP server and API connection to the tools a team already uses. # 25 Software Engineer Interview Questions With Follow-Ups > Use 25 software engineer interview questions for coding, debugging, system design and teamwork, with follow-ups and strong-answer signals. - URL: https://tracyinterview.ai/blog/software-engineer-interview-questions - Published: 2026-11-03 - Author: Tracy team ![Cover graphic introducing software engineer interview questions.](/blog/images/software-engineer-interview-questions-cover.png) The best software engineer interview questions do not end at the first correct sentence. They give the candidate something specific to reason about, then use a follow-up to test ownership, trade-offs or edge cases. Use this list as a question bank, not a script to finish in one sitting. A 30-minute first round usually fits three or four core questions with follow-ups. Pick the sections that match the job and write down what a strong answer should contain before the first candidate joins. ## How to use these questions 1. Pick four competencies the role needs, such as debugging, API design, reliability and collaboration. 2. Choose one opening question for each. 3. Keep the follow-up beside it. Ask it only after hearing the first answer. 4. Write observable signals, not personality labels. 5. Score the evidence after the interview, with the transcript or notes open. For a junior role, look for a sound approach and the ability to learn. For a staff role, expect clearer trade-offs, wider system effects and examples of influencing other teams. Do not use seniority to turn trivia into a harder interview. ![Pick role skills, ask a question and probe the engineer’s answer.](/blog/images/software-engineer-interview-questions-process.png) ## Project and ownership questions | # | Opening question | Follow-up | Strong-answer signals | | --- | --- | --- | --- | | 1 | Tell me about a system you shipped that you are proud of. | Which decision was yours, and what changed because of it? | Separates personal work from team work, names a constraint and gives a measurable result. | | 2 | Describe a technical decision you would make differently now. | What information did you have then, and what did you learn later? | Judges the decision in its original context, then names the new evidence. | | 3 | Tell me about the hardest production incident you handled. | What was the first signal, and what did you change to prevent a repeat? | Gives a timeline, distinguishes mitigation from root cause and names the follow-up work. | | 4 | Walk me through a feature from request to release. | Where did the original requirement change, and how did you respond? | Covers discovery, design, implementation, testing and rollout without pretending the plan stayed fixed. | | 5 | What piece of code have you removed or simplified? | How did you prove it was safe to remove? | Values less code, checks callers and telemetry, and explains the rollback plan. | Weak answers stay at team level: "we improved performance" or "we migrated to microservices." Ask what the person changed, why that option won and how the result was measured. A candidate can be collaborative and still state their own contribution clearly. ## Coding and data-structure questions These can be discussed at a whiteboard, in a shared editor or through a small live task. Let the candidate choose a familiar language unless the job genuinely depends on one. | # | Opening question | Follow-up | Strong-answer signals | | --- | --- | --- | --- | | 6 | Given a stream of events, return the ten most frequent values in the last hour. | What changes at 50,000 events per second? | Starts with a correct structure, then bounds memory with buckets or another window strategy. | | 7 | Merge overlapping time intervals. | How do you handle intervals that touch at the boundary? | States whether endpoints are inclusive, sorts deliberately and tests empty and single-item inputs. | | 8 | Design a least-recently-used cache. | What happens when two requests update the same key? | Combines a map with ordered access, then discusses locking, atomic operations or ownership. | | 9 | Find whether a directed graph contains a cycle. | How would your answer change for a graph too large for one machine? | Explains visited and active states, complexity and the limit of the local algorithm. | | 10 | Process a file that does not fit in memory and return duplicate records. | What if exact answers are required? | Distinguishes hashing, sorting and approximate methods, and accounts for disk or partitioning. | A strong coding answer goes past naming a familiar algorithm. Look for assumptions, test cases, complexity and a response when the constraint moves. If the candidate writes code, ask them to run or reason through at least one edge case before you do. ## Debugging and testing questions | # | Opening question | Follow-up | Strong-answer signals | | --- | --- | --- | --- | | 11 | An API's 99th-percentile latency doubled after a release. What do you do first? | The median did not move. What does that suggest? | Checks the change, narrows by endpoint and dependency, and reasons about a tail affecting a subset. | | 12 | A test passes locally and fails in CI once every twenty runs. How do you investigate? | How would you make the failure reproducible? | Considers shared state, order, timing and environment, then controls seeds, clocks or concurrency. | | 13 | Users report duplicate payments, but logs show one request. Where do you look? | How would you prevent another charge while investigating? | Checks retries and downstream processing, then adds idempotency or a temporary guard. | | 14 | Memory grows slowly for several days before a service restarts. How do you isolate the leak? | Which measurement would distinguish a leak from a larger working set? | Uses heap profiles or allocation data over time and compares retained objects after steady load. | | 15 | A daylight-saving change broke scheduled jobs. How would you test the fix? | What should be stored as local time, and what should be stored as an instant? | Separates business-local schedules from UTC instants and tests ambiguous and missing local times. | The useful signal is the order of investigation. Strong debuggers protect users, establish what changed, gather evidence and narrow the system. They do not jump straight to rewriting the suspected component. ## System-design questions Change the scale and reliability target to fit the level. The candidate should be free to ask questions before drawing boxes. | # | Opening question | Follow-up | Strong-answer signals | | --- | --- | --- | --- | | 16 | Design a notification service for email, SMS and push. | How do you prevent a retry from sending the same message twice? | Separates channel workers, persistence and delivery state, and uses idempotency at the right boundary. | | 17 | Design a rate limiter for a public API. | Where does state live across regions? | Chooses an algorithm for the requirement, then discusses consistency, latency and failure behaviour. | | 18 | Design file upload for large videos. | How does an interrupted upload resume safely? | Uses multipart upload, content validation and durable upload state, with security checks. | | 19 | Design search for a product catalogue. | How quickly must price and stock changes appear? | Clarifies ranking and freshness, separates the search index from the source of truth and plans reindexing. | | 20 | Design an audit log for sensitive account changes. | Who can alter or delete the log? | Makes events append-only, protects access, records actors and plans retention and verification. | Do not grade a diagram by how many fashionable components it contains. Grade whether the design meets the stated traffic, consistency, privacy and failure requirements, and whether the candidate can explain what they would build first. ## Collaboration and judgement questions | # | Opening question | Follow-up | Strong-answer signals | | --- | --- | --- | --- | | 21 | Tell me about a technical disagreement with another engineer. | What evidence changed the discussion? | Represents the other view fairly, uses a test or constraint and explains the eventual decision. | | 22 | A product manager wants a deadline you think is unsafe. What do you do? | What smaller version could you offer? | Makes risk concrete, proposes scope choices and shares ownership of the outcome. | | 23 | A junior engineer submits code that works but is hard to maintain. How do you review it? | What feedback belongs in the pull request and what belongs in a conversation? | Prioritises correctness and learning, gives actionable examples and avoids turning preference into law. | | 24 | You inherit a service nobody understands. What is your first month? | Which change would you deliberately postpone? | Maps users, owners, alerts and failure modes before a rewrite, then creates a safe improvement sequence. | | 25 | When should an engineer accept technical debt? | Name debt you accepted and how it was tracked. | Connects the shortcut to a real deadline or experiment and states the repayment trigger. | ## A simple scoring guide Score each competency from 1 to 4 with a written reason. | Score | Evidence | | --- | --- | | 1 | No relevant example or the approach would not work. | | 2 | Partly workable answer, but important assumptions or risks are missing. | | 3 | Sound answer with explicit trade-offs and a relevant example. | | 4 | Sound answer that also anticipates failure modes, measures the result and adapts when the constraint changes. | Avoid one overall "technical strength" number until reviewers have scored the separate competencies. A candidate can be strong at debugging and weaker at distributed design. That profile is more useful than an average stripped of context. For every score, keep one sentence of evidence. "3 because they proposed idempotency keys at the payment boundary and explained retry behaviour" is reviewable. "Good systems thinker" is not. ## Running these questions with an AI interviewer A two-way AI interviewer can ask the same opening questions and branch into the follow-ups when the candidate's answer creates a useful path. Tracy AI Interviewer's product page says it can introduce coding tasks, system design, debugging and code walk-through work, then probe the trade-offs behind an answer. That does not remove engineering review. A reviewer should inspect the candidate's words, code, transcript or recording behind a strong or weak score. The interviewer collects evidence; the hiring team decides. CodeSignal and HackerRank also offer technical assessments, and both describe AI interview products. A standard assessment is useful when you need a validated task at high volume. An interview is useful when you need to hear why the candidate chose an approach and how it changes under a follow-up. Some teams use both in sequence. See [AI technical interviews](/blog/ai-technical-interviews?utm_source=blog&utm_medium=seo&utm_campaign=software-engineer-interview-questions&utm_content=internal-link) for a complete task and follow-up chain, and [AI interview reports](/blog/ai-interview-reports?utm_source=blog&utm_medium=seo&utm_campaign=software-engineer-interview-questions&utm_content=internal-link) for the reviewer checklist. ![A checklist of the main actions for software engineer interview questions.](/blog/images/software-engineer-interview-questions-takeaways.png) ## Common questions ### What should I ask in a software engineer interview? Ask for one real project, one coding or debugging task, one design problem and one collaboration example. Follow up on the candidate's own trade-offs and contribution. ### How many questions fit in a 30-minute engineering screen? Usually three or four core questions with follow-ups. Depth is more useful than rushing through a long list. ### What makes a strong technical interview answer? A strong answer makes assumptions explicit, compares options, tests edge cases, admits uncertainty and ties claims to work the candidate personally did. ### Should every engineer get the same questions? Use the same core competencies and opening questions for comparable candidates, then allow follow-ups based on each answer. ### Can an AI interviewer ask software engineering questions? Yes. A two-way AI interviewer can present a coding, debugging or system-design task and ask about the candidate's solution. A person should review the evidence and decide. --- **[Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=software-engineer-interview-questions&utm_content=try-free)** or **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=software-engineer-interview-questions&utm_content=demo)** ## Sources All sources retrieved 2 Nov 2026. - Tracy AI Interviewer, [product page](https://tracyinterview.ai/): coding tasks, system design, debugging, code walk-through, adaptive follow-ups and report evidence. - CodeSignal, [technical assessments](https://codesignal.com/technical-assessments): assessment format and validation positioning. - HackerRank, [Screen](https://www.hackerrank.com/products/screen/): technical screening and question-library scope. - HackerRank, [Chakra](https://www.hackerrank.com/chakra): AI interview follow-ups on candidate work and score explanations. # AI Trainer Interview Questions: 25 Questions and the Follow-Ups That Matter > 25 AI trainer interview questions for domain experts in coding, maths, medicine, law and finance, with the follow-up to ask and strong vs weak answers. - URL: https://tracyinterview.ai/blog/ai-trainer-interview-questions - Published: 2026-11-02 - Updated: 2026-11-11 - Author: Tracy team > **In short** > - **These 25 AI trainer interview questions test domain expertise beyond a polished first answer.** Each includes a follow-up and signals to look for across coding, maths, medicine, law and finance. > - Tracy AI Interviewer can run an expert-level first round with every applicant, adapt the follow-up to each answer and return evidence reviewers can check. Pick four or five questions for a role and keep the follow-up even if you shorten the interview. **[Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=ai-trainer-interview-questions&utm_content=try-free)** or **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=ai-trainer-interview-questions&utm_content=demo)** **Contents** - [What should AI trainer interview questions test?](#what-should-ai-trainer-interview-questions-test) - [Why does the follow-up tell a real expert from a polished one?](#why-does-the-follow-up-tell-a-real-expert-from-a-polished-one) - [How can you ask these follow-ups for every applicant?](#how-can-you-ask-these-follow-ups-for-every-applicant) - [AI trainer interview questions and answers: domain depth](#ai-trainer-interview-questions-and-answers-domain-depth) - [Reasoning questions: can they think out loud?](#reasoning-questions-can-they-think-out-loud) - [Rubric questions: can they write a task a grader can use?](#rubric-questions-can-they-write-a-task-a-grader-can-use) - [Model error questions: will they catch a confident wrong answer?](#model-error-questions-will-they-catch-a-confident-wrong-answer) - [Integrity questions: is this the person on the profile?](#integrity-questions-is-this-the-person-on-the-profile) - [What does a full follow-up chain look like?](#what-does-a-full-follow-up-chain-look-like) - [Questions](#questions) - [Sources](#sources) An AI trainer, or domain expert, writes hard questions in their own field, grades model answers against a rubric, and catches answers that sound right and are wrong. So these AI trainer interview questions test the field first and AI second. Many lists of AI trainer interview questions ask how the candidate feels about AI, which tells you little about whether they can do the work. Use them as a bank. Pick four or five per interview, one from each group, and always ask the follow-up. ## What should AI trainer interview questions test? Public job pages describe the same tasks. Outlier lists "Write a Challenging Prompt," "Create Grading Rubrics," and "Rate and Rank Answers," and asks experts to "Think of a difficult question in your field of study." DataAnnotation describes coding work as "debugging code, evaluating implementations, reasoning about edge cases." The pay shows what a bad hire costs. DataAnnotation lists finance, legal and medical roles at "up to $125+/hr." At that rate, an expert who can only recite definitions costs hours of rework and puts the lab's data quality at risk. So test five things: - Domain depth: they know the field past the textbook. - Reasoning: they can show how they reached an answer and where it could break. - Rubric judgement: they can write a prompt with one right answer, and a rubric two graders would score the same way. - Catching model errors: they spot the fluent, confident mistake. - Integrity: the person answering is the person on the profile, doing the work themselves. ## Why does the follow-up tell a real expert from a polished one? Anyone applying has read the same question lists you have, and an AI assistant can draft a fine first answer in seconds. A follow-up built on what the candidate just said can't be rehearsed, so it shows how far their knowledge goes. "Which antibiotics worry you most?" only makes sense after they've mentioned antibiotics. At the second and third question, a real expert brings a case they saw, an exception to the rule or the number that changes the decision. If you're short of time, drop a question and keep the follow-up. ![Diagram of question 3 from this list: a patient on warfarin starts antibiotics. The follow-up asks which antibiotics worry them most. The strong answer names a rising INR, metronidazole and co-trimoxazole, and when to recheck. The weak answer says "bleeding risk, so check with a pharmacist."](/blog/images/ai-trainer-interview-questions-strong-vs-weak.svg) *Question 3 with its follow-up: the strong answer is specific enough for a reviewer to check.* ## How can you ask these follow-ups for every applicant? These questions need an interviewer who knows the field, and expert interviewers run out first. A finance interviewer can't vet a radiologist, and when several lab projects land in the same week, there aren't enough of them to go round. Tracy AI Interviewer runs this first round as a natural two-way video interview with role-specific questions, so your reviewers see real depth before they spend an hour on anyone. It "follows up on vague answers, digs into the interesting ones", which is where a shallow expert runs out of material. For engineers it can introduce a coding task and probe the trade-offs behind an answer, so you see how they work as well as what they say. Our page on [AI technical interviews](/blog/ai-technical-interviews?utm_source=blog&utm_medium=seo&utm_campaign=ai-trainer-interview-questions&utm_content=internal-link) walks through one. Your questions are drafted for you, and you can add the questions that matter, such as the 25 below, so the interview matches your projects. Each applicant opens a link in their browser and joins with no sign-up, whenever it suits them, so a new lab project can be staffed without adding interviewers. Candidates rate Tracy 4.7 out of 5 on average, so experts stay in your funnel. Pricing is $1 per credit; one credit covers up to 20 minutes, and "a session with no answers is never charged". Enterprise pricing is custom, based on your volume, so cost stays in line at thousands of interviews. ![A Tracy result report for a clinical reviewer. Three fields, such as drug interactions and reasoning on a hard case, are marked Stated, each with the expert's own words as evidence. The recording and timestamped transcript sit above them.](/blog/images/lib-result-traced-clinical-expert.png) *Every point in a Tracy result is the expert's own words, next to the video and transcript, and your team decides (illustration with sample data).* To compare tools, see our list of the [best AI interview software for hiring teams](/blog/best-ai-interview-software?utm_source=blog&utm_medium=seo&utm_campaign=ai-trainer-interview-questions&utm_content=internal-link) or [Tracy vs HireVue](/compare/tracy-vs-hirevue?utm_source=blog&utm_medium=seo&utm_campaign=ai-trainer-interview-questions&utm_content=internal-link) on follow-up depth. ## AI trainer interview questions and answers: domain depth Do they know the field past the definition? Swap in your own field where you need to. | # | Question | Follow-up to ask | Strong answer | Weak answer | |---|---|---|---|---| | 1 | Coding: walk me through a bug you shipped to production and how you found it. | What would have caught it earlier, and why wasn't that in place? | Names the bug (a race condition in a cache refresh, say), the signal that exposed it and the test added after. Admits the gap. | "I use logging and debugging tools." No specific bug, nothing changed. | | 2 | Maths: why does a proof by induction need a base case? | Find the flaw in the "proof" that all horses are the same colour. | Locates the step from one horse to two, where the two groups don't overlap. | "The induction step is wrong," without saying where. | | 3 | Medicine: a patient on warfarin starts a course of antibiotics. What do you watch for? | Which antibiotics worry you most, and what would you change in monitoring? | Expects the INR to rise and bleeding risk with it; names metronidazole and co-trimoxazole as the highest-risk antibiotics; says when to recheck the INR. | "Bleeding risk, so check with a pharmacist." No mechanism, no names, no plan. | | 4 | Law: what's the difference between a condition and a warranty in a contract? | A supplier delivers late. What can the buyer do, and what would you check in the contract first? | Breach of a condition can allow termination plus damages; a warranty gives damages only. Asks which law governs and whether time is "of the essence". | Recites the two definitions. Never asks about jurisdiction or wording. | | 5 | Finance: how would you value a company with negative earnings? | Your DCF says $40 a share and comparables say $15. What do you check first? | Revenue multiples or a DCF with a path to margins. Checks how much sits in terminal value and whether the peers compare. | "Use a P/E ratio," which fails with losses, or picks one number without reconciling. | | 6 | Any field: what's a question in your field that most people, and most AI models, get wrong? | What's the right answer, and how would a grader spot the wrong one? | A narrow question with one checkable answer, plus the specific mistake. | A broad topic ("ethics in medicine") or trivia. | On question 3: in a study summarised by NHS UK Medicines Information, the antibiotics that interfere with warfarin metabolism, metronidazole and trimethoprim-sulfamethoxazole (co-trimoxazole), "posed the greatest risk for an INR of 5.0 or more". ## Reasoning questions: can they think out loud? Expert data is valuable because it shows the reasoning, so the expert has to show theirs. | # | Question | Follow-up to ask | Strong answer | Weak answer | |---|---|---|---|---| | 7 | When did you last change your mind on a question in your field? What changed it? | What would change it back? | Names the evidence and the condition that would reverse it. | "I'm always open to new ideas." No example. | | 8 | Two experts in your field disagree. How do you decide who is right? | Give me a case where the textbook answer was wrong in practice. | Goes to the primary source and each side's assumptions. Has a real case. | "I'd go with the more senior person." | | 9 | Solve this problem from your field and state each assumption as you go. | Drop your second assumption. Does the answer hold? | States assumptions unprompted and knows which one the answer rests on. | One answer, then can't rework it. | | 10 | Before answering "Is this clause enforceable?" (or "Is this drug safe in pregnancy?"), what do you need to know? | The user won't tell you. What do you write? | Lists the missing facts, then writes a conditional answer. | Picks a default silently and answers as if certain. | | 11 | How confident are you in that answer, as a percentage? | What would get you to 95%? | A number below 100 and a concrete check that would raise it. | 100% on everything, or no number. | ## Rubric questions: can they write a task a grader can use? Most of the paid work is writing. These show whether the candidate writes for a grader or for a reader who already agrees with them. | # | Question | Follow-up to ask | Strong answer | Weak answer | |---|---|---|---|---| | 12 | Write a prompt in your field that a current model gets wrong. | Why does it fail, and what must the right answer include? | One verifiable answer; explains the failure (the model mixes up two similar rules, say). | A trick question, an ambiguous one, or trivia. | | 13 | Turn this model answer into a rubric of five checkable items. | Which item would two graders score differently, and how would you reword it? | Yes-or-no items. Rewrites "clear explanation" into something like "states the maximum dose in mg/kg". | "Accurate, helpful, well written, safe, concise." | | 14 | Rank these two responses and justify it in three sentences. | The shorter one has a small error and the longer one is padded. Does length change your ranking? | Correctness first: names the error and its severity, then the padding. | Prefers the longer, friendlier answer and misses the error. | | 15 | Rewrite this paragraph for a reader with no background, keeping the exact claim. | Which word did you almost cut, and why did you keep it? | Keeps the qualifier that carries the meaning ("in adults", "before tax"). | Simplifies the caveat away and changes the claim. | | 16 | How would you write the ideal answer to a question with no single right answer? | What would make an answer fail anyway? | Defines must-haves and disqualifiers, then the acceptable range. | "Anything reasonable passes." | ## Model error questions: will they catch a confident wrong answer? Show a real model answer from your project with one planted error. | # | Question | Follow-up to ask | Strong answer | Weak answer | |---|---|---|---|---| | 17 | Coding: here's a model's fix for a failing test. Does it fix the bug? | What input would break it? | Names an edge case: an empty list, an off-by-one at the boundary, a time zone change. | Traces the happy path and says yes. | | 18 | Law or medicine: this answer cites a case (or a guideline). How would you check it? | The source exists but says something else. How do you mark it? | Checks the primary source; treats a misquoted source as worse than no citation. | Trusts it because it looks real. | | 19 | Maths: this solution has the right final answer. Is it correct? | Which step would you mark down, and by how much? | Finds the invalid step and grades the method too. | "The answer matches, so it's correct." | | 20 | What's the most harmful mistake a model can make in your field? | Show me how it would look inside a fluent, confident answer. | Specific: mg written where mcg is meant, a sign error in a cash flow, the right rule from the wrong jurisdiction. | "Hallucination," with no example. | | 21 | This answer is correct. Would you let it ship? | What's missing? | Spots the missing safety caveat, the outdated rule or the unstated assumption. | Correct means done. | ## Integrity questions: is this the person on the profile? A fake or shallow expert who passes screening costs a platform the lab's trust. Ask these in the same calm tone as the rest. | # | Question | Follow-up to ask | Strong answer | Weak answer | |---|---|---|---|---| | 22 | Walk me through the credential on your profile: where, when and what you worked on. | What was the hardest case or project in that time? | Details come easily and stay consistent. | Vague, or details shift between answers. | | 23 | Explain your last answer again, for a junior colleague. | Why did you pick that example? | Rephrases freely with a new example of their own. | Repeats the same sentences, or pace and vocabulary change sharply between answers. | | 24 | Which parts of this work would you use AI tools for, and which not? | A project bans AI help. How would you work on it? | Honest and specific; knows project rules differ. | Evasive, or claims never to use AI tools at all. | | 25 | What would you do if a task's instructions looked wrong? | And if the deadline is in an hour? | Flags it with evidence through the agreed channel, then follows the decision. | Quietly changes the task, or ignores it. | Tracy's interview integrity detection also flags the signals around each answer: tab changes, abnormal eye tracking, live face detection, average delay before an answer, external screens and other unusual patterns. Each one points the reviewer to a moment worth a second look, and your team decides what it means. Our guide to [how AI interview cheating detection works](/blog/ai-interview-cheating-detection?utm_source=blog&utm_medium=seo&utm_campaign=ai-trainer-interview-questions&utm_content=internal-link) covers each signal. ## What does a full follow-up chain look like? Example: an illustrative first round for a finance expert, with test data, not a real candidate. It starts from question 5. > **Tracy:** How would you value a company with negative earnings? > > **Candidate:** I'd use an EV to revenue multiple from comparable companies. > > **Tracy:** Which comparables would you pick for a mid-size B2B software company, and why those? > > **Candidate:** Listed SaaS companies of similar size. But growth matters more than size, so I'd filter on revenue growth first. > > **Tracy:** Say your peers grow at 40% a year and this company grows at 12%. What does that do to the multiple? > > **Candidate:** It should come down a lot. I'd regress multiple against growth across the peers and read off the value at 12%, then check it with a DCF where margins reach the peer median in year five. > > **Tracy:** Your DCF comes out at almost three times the multiple. What do you check first? > > **Candidate:** How much of the DCF is terminal value. If it's 80% or more, the gap is mostly my long-run growth assumption, and I'd say so in the write-up. The first answer is the textbook one. By the third question this candidate is working with the company's own numbers and naming the assumption that drives the result. A candidate who only knew the definition would usually stall at "which comparables". The reviewer gets a summary with the candidate's own words attached, and the AI explains each answer score (test data): | Field | What the candidate said | |---|---| | Valuation method | "I'd regress multiple against growth across the peers and read off the value at 12%." | | Checks own work | "If it's 80% or more, the gap is mostly my long-run growth assumption, and I'd say so in the write-up." | With the quotes next to the video and transcript, a reviewer can confirm a strong expert quickly and move them onto the lab project the same day. ## Common questions ### What are common AI training interview questions? Most first rounds cover depth in the candidate's field, reasoning out loud, writing a prompt and rubric a grader can use, and catching errors in a model's answer. The 25 questions above cover all four, plus integrity checks. ### How long should an AI trainer interview be? About 20 minutes is enough for four or five questions with their follow-ups. ### Do I need different domain expert interview questions for each field? Keep the five groups and swap the domain questions (1 to 6) and the planted errors (17 to 21) for your field. Tracy drafts role-specific questions for you, and you add your own for each domain. ### Can an AI interviewer judge a domain expert fairly? Tracy asks every applicant your core questions and follows up on vague answers. Each summary has "the participant's own words attached", direct quotes for context and an AI explanation of each answer score, alongside the video and transcript. Your reviewers check the evidence and decide. ### Do these questions work for founders too? The principle carries over: ask, then follow up on the answer. Our guide to the accelerator interview has a first round built for founders. --- **[Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=ai-trainer-interview-questions&utm_content=try-free)** or **[Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=ai-trainer-interview-questions&utm_content=demo)** ## Sources - Outlier, [outlier.ai](https://outlier.ai/), retrieved 29 Oct 2026: "Write a Challenging Prompt," "Create Grading Rubrics," "Rate and Rank Answers"; "Think of a difficult question in your field of study." - DataAnnotation, [dataannotation.tech](https://www.dataannotation.tech/), retrieved 29 Oct 2026: "debugging code, evaluating implementations, reasoning about edge cases"; finance, legal and medical roles listed "up to $125+/hr". - NHS UK Medicines Information, [Warfarin Interactions With Antibiotics in the Ambulatory Care Setting](https://www.medicinesresources.nhs.uk/en/Medicines-Awareness/Primary-Research/Other-primary-research/Warfarin-Interactions-With-Antibiotics-in-the-Ambulatory-Care-Setting), summary of a JAMA Internal Medicine study, retrieved 29 Oct 2026: "Among antibiotics, those interfering with warfarin metabolism (8.6%; metronidazole and trimethoprim-sulfamethoxazole)" ... "posed the greatest risk for an INR of 5.0 or more". - Qode, [qode.world/tracy](https://qode.world/tracy), retrieved 29 Oct 2026: "4.7/5 average candidate rating across interviews". - Qode, [qode.world/pricing](https://qode.world/pricing), retrieved 29 Oct 2026: "Interview integrity detection: Tab changes, Abnormal eye tracking, Live face detection, Average delay time before answer, External screens, Other unusual patterns"; "Video & chapter breakdown", "Transcript", "Answer summary & score", "Explain answer score by AI Agent"; Enterprise "Custom based on volume". - Qode, 28 Oct 2026: enterprise pricing is custom, based on volume. - Tracy AI Interviewer, [tracyinterview.ai](https://tracyinterview.ai), retrieved 29 Oct 2026: "Tracy AI Interviewer runs natural two-way video interviews, follows up on vague answers, digs into the interesting ones"; "Tracy AI Interviewer can ask role-specific questions, introduce a coding task, and probe the trade-offs behind an answer"; "Your questions are drafted for you." and "add the questions that matter"; "Each person opens a link in their browser and joins with no sign-up."; "People join in their browser whenever it suits them."; "Each interview becomes a readable summary with the participant's own words attached"; "Direct quotes for context". - Tracy AI Interviewer pricing, [tracyinterview.ai/pricing](https://tracyinterview.ai/pricing), retrieved 29 Oct 2026: "$1 per credit, in USD. Credits never expire."; "1 credit for each started 20 minutes, charged when the result is ready."; "A session with no answers is never charged." # What Is a Structured Interview? > What is a structured interview? Every candidate gets the same questions and is scored on set criteria. Definition, example, evidence and how to run one. - URL: https://tracyinterview.ai/blog/what-is-a-structured-interview - Published: 2026-11-10 - Author: Tracy team > **In short** Tracy AI Interviewer generates the questions from your job description, follows up on each answer in the candidate's own field and writes an assessment for each question with its reasoning, with evidence per result. > > A structured interview asks every candidate the same questions and scores each answer against criteria set in advance. > > In a 2022 re-analysis, structured interviews had a mean validity of .42 for predicting job performance, against .19 for unstructured. > > Next step: send us one role and we build the interview plan. ## What is a structured interview? A structured interview is an interview in which every candidate for a role is asked the same questions, in the same order, and each answer is scored against criteria written before the interviews began. The structure sits in two places: the questions and the scoring. The structured interview process takes the interviewer's mood, memory and chemistry with the candidate out of the decision as far as a conversation can. The opposite is an unstructured interview, where the interviewer picks questions as the conversation goes and ends with an overall impression. Most real interviews sit somewhere between the two. Moving toward the structured end is the point. ## How a structured interview process works Three parts make an interview structured. - **Job-related questions.** Each question comes from what the role needs, such as the job description and the work the person will do. Every candidate for the role gets the same core questions. - **Scoring anchors.** Before the interviews, write what a weak, an adequate and a strong answer sounds like for each question, on a scale such as 1 to 5. Interviewers score against the anchors, not against each other. - **Consistent conditions.** Same questions, same order and the same scale for every candidate. Each interviewer scores alone before the debrief, with one piece of evidence for each score. Structure is a matter of degree. Some versions fix the follow-up questions too. Others let the interviewer follow up on what the candidate said, as long as the scoring stays tied to the criteria. ![A glossary card for structured interview: same questions for every candidate, each answer scored against criteria set in advance.](/blog/images/what-is-a-structured-interview-1.png) *The definition in one card, with a one-line example. Illustration with sample data* A structured interview is defined by shared questions and scoring set in advance, not by a form. ## A structured interview example A hiring team for a customer success manager chooses three questions and writes the anchors first. Every candidate gets all three. | Question | 1: weak | 3: adequate | 5: strong | | --- | --- | --- | --- | | Tell me about a customer you nearly lost. What did you do? | Cannot give a real case, or blames the customer. | A real case with the action they took and an outcome. | A real case, the early signal they saw, what they changed afterwards and the result. | | How do you decide which accounts get your time this week? | No method, or "the loudest first". | A method with a stated criterion, such as renewal date or risk. | A method they have adjusted after a miss, with an example. | | Describe a time you had to say no to a customer. | Avoids the question. | A case where they said no and kept the account. | A case with the reason, the alternative they offered and what the customer said. | Each interviewer writes a score and a quote for every row, then the team compares. Nobody argues about whether a candidate had "good energy". They argue about whether a quote meets the 3 or the 5. Our [AI behavioral interviews](/blog/ai-behavioral-interviews) guide covers how to write the situation-and-result questions. ## Structured vs unstructured interviews | | Structured | Unstructured | | --- | --- | --- | | Questions | The same core questions for every candidate | Chosen as the conversation goes | | Scoring | Against anchors written in advance | An overall impression | | Evidence | One note or quote per score | Memory and general notes | | Comparing candidates | Row by row, on the same criteria | Hard, because the questions differed | | Effort to set up | Hours, once per role | None | The evidence favours structure. Sackett and colleagues' 2022 re-analysis in the Journal of Applied Psychology put the mean validity of structured interviews for predicting job performance at .42 and of unstructured interviews at .19. It re-analysed earlier published meta-analyses: 105 studies and 7,864 people for structured interviews and 32 studies and 3,125 people for unstructured ones. Validity here is a correlation between the interview rating and job performance ratings, so it describes how strong the link is and says nothing about how often a hire works out ([source](https://doi.org/10.1037/apl0000994)). The authors also lowered earlier estimates for most selection methods, and structured interviews still ranked first. ![Bar chart of mean validity for predicting job performance: unstructured interviews 19, structured interviews 42.](/blog/images/what-is-a-structured-interview-2.png) *Mean validity of structured and unstructured interviews in Sackett and colleagues' 2022 re-analysis, multiplied by 100. Source: Sackett, Zhang, Berry and Lievens (2022), Journal of Applied Psychology 107(11), Table 3 and Table 1.* The link between interview rating and job performance is about twice as strong when the interview is structured. ## What a structured interview does not do Structure narrows how much the interviewer's mood decides the outcome. It does not make the criteria correct, and it does not remove bias. A badly written question or an anchor that rewards one background still scores badly, now consistently. The people who write the criteria should be the people who know the job, and someone outside the hiring team should read them. A structured interview also does not replace the decision. A person still weighs the scores, the evidence and the parts of the role no interview can test. See how that works when the first round is run by an [AI interviewer](/blog/what-is-an-ai-interviewer), and how [AI and human interviews](/blog/ai-interview-vs-human-interview) compare. ## Running a structured interview with Tracy Keeping an interview structured across many candidates is the hard part. - **Questions from the job description.** Tracy generates them from the role. Edit them or use your own. - **Skills you choose.** You choose the skills to evaluate as tags when you set up a role, and Tracy writes an assessment for each question with its reasoning. - **Follow-ups.** She asks in the candidate's own field and follows up on what they said. A Director at a top-tier early-stage VC said: "Tracy sometimes brings up exactly those questions I might have forgotten to ask, which has been a pleasant surprise." - **Evidence per result.** A recording of both sides, a timestamped transcript and a written assessment for each question with its reasoning, so a reviewer can check the evidence behind a result. A person decides from the evidence. Tracy's role is the interview and the report. Learn the mechanics in [how AI interviews work](/blog/how-do-ai-interviews-work) and see [how to screen candidates](/blog/how-to-screen-candidates) for the steps around it. The first 3 results are free with no card. After that it is $1 per credit, one credit per started 20 minutes of an interview that produced a result, with no seats and no upfront cost. See [pricing](/pricing). To see the method applied, set up one role and read the evidence from its first interviews. ![Four steps in Tracy: your questions, your chosen skills, live follow-ups, evidence with recording, transcript and assessment.](/blog/images/what-is-a-structured-interview-3.png) *Tracy generates the questions from your job description, follows up on each answer and writes an assessment for each question, so the evidence stays comparable across candidates. Illustration with sample data* Tracy generates the questions from your job description, follows up on each answer and writes an assessment for each question. ## Related terms - **Behavioral interview.** Questions about what the candidate did in a past situation. It can be structured or not. - **Situational interview.** Questions about what the candidate would do in a described situation. - **Scorecard.** The sheet where scores and evidence are written down, one row per criterion. - **Scoring anchor.** The written description of what a 1, a 3 and a 5 sound like for a question. - **Follow-up question.** A question that builds on the answer just given. Which path fits you: - Screening applicants: start with [applications](/use-cases/applications). - Hiring at volume: see [hiring at volume](/use-cases/volume-hiring). - Vetting specialists: see [expert vetting](/use-cases/expert-vetting). ## Common questions ### What is a structured interview process? Write the criteria from the job, choose questions that test them, write what a weak, adequate and strong answer sounds like, ask every candidate the same core questions, score each answer alone with evidence, then compare in a debrief. ### Is a structured interview better than an unstructured one? On the evidence, yes for predicting job performance. The 2022 re-analysis above found a stronger link between structured interview ratings and job performance. Validity is a correlation, so it shows a stronger link and no guarantee. ### Can you ask follow-up questions in a structured interview? Yes. Some versions fix them in advance and others let the interviewer follow up on the answer. Either way the scoring stays tied to the criteria written beforehand. ### How many questions should a structured interview have? Enough to cover each criterion once or twice. Four to eight is a workable range for a first round. That is our rule of thumb, not a researched figure. ### Does a structured interview remove bias? No. It limits how much the interviewer's mood and chemistry decide, but the criteria have to be fair too. Tracy publishes no bias audit, and a person decides from the evidence report. ### Can an AI interviewer run a structured interview? Yes. Tracy generates the questions from the job description, follows up on each answer and writes an assessment for each question, with a recording and written assessment per question for a person to review. --- See the method applied in one interview for your own role. Send us one role and we will build the interview plan in days, with a recording and report you can check. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=what-is-a-structured-interview&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=what-is-a-structured-interview&utm_content=demo) ## Sources - [Sackett, Zhang, Berry and Lievens (2022), Revisiting meta-analytic estimates of validity in personnel selection, Journal of Applied Psychology 107(11)](https://doi.org/10.1037/apl0000994), retrieved 9 Nov 2026. - [Tracy AI Interviewer home page, customer quotes](https://tracyinterview.ai/), retrieved 9 Nov 2026. - [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 5 Nov 2026. - [Tracy AI Interviewer Connect page](https://tracyinterview.ai/connect), retrieved 5 Nov 2026. # What Is Interview Fraud? > Interview fraud is misrepresenting who you are or what you can do in a hiring interview. Definition, five common types and what a reviewer can check. - URL: https://tracyinterview.ai/blog/what-is-interview-fraud - Published: 2026-11-10 - Author: Tracy team > **In short** **Tracy AI Interviewer** shows a reviewer the evidence behind an interview, such as tab changes, gaze and delay before answers, each with its times. > > **Interview fraud** is misrepresenting who you are or what you can do in a hiring interview, so the employer decides on a false picture. > > Signals are for a person to review and are not proof. Nothing is decided in the interview, and candidates are told on the join screen that it is recorded and transcribed. > > Next step: send us one role and see the signals on your own interview. ![Glossary card defining interview fraud as misrepresenting who you are or what you can do in an interview.](/blog/images/what-is-interview-fraud-1.png) *The definition in one card: the concept Tracy gives reviewers evidence for. Illustration. Illustration with sample data* Interview fraud means the employer decides on a false picture of the person. Interview fraud is misrepresenting who you are or what you can do in a hiring interview, so that the employer decides on a false picture. It covers a person who is not who they say they are, and a person who is who they say but cannot do what they claim. People also call it candidate fraud or interview cheating. The terms overlap, but fraud is the wider one: it includes a stolen identity or a made-up work history, not only a prompt whispered off camera. A remote interview is where it is easiest to try and where it is hardest to see. The reviewer cannot see the desk, the second screen or the person sitting out of frame. That is why the term matters most to teams who hire or vet people over video, such as [expert vetting](/use-cases/expert-vetting) for AI projects. If the interviewer is itself an AI, our guide to [what an AI interviewer is](/blog/what-is-an-ai-interviewer) explains how it works. ## Five common forms of interview fraud These are the common forms as this guide groups them. They are not exclusive, and one interview can contain more than one. - **Someone else answers.** A proxy takes the interview, and a different person does the job. The face and voice on the call are not the hire's. - **Undisclosed outside help.** The person is real, but another person or an AI tool is feeding answers. This is the case employers argue about most, because the rules on tools differ by employer and by role. - **A deepfake or cloned voice.** The video or audio is altered to look or sound like someone else. - **A stolen or false identity.** Another person's name and details are used to apply, and sometimes to pass a background check. - **Invented credentials or work history.** The person is real and present, and the CV or the story is not true. The first four are about who is in the interview. The fifth is about what they say. A good process treats them differently, because a check that helps with one does little for another. ## What counts as fraud, and what does not Fraud needs misrepresentation: the employer is led to believe something false about the person or their ability. A nervous pause, notes on the desk or a repeated question are not that. The line on tools depends on the employer. A company that allows an assistant for a coding task has no fraud if the candidate uses one. A company that forbids it, and says so before the interview, does. We recommend stating your rule in the invitation, because a rule that was never said is not one a person can break on purpose. ## The FBI has warned about deepfakes and stolen identities in remote job applications The FBI has documented the problem. In a public service announcement dated 28 June 2022, the FBI's Internet Crime Complaint Center warned of "an increase in complaints reporting the use of deepfakes and stolen Personally Identifiable Information (PII) to apply for a variety of remote work and work-at-home positions". The announcement says some of the positions include access to customer data, financial data or corporate IT databases. It also describes what reviewers noticed. In some interviews, "the actions and lip movement of the person seen interviewed on-camera do not completely coordinate with the audio of the person speaking". And in some cases, "pre-employment background checks discovered PII given by some of the applicants belonged to another individual." The announcement gives no rate and does not say how common this is. It warns that the pattern exists, and it comes from the agency that receives the complaints. ## No single check proves interview fraud, so reviewers look at several Interview fraud detection is not one test. Teams combine what they can see: - **The content of the answers.** Does the person explain how they got to a result, and can they go one level deeper when asked? - **How the answers arrive.** Long, regular pauses before each answer can mean a person is waiting for something. - **What the person does on the screen.** Switching tabs, leaving fullscreen or a second display. - **Whether the face and voice stay consistent.** A video that stops matching the audio is a reason to look closer. - **What the CV and the interview say about each other.** A claimed specialty the person cannot discuss is a mismatch. - **A separate identity or credential check.** Verifying an ID or a licence is its own step in your process. Each of these is evidence for a person to weigh. None of them settles the question alone, and a team that treats a single signal as proof will reject honest people. For the full product, see [how Tracy works](/product). ## How Tracy shows the evidence to a reviewer Tracy AI Interviewer captures tab changes, abnormal eye tracking (gaze), live face detection, delay before answers, leaving fullscreen and external screens, and shows that evidence to a person in the interview report. She reports them with their times, and nothing is decided in the interview. The person interviewed is told on the join screen that the conversation is recorded and transcribed, and the decision stays with the customer. The follow-up is part of the evidence as well. She asks about what the person just said and waits for the answer. The reviewer reads how each answer holds up one question deeper. The recording and the timestamped transcript sit behind each result, so the reviewer can go to the minute a signal appeared and listen. ![Recreated signals panel, sample data: delay before answers four times, fullscreen exits twice, with times; three signals clear.](/blog/images/what-is-interview-fraud-2.png) *Tracy shows a reviewer each signal with its times, so a person can listen to those minutes and decide. Illustration with sample data* The signals come with their times, so a reviewer can listen to those minutes and decide. ## What should a reviewer do when a signal appears? Take the sample above. A pause appears four times, and the person leaves fullscreen twice, at 05:36 and 09:28, a few seconds before two of the pauses. That pattern is a reason to listen to those minutes and ask the person about them, and no more than that. A long pause can be a bad connection, a notification or a candidate who thinks slowly. Leaving fullscreen can be an accident. So the reviewer does three things: reads the transcript around the signal, asks the person to explain the answer in their own words, and compares it with the rest of the interview. We would not reject anyone on a signal alone, and a good process does not need to. Our guide on [how integrity signals work in an AI interview](/blog/ai-interview-cheating-detection) explains what each one means and what a reviewer should do with it, and [what an AI interview report shows](/blog/ai-interview-reports) covers where the evidence sits. To see the signals on one of your own interviews, send us one role. ## Common questions ### What is interview fraud? It is misrepresenting who you are or what you can do in a hiring interview, so the employer decides on a false picture. Examples are a proxy answering, undisclosed outside help, a deepfake, a stolen identity or invented credentials. ### How is interview fraud detected? No single check proves it. Teams combine the content of the answers, how the answers arrive, behaviour on the screen, consistency of face and voice, and a separate identity or credential check. Each is evidence for a person to weigh, not a verdict. ### Is using AI in an interview fraud? It depends on the employer's rule. If the rule is stated and the person breaks it on purpose, it is a misrepresentation. If no rule was stated, treat it as a question to ask, not an accusation. ### Can an AI interviewer help with interview fraud? It can show a reviewer evidence. Tracy captures signals such as tab changes, gaze, live face detection and delay before answers, and asks follow-ups that show how an answer holds up. A person reviews the evidence and decides. ### Who should decide when a signal appears? A person. Read the transcript around the signal, ask the candidate to explain, and weigh it against the rest of the interview. ### What is the difference between interview fraud and interview cheating? Cheating is the narrower term, usually about outside help during answers. Fraud is wider and also covers a stolen identity or a made-up work history. The terms overlap. --- See the signals and evidence in one real interview: send us one role and we will build the interview plan in days, with a recording and report you can check. The first 3 results are free, with no card. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=what-is-interview-fraud&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=what-is-interview-fraud&utm_content=demo) ## Sources - [FBI IC3 public service announcement I-062822-PSA, 28 June 2022: deepfakes and stolen PII used to apply for remote work](https://www.ic3.gov/PSA/2022/PSA220628), retrieved 9 Nov 2026. - [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 5 Nov 2026. # What Is a Competency-Based Interview? > What is a competency-based interview? Ask for evidence of the skills a role needs and score every answer the same way. Definition, example, limits. - URL: https://tracyinterview.ai/blog/what-is-a-competency-based-interview - Published: 2026-11-09 - Author: Tracy team > **In short** **Tracy AI Interviewer** runs a competency-based first round for every applicant: you choose the skills to evaluate for each role, and she asks about each skill and follows up on the answer. > > A **competency-based interview** asks for evidence of the skills a role needs and scores every answer against the same criteria. > > Each result keeps a recording and timestamped transcript, plus a written assessment per question. > > Next step: send us one role and we build the interview plan. ## What is a competency-based interview? A competency-based interview asks each candidate for evidence of the specific skills and behaviours a role needs, then scores every answer against the same criteria. What does that look like in practice? The interviewer picks a handful of competencies from the job, asks for a real example of each, and rates the example instead of the impression the person made. A competency is a skill or behaviour you can observe and measure. The UK Civil Service uses the word behaviours for the same idea and describes them as "the actions and activities that people do which result in effective performance in a job". It adds that behaviours "can be observed and measured". That is the whole approach in one line: you look for evidence of named competencies and not for a personality. ![Glossary card defining a competency-based interview: evidence of the skills a role needs, scored against the same criteria.](/blog/images/what-is-a-competency-based-interview-1.png) *The definition in one card, with a sample question. Illustration with sample data* A competency-based interview is defined by named skills, real examples and shared scoring. ## How a competency-based interview works The scoring is decided before the first interview, and the order of the steps shows it. 1. **Choose the competencies.** Take four to six from the job description, such as ownership, problem solving or working with others. Fewer and you miss parts of the job; more and each gets a thin question. 2. **Write the anchors.** For each competency, write what a weak, adequate and strong answer contains. 3. **Ask for past behaviour.** Questions that start with "tell me about a time you" begin from something the person did. 4. **Probe the example.** The Civil Service guide to applying suggests the STAR method, which it describes as setting the scene, showing what and how you did, and the overall outcome. A good interviewer asks for whichever part is missing, usually the action and the result. 5. **Score each answer against its anchors**, one competency at a time, and compare candidates on the same scale. ## A competency-based interview example Here is one role's worth of competencies. The questions are set per competency before the first interview. The follow-up depends on what the person says. | Competency | Question | A strong answer has | Follow-up | |---|---|---|---| | Ownership | Tell me about a project that went wrong after you handed it over. | A named project, what the person did next and what changed afterwards | What would you have checked before the handover? | | Problem solving | Describe a problem where your first fix did not work. | The first approach, why it failed and how the cause was found | What did you rule out, and how? | | Working with others | Tell me about a disagreement with a colleague over how to do a task. | The disagreement in concrete terms, the person's own move and the outcome | What did the other person say when you proposed it? | This table shows three competencies from a sample role: the question, the evidence a strong answer contains and one follow-up. It is sample content. Each answer is scored on its own, against the anchors for that competency. The follow-up carries most of the weight. A rehearsed story survives the first question and rarely survives a request for the exact thing the person did, said or decided. ## Competency-based, behavioural and structured interviews The terms overlap and people use them loosely. In this guide, a behavioural question asks about something you did in the past. A competency-based interview organises the whole interview around a defined list of competencies. A structured interview asks every candidate the same questions and scores them against criteria set in advance. A competency-based interview is usually both behavioural and structured, but each word describes something different: the type of question, the content of the interview and the way it is scored. For the behavioural side, see our guide to [AI behavioural interviews](/blog/ai-behavioral-interviews). ## Where a competency-based interview falls short The method measures what people say they did. A person can tell a vivid story about work someone else did, and a nervous person can tell a real one badly. Three habits limit the damage: ask for detail only the person who did the work would have, score the evidence in the answer and not the delivery, and keep a person in the decision. The scores also depend on the competencies you chose. If the list does not match the job, a well-run interview measures the wrong things. And a rating from one interviewer on one day is still one rating, which is why the evidence behind each score should be easy to check. ## Running a competency-based interview with Tracy Tracy AI Interviewer applies the method to every applicant. You choose the skills to evaluate for each role, and she generates the questions from your job description. You can edit them or use your own. She asks each applicant, in the applicant's own field, about each skill, follows up on what they said and waits for the answer. Nobody schedules anything: each person starts from a link when it suits them, which is what makes the method workable for [hiring at volume](/use-cases/volume-hiring). Each result keeps a recording and a timestamped transcript, plus a written assessment per question with its reasoning, so a reviewer can check what the person said against the competency. A recruiter keeps the last word: they can move someone forward or invite them to interview again. Our guide to [what an AI interview report shows](/blog/ai-interview-reports) covers where each piece of evidence sits. Tracy also raises questions an interviewer may skip. A Director at a top-tier early-stage VC says: "Tracy sometimes brings up exactly those questions I might have forgotten to ask, which has been a pleasant surprise." If you want to see a chain on your own role, send us the role and we build the interview plan. The first 3 results are free, with no card. ![Sample chain: Tracy asks about a plan that went off track, the candidate answers, Tracy asks who was told first.](/blog/images/what-is-a-competency-based-interview-2.png) *Tracy asks about one competency and follows up on the answer, so the evidence is specific. Illustration with sample data* Tracy asks about one competency, then follows up on what the applicant said. ## Related terms - **Behavioural interview:** questions about past behaviour. See [AI behavioural interviews](/blog/ai-behavioral-interviews). - **Anchor:** a written description of what a weak, adequate or strong answer to a question looks like. - **Screening interview:** the first-round conversation that decides who moves on. See [AI screening interviews](/blog/ai-screening-interviews). - **STAR:** situation, task, action, result, a frame for telling an example. ## Common questions ### What is a competency-based interview? It is an interview built around a defined list of competencies. The interviewer asks for a real example of each and scores every answer against the same criteria. ### What are competency-based interview questions? Most begin with a past situation, such as "tell me about a time you handled a disagreement with a colleague". Each question maps to one competency, and the follow-up asks for the detail behind the answer. ### What is the STAR method? STAR means situation, task, action, result. The UK Civil Service describes it as setting the scene, showing what and how you did, and the overall outcome. Interviewers use it to check an answer has all four parts. ### How many competencies should an interview cover? Four to six is a workable range for a first round. ### Can an AI interviewer run a competency-based interview? Yes. Tracy generates questions from the job description, follows up on each answer and assesses on the skills you choose. A person reviews the recording, transcript and written assessment and decides. --- See the method applied in one real interview. Send us one role and we will build the interview plan in days, with a recording and report you can check. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=what-is-a-competency-based-interview&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=what-is-a-competency-based-interview&utm_content=demo) ## Sources - [UK Cabinet Office, Success Profiles: Civil Service behaviours (updated 29 January 2025)](https://www.gov.uk/government/publications/success-profiles/success-profiles-civil-service-behaviours), retrieved 9 Nov 2026. - [UK Cabinet Office, Success Profiles: Candidate overview (STAR method)](https://www.gov.uk/government/publications/success-profiles/success-profiles-candidate-overview), retrieved 9 Nov 2026. - [Tracy AI Interviewer home page, customer quotes](https://tracyinterview.ai/), retrieved 9 Nov 2026. - [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. # What Is a Deepfake Candidate? > What is a deepfake candidate? An applicant whose video or voice is altered to hide who they are. Definition, what the FBI says, what an interview shows. - URL: https://tracyinterview.ai/blog/what-is-a-deepfake-candidate - Published: 2026-11-09 - Author: Tracy team > **In short** **Tracy AI Interviewer** gives a reviewer evidence from the interview: follow-ups on the person's own work, and signals such as delay before answers and tab changes, each with its evidence. > > A **deepfake candidate** is an applicant whose video, voice or both are altered or generated to misrepresent who is on the call. > > Signals are for a person to weigh. They are not proof, and the decision stays with your team. > > Next step: send us one role and see the evidence on your own interview. ## What is a deepfake candidate? What is a deepfake candidate? It is a job applicant who uses a deepfake to pass a remote interview as someone they are not. The FBI's Internet Crime Complaint Center defines deepfakes as "a video, an image, or recording convincingly altered and manipulated to misrepresent someone as doing or saying something that was not actually done or said". The word covers two different things. A face swap changes what the interviewer sees, and a spoofed or cloned voice changes what the interviewer hears. The FBI's 2022 alert describes complaints about voice spoofing, or potentially voice deepfakes, in online interviews, and about stolen personal data used in the same applications. ![Glossary card: a deepfake candidate is an applicant whose video or voice is altered to misrepresent who is on the call.](/blog/images/what-is-a-deepfake-candidate-1.png) *The definition in one card. Illustration with sample data* A deepfake candidate hides who is on the call by altering the video or the voice. ## What the FBI has said about deepfakes in job interviews The FBI warned on 28 June 2022 of "an increase in complaints" about deepfakes and stolen personal data used to apply for remote work. The positions named were in information technology, programming, database and software roles, and some had access to customer personal data, financial data or corporate IT databases. The complaints describe interviews where on-camera lip movement and actions do not completely match the audio, and where coughing or sneezing does not line up with the picture. On 23 January 2025 the FBI added that North Korean IT workers "have been observed using artificial intelligence and face-swapping technology during video job interviews to obfuscate their true identities". That alert is about one scheme. It does not say how common deepfake candidates are in general, and neither alert gives a rate. They tell hiring teams the pattern exists. ## Deepfake candidate, proxy, stolen identity and AI help People use several terms for people who are not who they seem in an interview. They are different problems, and each leaves different evidence. Four deepfake-related terms: what is false and what a reviewer may notice (our grouping). | Term | What is altered or false | What a reviewer may notice | |---|---|---| | Deepfake candidate | The face, the voice or both are altered or generated | Video and audio that drift apart. The FBI describes lip movement and actions that do not fully match the audio | | Proxy or stand-in | A different real person answers | Answers that do not fit the CV | | Stolen identity | The name and personal data belong to someone else | A background check that returns another person's data | | AI-fed answers | A real person reads answers written by a tool | Delay before answers or gaze that looks abnormal, which can also have innocent causes | The first two rows are about who is on the call. For how the whole family of problems is organised, see our guide to [AI interview cheating detection](/blog/ai-interview-cheating-detection), which covers the AI-fed-answers case in depth. ## What an interview can and cannot show about a deepfake A conversation cannot verify who someone is. An interview can produce evidence that something is off, and confirming identity belongs to a separate step. The FBI's January 2025 advice is to "implement identity-verification processes during interviewing, onboarding, and throughout the employment of any remote worker". What an interview can show is narrower, and still useful: - **A video that stops matching the audio.** The mismatch the FBI describes is something a reviewer can see and hear. - **Answers that need the live person.** A question about the answer just given is hard to prepare for. A person using a face swap still has to answer it from their own knowledge. - **Behaviour on the screen.** Tab changes, leaving fullscreen and extra screens can sit around an answer. - **Timing.** A pause before each answer can mean a person is waiting for something, and it can also be a poor connection. Each item is a reason to look. None of them settles the question alone. ## How Tracy shows a reviewer the evidence Tracy AI Interviewer captures tab changes, abnormal eye tracking (gaze), live face detection, delay before answers, leaving fullscreen and external screens, and shows the evidence to a person in the interview report. Live face detection is one signal among six that a reviewer reads. It is not a deepfake verdict. The signals go to a person, who decides. Nothing is decided in the interview. The person interviewed is told on the [join screen](/participant-experience) that the conversation is recorded and transcribed, and the decision stays with the customer. The follow-up carries the rest. Tracy asks about what the person just said, in the person's own field, and waits for the answer. For teams that vet experts remotely, see [expert vetting](/use-cases/expert-vetting). A recording of both sides and a timestamped transcript sit behind each result, with a written assessment per question, so a reviewer can read the transcript around the part they want to check. For the full report, see [what an AI interview report shows](/blog/ai-interview-reports), and for the interview itself, [how AI interviews work](/blog/how-do-ai-interviews-work). ![Four steps: follow-ups, signals with evidence, recording with transcript, a person decides.](/blog/images/what-is-a-deepfake-candidate-2.png) *Tracy gives a reviewer four kinds of evidence from one interview, and a person decides. Illustration with sample data* Tracy gives a reviewer follow-ups, signals, a recording and a transcript, and a person decides. ## What should a reviewer do when something looks wrong? Suppose a signal appears and something in the video looks wrong. A reviewer can do four things, in this order. Read the transcript around that minute. Ask the person a different follow-up on the same topic, in a short live call. Run the separate identity check your process already has (see [how Tracy fits your process](/product)). Then weigh all of it against the rest of the interview. Apply the same steps to every candidate. A check that only some people get is a check on those people, not on the process. Say in the invitation that the interview is recorded and that you review signals, so the rule exists before anyone can break it. To see the evidence on one of your own interviews, send us one role. The first 3 results are free, with no card. ## Common questions ### What is a deepfake candidate? An applicant whose video, voice or both are altered or generated so the interviewer sees or hears someone else. The FBI defines a deepfake as a video, image or recording convincingly altered to misrepresent someone. ### What is deepfake detection? Usually software that analyses video or audio for signs of manipulation. Tracy does not claim to do that. She shows a reviewer interview signals and the evidence behind them, and a person decides. ### How does a deepfake work? In a face swap, software puts another face on the person seen on camera, and a voice clone does the same for sound. The FBI's 2025 alert says face-swapping technology has been observed in video job interviews. ### Can you spot a fake candidate in a video interview? Sometimes. Lip movement that does not match the audio, answers that fall apart one question deeper and unusual behaviour on screen are reasons to look closer. None of them is proof, so a person weighs them. ### Is a deepfake candidate the same as a proxy? No. A deepfake alters the face or voice of the person on the call. A proxy is a different real person answering. Both are about who is in the interview, and each needs its own check. --- See the method applied in one real interview: send us one role and we will build the interview plan in days, with a recording and report you can check. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=what-is-a-deepfake-candidate&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=what-is-a-deepfake-candidate&utm_content=demo) ## Sources - [FBI IC3 public service announcement I-062822-PSA, 28 June 2022: deepfakes and stolen PII used to apply for remote work](https://www.ic3.gov/PSA/2022/PSA220628), retrieved 9 Nov 2026. - [FBI IC3 public service announcement I-012325-PSA, 23 January 2025: North Korean IT workers conducting data extortion](https://www.ic3.gov/PSA/2025/PSA250123), retrieved 9 Nov 2026. - [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. # What Is a Pre-Screening Interview? > What is a pre-screening interview? The short first conversation that checks a role's basics. Definition, example and how to run one for every applicant. - URL: https://tracyinterview.ai/blog/what-is-a-pre-screening-interview - Published: 2026-11-09 - Author: Tracy team > **In short** Tracy AI Interviewer runs the pre-screen for every applicant, starting on click at any hour, with follow-ups on what each person says and a recording, transcript and written assessment per question. > > A pre-screening interview is the short first conversation that decides who moves on before a hiring manager spends time. > > A Director at a top-tier early-stage VC went from about six hours a day on screening to about 1.5. > > Next step: send us one role and we build the interview plan. ## What is a pre-screening interview? A pre-screening interview is a short first conversation that checks whether an applicant meets the basic requirements of a role before the team spends longer interviews on them. It is also called a preliminary screening interview or a screening call. The questions come straight from the job: the experience and skills the role cannot work without, availability, location or work authorisation, and sometimes a salary range. The output is a decision, not a hire. Each applicant is moved on, held or stopped, and the people who move on go to the hiring manager with what they said written down. ## How a pre-screening interview works The format varies, but the steps are the same. 1. **Write the must-haves.** List the three to six things an applicant needs for the role, taken from the job description. 2. **Ask each applicant about the same must-haves.** A recruiter does it in a pre screening phone interview, or the applicant answers a recorded or AI-run interview. 3. **Write down what was said.** Notes or a transcript, one line per must-have. 4. **Decide.** Move on, hold or stop, and pass the people who move on to the next round. A pre-assessment interview is the same idea placed before a test or assignment: it checks the applicant is worth the time of the assessment. The pre screening interview call is the phone version, usually run by a recruiter. ![Glossary card: pre-screening interview, a short first conversation that checks a role's basics before longer interviews.](/blog/images/what-is-a-pre-screening-interview-1.png) *The definition in one card, with a one-line example. Illustration with sample data* A pre-screening interview checks the basics of a role and ends in a decision for each applicant. ## A pre-screening interview example A team hiring a customer support agent writes four must-haves and one question for each. Sample pre-screen for a support agent role: four must-haves, one question each, what a pass sounds like and when to stop (fictional example). | Must-have | Question | A pass sounds like | Stop if | | --- | --- | --- | --- | | Support experience | Describe the last customer problem you solved. | A real case, the steps and the outcome. | No real case after one follow-up. | | Tools | Which help desk tools have you used, and for how long? | Names the tools and what they did in them. | Has not used any, and the role needs one from day one. | | Shifts | Can you work weekend shifts? | A clear yes or no. | A no, when the role is weekend cover. | | Written English | Reply to this upset customer in two sentences. | Clear, polite and specific. | Cannot write a clear reply. | The recruiter scores each row pass, hold or stop and writes one line of evidence for it. Nobody needs a second call to remember what was said. ## What a good pre-screening question looks like - **Tied to a must-have.** If the answer would not change the decision, drop the question. - **Open enough to need a real answer.** "Describe the last customer problem you solved" shows more than "Do you have support experience?", which anyone can answer yes. - **One thing at a time.** Two questions in one get half an answer. - **About the job.** Ask what the role requires. Leave out age, family, health, religion and other personal topics unrelated to the work. Keep the list short. A pre-screen that feels like an exam loses applicants before the team has met them. ## Why pre-screening usually covers only a short list A recruiter call takes time to schedule and time to run, so many teams call only the applicants whose résumés look strongest and leave the rest. The pre-screen is meant to find the people a résumé hides, and it only reaches the people the résumé already favoured. Two changes help. Let the applicant start the conversation when they choose, so nothing is scheduled. And score every applicant against the same must-haves, so the results can be compared. An [AI screening interview](/blog/ai-screening-interviews) works this way, and it is the idea behind [screening hundreds of applicants](/blog/how-to-screen-hundreds-of-applicants). ## Running the pre-screen for every applicant with Tracy Tracy runs the pre-screening interview as a conversation the applicant starts from a link. - **Starts on click.** Each applicant starts at any hour, with no scheduling and no recruiter present. Nothing to download and no account. - **Your must-haves.** Tracy generates questions from your job description. Edit them or use your own. - **Follow-ups.** She asks in the applicant's own field, follows up on what they said and waits for the answer. When the CV check is on, she reads the CV too. - **Evidence per result.** A recording of both sides, a timestamped transcript and a written assessment per question with its reasoning. - **One record per applicant.** The recruiter reads the same kind of evidence for everyone who took the interview, so answers can be compared must-have by must-have. A recruiter keeps the last word, moves a candidate forward or invites them to interview again. See [how AI interviews work](/blog/how-do-ai-interviews-work) for the mechanics and [hiring at volume](/use-cases/volume-hiring) for the full use case. The first 3 results are free with no card. After that it is $1 per credit, one credit per started 20 minutes of an interview that produced a result, with no seats and no upfront cost. See [pricing](/pricing). To see a pre-screen for one real role, set it up and read the first results. ![Four steps in Tracy: your must-haves, starts on click, live follow-ups, evidence with recording and transcript.](/blog/images/what-is-a-pre-screening-interview-2.png) *Tracy generates the questions from your job description, starts each interview on click and follows up on what applicants say, so the recruiter reads the evidence instead of making calls. Illustration with sample data* Tracy pre-screens every applicant who clicks the link and hands the recruiter evidence to read. ## Related terms - **Screening interview.** The broader term. A pre-screen is the first and shortest one. - **Phone screen.** A pre-screen done by a recruiter on a call. - **Knockout question.** A question with a required answer, such as work authorisation. A wrong answer ends the process. - **Technical screen.** A first interview that tests the skills of a technical role. - **Structured interview.** One where every candidate gets the same questions and is scored against criteria set in advance. Which path fits you: - Hiring at volume: see [hiring at volume](/use-cases/volume-hiring). - Screening applications: see [applications](/use-cases/applications). - Technical roles: see [technical interviews](/use-cases/technical-interviews). ## Common questions ### What is a pre-screening interview? A short first conversation that checks whether an applicant meets the basic requirements of a role, before longer interviews. It ends in a decision: move on, hold or stop. ### What is a preliminary screening interview? The same thing under another name. It is the first interview in a hiring process and covers the must-haves of the role. ### What is a pre screening phone interview? A pre-screening interview run by a recruiter on a call. It works for a short list and gets slow when applicants number in the hundreds. ### What is a pre-assessment interview? A short interview placed before a test or assignment, to check the applicant is worth the assessment time. It is a pre-screen with a specific next step. ### How many questions should a pre-screening interview have? Enough to cover each must-have once, usually three to six. ### Can an AI interviewer run a pre-screening interview? Yes. Tracy runs it for every applicant on click, follows up on what each person says and leaves a recording and written assessment per question for a person to review. --- See the method applied in one real interview. Send us one role and we will build the interview plan in days, with a recording and report you can check. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=what-is-a-pre-screening-interview&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=what-is-a-pre-screening-interview&utm_content=demo) ## Sources - [Tracy AI Interviewer home page, customer quotes](https://tracyinterview.ai/), retrieved 9 Nov 2026. - [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer Connect page](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. # What Is a Technical Screen? > What is a technical screen? The early interview that tests technical skills before a longer loop. Definition, who runs it, an example and how to scale it. - URL: https://tracyinterview.ai/blog/what-is-a-technical-screen - Published: 2026-11-09 - Author: Tracy team > **In short** Tracy AI Interviewer runs the technical screen for every applicant: a coding task inside the interview, then follow-ups on approach and trade-offs, with a recording and written assessment per question. > > A technical screen is the early interview that tests whether a candidate has the skills a technical role needs, before a longer loop. > > The first 3 results are free, then $1 per credit, with no seats and no upfront cost. > > Next step: send us one role and we build the interview plan. ## What is a technical screen? A technical screen is an early interview that tests whether a candidate has the technical skills a role needs, before the team schedules a longer technical loop. It is also called a technical screening round, a technical phone screen or a technical screen interview. It sits after the résumé review and before the on-site or final interviews. The output is a decision with notes: pass the candidate to the full loop, hold, or stop. Many technical screens combine three things: fundamentals questions, a short coding or problem-solving task, and a conversation about past work. ## Who runs a technical screen? Four setups are common. Four ways to run a technical screen, what each looks like and where it strains. | Who runs it | What it looks like | Where it strains | | --- | --- | --- | | An engineer from the team | A call, a shared editor and a problem to solve together. | Engineer hours. Each screen takes the time of someone who is also building the product. | | A recruiter with a script | Fixed fundamentals questions and a pass mark. | Limited depth. A script covers the questions it lists, so an unexpected approach goes unexplored. | | An online assessment | Timed tasks the candidate completes alone. | It scores the submitted answer. Nobody asks why unless a reviewer adds a call. | | An AI interviewer | A live conversation with a coding task and follow-ups. | Needs a clear set of skills to test and a human to decide. | Teams often mix them: an assessment to filter, then a screen to understand how the candidate thinks. ![Glossary card for technical screen: an early interview that tests the skills a technical role needs before a longer loop.](/blog/images/what-is-a-technical-screen-1.png) *The definition in one card, with a one-line example. Illustration with sample data* A technical screen tests the skills a technical role needs, early, and ends in a decision. ## A technical screen example A backend engineer screen might run in this order. Sample backend engineer screen: four parts, an example prompt and what a strong answer includes (fictional example). | Part | Example prompt | A strong answer includes | | --- | --- | --- | | Fundamentals | What is the difference between a process and a thread? | A correct distinction and when each is the right tool. | | Coding task | Write a function that removes duplicates from a list. | Working code, a stated approach and an edge case. | | Follow-up on the task | What is its time complexity? What changes if the list does not fit in memory? | A cost for the approach and a different design under the new limit. | | Past work | Tell me about a performance problem you fixed. | What they measured, what they changed and the result. | The follow-ups are where the screen earns its time. A correct function shows the candidate can solve the problem; the follow-up on memory shows whether they can reason past it. Our [software engineer interview questions](/blog/software-engineer-interview-questions) page has more prompts of this kind. ## Technical screen vs the full technical loop The screen is the first filter and the loop is the full evaluation. A screen is shorter and narrower: it asks whether the candidate has the core skills at all. The loop comes after and goes deeper, with several rounds that may include system design, a longer coding problem and a conversation with the hiring manager. A good screen saves the loop for the people most likely to pass it. A weak screen lets too many through, and engineers spend loop time on candidates who should have stopped earlier. It can also stop strong candidates, if the questions test trivia instead of the work. ## What makes a technical screen work - **Test the skills the job uses.** Pick the three or four skills the role needs this year and write a question for each. Skip trivia the team would look up. - **Ask each candidate about the same skills.** It lets you compare people on the same ground. - **Follow up on the approach.** Ask why, what it costs and what breaks. This separates a memorised answer from understanding. - **Write down evidence.** One line or quote per skill, so a second person can check the decision. A screen that skips the follow-ups becomes a quiz. A screen that skips the evidence becomes an opinion. ## Running the technical screen with Tracy Tracy runs the technical screen as a conversation the candidate starts from a link, so no engineer has to be on the call. - **Coding task inside the interview.** Live coding, SQL, debugging or multiple choice, in the same interview as the questions. - **Follow-ups on approach and trade-offs.** She asks in the candidate's own field and follows up on what they said. - **Questions from the job description.** Edit them or use your own, and set the skills and weights for the role. - **Evidence per result.** A recording of both sides, a timestamped transcript and a written assessment per question with its reasoning. A person decides from the evidence. Tracy runs the interview and writes the report. See [AI technical interviews](/blog/ai-technical-interviews) for the full workflow and [technical interviews](/use-cases/technical-interviews) for the use case. The first 3 results are free with no card. After that it is $1 per credit, no seats, no upfront cost. See [pricing](/pricing). To see the method applied, set up one engineering role and read the first results. ![Sample interview: Tracy asks the complexity of a deduplication function, then what changes if the list outgrows memory.](/blog/images/what-is-a-technical-screen-2.png) *A sample follow-up chain: Tracy asks the cost of an approach and what changes at scale, so a reviewer sees how the candidate thinks. Illustration with sample data* Tracy follows a coding answer with a question about its cost, so the screen tests reasoning as well as output. ## Related terms - **Technical interview.** The wider term for any interview that tests technical skill. A screen is the first one. - **Coding interview.** An interview where the candidate writes code to solve a problem. - **Take-home test.** A task completed alone and submitted later. - **System design interview.** A longer round on how to build a large system, usually later in the loop. - **Pre-screening interview.** A shorter first conversation on the basics, before the technical screen. Which path fits you: - Technical roles: see [technical interviews](/use-cases/technical-interviews). - Many applicants: see [hiring at volume](/use-cases/volume-hiring). - Reading the output: see [AI interview reports](/blog/ai-interview-reports). ## Common questions ### What is a technical screen? An early interview that tests whether a candidate has the skills a technical role needs, before a longer technical loop. It ends in a decision to move on, hold or stop. ### What happens in a technical screening round? Usually fundamentals questions, a short coding or problem-solving task and a discussion of past work, with follow-ups on how the candidate approached the task. ### What is a technical phone screen? A technical screen run on a call, often with a shared editor. When an engineer runs it, it takes engineer time for each candidate. ### Who runs a technical screen interview? An engineer from the team, a recruiter with a script, an online assessment or an AI interviewer such as Tracy. The choice depends on how many candidates you have and how deep you need the follow-ups. ### Can an AI interviewer run a technical screen? Yes. Tracy runs a coding task inside the interview, with follow-ups on approach and trade-offs, and leaves a recording and written assessment per question for a person to review. --- See the method applied in one real interview. Send us one role and we will build the interview plan in days, with a recording and report you can check. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=what-is-a-technical-screen&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=what-is-a-technical-screen&utm_content=demo) ## Sources - [Tracy AI Interviewer home page, customer quotes](https://tracyinterview.ai/), retrieved 9 Nov 2026. - [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer Connect page](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. # What Is an Asynchronous Video Interview? > What is an asynchronous video interview? Candidates record answers to preset questions and the team watches later. Meaning, questions and limits. - URL: https://tracyinterview.ai/blog/what-is-an-asynchronous-video-interview - Published: 2026-11-09 - Author: Tracy team > **In short** Tracy AI Interviewer keeps what asynchronous interviews do well, since each applicant starts on click at any hour, and adds what a recording cannot: follow-ups on what the person said. > > An asynchronous video interview is one where the candidate records answers to preset questions in their own time and the team watches them later. > > The first 3 results are free, then $1 per credit, with no seats and no upfront cost. > > Next step: send us one role and we build the interview plan. ## What is an asynchronous video interview? An asynchronous video interview is an interview where the candidate records answers to preset questions in their own time, and the hiring team watches the recordings later. Nobody is on the other end while the candidate answers. It is also called a one-way video interview, and asynchronous interview software is the tool that sends the questions and collects the recordings. The word asynchronous describes timing: the two sides are not in the conversation at the same moment. In hiring the term also implies one-way: the candidate speaks and no one answers back. ## How an asynchronous video interview works 1. **The team writes the questions.** Usually three to six, each with a time limit for the answer. 2. **The candidate gets a link.** They open it when it suits them and need no scheduled slot. 3. **The candidate records each answer.** Many tools allow a practice question and a limited number of retakes. 4. **The team reviews later.** Reviewers watch the recordings, often at the same time as other applicants, and rate or comment on each one. The appeal is practical. There are no time zones to align, every candidate sees the same prompts, and reviewers watch when they have time. The cost shows up at review: someone still has to watch every recording. ![Glossary card for asynchronous video interview: the candidate records answers to preset questions and the team watches later.](/blog/images/what-is-an-asynchronous-video-interview-1.png) *The definition in one card, with a one-line example. Illustration with sample data* An asynchronous video interview is defined by recorded answers to preset questions and review later. ## Asynchronous video interview questions The questions are written to be answered once, in a minute or two, without a reply. These are typical. Typical asynchronous video interview questions, what each is for and why it suits a recording. | Purpose | Example question | Why it suits a recording | | --- | --- | --- | | Motivation | Why did you apply for this role? | A short, personal answer. | | Experience | Tell us about the work you did in your last role. | The candidate can structure it alone. | | Situation | Describe a time a plan changed at the last moment. | A story with a beginning and an end. | | Role basics | Which tools have you used for this work? | A list with a sentence of context. | Questions that need a back-and-forth fit a recording badly. If the honest next question is "what do you mean by that?", the format cannot ask it. ## What a one-way recording misses - **No follow-up.** A vague answer stays vague. The reviewer sees the gap and cannot ask about it. - **No clarification.** A candidate who misreads a question has no one to check with, and may be scored down for it. - **Review time grows with every applicant.** Someone has to watch each recording, so 200 applicants means 200 recordings to get through. - **Answers can be rehearsed.** With retakes, candidates can give a polished answer to a prompt they have seen once. The format still suits short, fixed answers, such as motivation and availability, when your team has time to watch. See [one-way vs two-way video interviews](/blog/one-way-vs-two-way-video-interviews) for when each fits, and [AI vs human interviews](/blog/ai-interview-vs-human-interview) for the wider comparison. ## When an asynchronous video interview is the right choice A recorded one-way interview is the better pick when you need short, fixed answers from a limited number of candidates and your team has the time to watch them. Examples are a motivation check, availability or a quick view of how someone presents themselves for a role with customer contact. It is a weaker pick when the decision depends on how a candidate reasons: a technical role, a specialist or any case where the first answer is only the start of the conversation. There the follow-up carries the information, and a recording does not have one. A middle path is to keep the asynchronous timing and add the follow-up. The candidate still answers when it suits them, but the questions respond to what they say, and the team reads a written assessment before it decides which conversations to review in full. ## Asynchronous in timing, two-way in content, with Tracy Tracy is asynchronous in the way that matters to a busy team: the applicant starts the interview on click at any hour, with no scheduling and no recruiter present. It is two-way in content, because she answers back. - **Follow-ups.** Tracy asks in the applicant's own field, follows up on what they said and waits for the answer. A vague answer gets a question. - **Questions from the job description.** Edit them or use your own. - **Review by evidence.** A recording of both sides, a timestamped transcript and a written assessment per question with its reasoning, so a reviewer reads a report before deciding what to watch. - **A browser link.** The applicant needs no account. - **A candidate experience that holds up.** Steven Perry, Account Executive at IT Resources, an IT staffing firm in Tampa, US, said: "By the time they finish with this platform, they're even laughing and joking." A person still decides from the evidence. See [AI interview candidate experience](/blog/ai-interview-candidate-experience) and [how AI interviews work](/blog/how-do-ai-interviews-work). The first 3 results are free with no card. After that it is $1 per credit, no seats, no upfront cost. See [pricing](/pricing). To see the difference on one role, create an interview and compare it with the recordings you collect today. ![Sample interview: Tracy asks why the candidate left, then what ownership would look like in the role they want.](/blog/images/what-is-an-asynchronous-video-interview-2.png) *A sample follow-up chain: a general answer gets a question about what the candidate wants, which a recorded one-way answer would leave unasked. Illustration with sample data* Where a recording ends at the answer, Tracy asks the next question. ## Related terms - **One-way video interview.** Another name for the same format. - **Two-way interview.** Both sides speak in the same session, live or not. - **Live video interview.** A scheduled call with a person on the other end. - **Asynchronous interview software.** The tool that sends questions and stores the recordings. - **Pre-screening interview.** The short first conversation, which an asynchronous interview can serve. Which path fits you: - Hiring at volume: see [hiring at volume](/use-cases/volume-hiring). - Screening applications: see [applications](/use-cases/applications). - Moving off a recorded-answers tool: see [VidCruiter alternatives](/compare/vidcruiter-alternatives). ## Common questions ### What is an asynchronous video interview? An interview where the candidate records answers to preset questions in their own time and the hiring team watches the recordings later. It is also called a one-way video interview. ### What does asynchronous video interview mean? Asynchronous means the two sides are not in the conversation at the same moment. The candidate answers when it suits them and the reviewer watches later. ### What are good asynchronous video interview questions? Short, self-contained prompts that can be answered once in a minute or two: motivation, experience, a situation and role basics. Avoid questions that need a back-and-forth. ### What is asynchronous interview software? The tool that sends the questions, lets candidates record answers and stores the recordings for review. Some add a rating form for reviewers. ### Can an asynchronous interview ask follow-up questions? A recorded one-way interview cannot. Tracy starts on click like an asynchronous interview but answers back, so a vague answer gets a follow-up. --- See the method applied in one real interview. Send us one role and we will build the interview plan in days, with a recording and report you can check. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=what-is-an-asynchronous-video-interview&utm_content=try-free) · [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=what-is-an-asynchronous-video-interview&utm_content=demo) ## Sources - [Tracy AI Interviewer home page, customer quotes](https://tracyinterview.ai/), retrieved 9 Nov 2026. - [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 9 Nov 2026. - [Tracy AI Interviewer Connect page](https://tracyinterview.ai/connect), retrieved 9 Nov 2026. # AI Interviews in ChatGPT: Set Up a Tracy Interview From a Chat > Install Tracy AI Interviewer in ChatGPT, describe who you want to interview, review the draft and read results in the same chat. Steps, prompts and limits. - URL: https://tracyinterview.ai/blog/tracy-chatgpt-integration - Published: 2026-11-05 - Updated: 2026-11-11 - Author: Tracy team > **In short** > - Describe the interview you want in ChatGPT and Tracy AI Interviewer drafts it. Candidates take a video interview with CV-based questions and follow-ups, and you get a recording, a timed transcript and integrity signals. > - ChatGPT returns a draft with the questions and result fields (Tracy is listed as Tracy at Qode), and nothing is created until you approve it. ChatGPT reaches Tracy through MCP, and teams can also connect through the API. > - Read the results later in the same chat by asking for them. ChatGPT gets one permission, named tracy, and acts only on your own account. > - Candidates install nothing. Your first 3 results are free. Fact-checked on 4 Nov 2026 against Tracy's own ChatGPT guide, which is listed under Sources. With Tracy AI Interviewer in ChatGPT, you set up and start interviews by describing them. Tracy can read the CV and tailor questions to it and to the role, follow up on what the person says and wait for the end of their answer, and leave a recording, a transcript timed to the recording and integrity signals with evidence for a person to review. ChatGPT reaches Tracy through MCP, and teams can also connect through the API. If you searched for a ChatGPT interview app, you may be looking for a way to practise your own answers. This page is for the other side of the table: someone who runs interviews and wants to set one up by describing it. If you are weighing what an AI interviewer is in the first place, start with [what an AI interviewer is and how it works](/blog/what-is-an-ai-interviewer). Tracy's [ChatGPT guide](https://tracyinterview.ai/help/chatgpt) puts it in one line: install Tracy AI Interviewer in ChatGPT, set up an interview, and read its results by asking for them in a chat. ## What you can do from the chat - Describe who you want to interview and what you need to learn. Tracy drafts the interview, and ChatGPT shows you the title, the questions and the result fields. - Ask for changes to the draft, then approve it. - Come back later and ask for the results. The interview is a two-way video interview. Candidates join from a link in the browser, and the person being interviewed installs nothing. The report that comes back is organised around the fields you chose, with the transcript and recording behind it, which [AI interview reports](/blog/ai-interview-reports) explains from a reviewer's side. ## Install it from ChatGPT Plugins 1. **Open the listing.** Use the link on the guide, or go to ChatGPT Plugins and search for "Tracy at Qode". The listing name is Tracy at Qode. 2. **Click Install plugin.** 3. **Continue to the sign-in and click Allow.** When ChatGPT asks to connect the plugin, click Continue, then complete the Tracy AI Interviewer sign-in there. 4. **Start a new conversation**, type @ in the prompt box and select Tracy at Qode. 5. **Send a starter prompt**, then review the draft. The draft comes back in the chat with the title, the questions and the result fields. Nothing is created until you approve it. ## Prompts to try These are the starter prompts on the guide: > Set up a screening interview for senior backend engineer candidates. > Interview 5 customers who cancelled last month about why they stopped using our product. > Let me try a practice video interview first, so I can see what my candidates will go through. The third one lets you take the interview yourself, so you see what a candidate sees before you invite anyone. ## What you approve, and what ChatGPT can do ChatGPT shows the interview as a draft before anything is created. Read it, ask for changes and say yes when it is right. When ChatGPT opens the Tracy AI Interviewer sign-in, you click Allow to approve access. That asks for one permission, named tracy, and ChatGPT acts only on your own account. You can cut ChatGPT off at any time: open your account, then Connected AI apps, and revoke it. That check is also how you confirm the connection worked, because the app is listed there with the time you connected it. ## If you cannot find it: add it as a custom MCP server Use this if Tracy at Qode does not show up in ChatGPT Plugins for you. You need ChatGPT on the web. Account and workspace policies apply, so on a work or school workspace an admin may need to allow custom MCP servers first. If you do not see Add custom MCP server, turn on Developer mode in your ChatGPT settings. 1. Go to ChatGPT Plugins, select Add, then Add custom MCP server. 2. Enter the name `Tracy at Qode` and paste `https://tracyinterview.ai/mcp` as the Server URL. 3. For Authentication, choose OAuth. 4. Review the risk warning and select "I understand and want to continue". 5. Select Create as a plugin, install it and start a new conversation. Finish the Tracy sign-in when ChatGPT opens it. ## Limits - Account and workspace policies apply to adding and using custom MCP servers, and a work workspace may need an admin first. - If ChatGPT is not listed under Connected AI apps in your account, the sign-in was not finished. Start a new conversation, select Tracy at Qode with @ and finish it. - Tracy's home page notes that listing a tool there does not mean a partnership or an endorsement. If your candidates live in Workable, the same chat can bring them in: see [AI interviews with Workable](/blog/tracy-workable-integration). If your team talks in Slack, [Tracy also works there](/blog/tracy-slack-integration). ## What it costs Tracy's first 3 results are free. After that one credit covers each started 20 minutes of an interview that produced a result, at $1 per credit. A session with no answers is never charged. The [AI interview cost guide](/blog/ai-interview-cost) sets that beside other tools' published prices. ## Common questions ### Is this an app for practising my own job interview? No. Tracy AI Interviewer is for the person running interviews. It sets up an interview, invites participants and writes up what they said. One starter prompt lets you try a practice video interview first, so you can see what your candidates will go through. ### Will ChatGPT create an interview without my approval? No. ChatGPT shows the interview as a draft, and nothing is created until you approve it. ### What permission does ChatGPT get? One permission, named tracy. It acts only on your own account, and you can revoke it under Connected AI apps in your Tracy account. ### What if Tracy at Qode is not in ChatGPT Plugins for me? Add it as a custom MCP server with the server address `https://tracyinterview.ai/mcp`, the name Tracy at Qode and OAuth. If you do not see Add custom MCP server, turn on Developer mode, and on a work or school workspace an admin may need to allow custom MCP servers. ### Do my candidates need ChatGPT? No. ChatGPT is for the person running the interview. The person being interviewed installs nothing and joins from a link in the browser. --- **[Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=tracy-chatgpt-integration&utm_content=try-free) with your first 3 results on us, or [book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=tracy-chatgpt-integration&utm_content=demo) to see it run.** ## Sources All sources retrieved 4 Nov 2026. - Tracy AI Interviewer, [Add Tracy AI Interviewer to ChatGPT](https://tracyinterview.ai/help/chatgpt): install steps, starter prompts, draft before creation, the one permission, custom MCP server steps, troubleshooting. - Tracy AI Interviewer, [Help](https://tracyinterview.ai/help): revoking a connected AI app under Connected AI apps. - Tracy AI Interviewer, [Pricing](https://tracyinterview.ai/pricing): first 3 results free, $1 per credit, one credit per started 20 minutes, no charge for a session with no answers. - Tracy AI Interviewer, [Home](https://tracyinterview.ai/): evidence behind every result, each person starts in the browser, CV check adds 3 to 5 questions from each CV, integrity signals with their evidence, the note that a listed tool is not a partnership or an endorsement. - Tracy AI Interviewer, [Product](https://tracyinterview.ai/product): no install for the person being interviewed, ready in under 10 seconds at any hour, recorded on both sides with a transcript timed to the recording, CV check, three ways to start an interview through the API or MCP. # AI Interviews in Slack: Check, Draft and Invite With Tracy > Add Tracy AI Interviewer to Slack to list your interviews, check where results stand, draft an interview and invite people, with an Approve step before anything is sent. - URL: https://tracyinterview.ai/blog/tracy-slack-integration - Published: 2026-11-05 - Updated: 2026-11-11 - Author: Tracy team > **In short** > - Draft a Tracy AI Interviewer interview in Slack, by message or @mention, and invite people by email. Each person takes a video interview with follow-ups, and you get a recording, a transcript timed to it and integrity signals. > - You can also list your interviews and see where results stand. > - She shows a preview first and creates or sends nothing until you press Approve. In a channel, only one short line is visible to others; the answer goes to your direct message. > - You add her from Tracy's own page, not Slack's app directory. Your first 3 results are free. Fact-checked on 4 Nov 2026 against Tracy's own Slack guide, which is listed under Sources. With Tracy AI Interviewer in Slack, you start interviews from the workspace you already use. The interview itself is a two-way video interview in which Tracy can read the CV and tailor questions to it, follow up on what the person says, and leave a recording, a transcript timed to the recording and integrity signals with evidence for a person to review. Teams can also connect Tracy through the API and MCP. "AI interview in Slack" can mean two things. One is an interviewer that talks to candidates inside Slack. Tracy does not do that: each interview is a two-way video interview, and each person joins from a link in the browser with nothing to install. The other is a way to run your interviews from Slack, and that is what the Slack app does. You message Tracy, she answers, and she acts on your Tracy account. ## What you can ask her Tracy's [Slack guide](https://tracyinterview.ai/help/slack) lists these, all in plain words: - List your interviews. - See who is on an interview and where their results stand. - Check how many free results and credits you have left. - Draft an interview. - Invite people by email. Typing `/tracy help` in Slack shows the commands that do the same. Three questions the guide suggests to start: "Which interviews do I have?", "Who is on my customer success interview, and where do their results stand?" and "How many free results and credits do I have left?" Drafting and inviting both stop at a preview. Nothing is created or sent until you press Approve. The guide lists checking where results stand and optional alerts when a result is ready. Reading a full report inside Slack is not on that list. ## Add Tracy to Slack You need a Tracy account, and a Slack workspace you can add apps to. If your workspace asks an admin to approve new apps, ask one to approve Tracy. You do not search Slack's app directory for her. The Add to Slack button on the guide is where it starts, and if you are signed out it takes you to sign in or sign up first. 1. **Click Add to Slack** on the guide. Slack opens and asks which workspace to add Tracy to. 2. **Read what Slack says Tracy will do**: use slash commands, send messages as Tracy, read the direct messages sent to her and see the messages that mention her. Then click Allow. 3. **Land back on the Integrations page**, where the Slack card says the Tracy bot was added. Adding her also links your Slack account to your Tracy account, so nothing more needs connecting. 4. **Open Slack and find Tracy under Apps** in the sidebar. She sends a first direct message that starts "Hi, I'm Tracy, an AI interviewer". 5. **Ask her something**, such as "Which interviews do I have?" She answers from your Tracy account, or tells you there are none yet. ## Use her in a channel Add her to the channel first: type `/invite @Tracy`, or open the channel's name, choose Integrations and add her from Apps. Then @mention her with a question, such as "@Tracy how many credits do I have left?" Under your message she replies with one short line: "I replied to you in a direct message." The answer, the Connect Tracy button if you have not connected yet, and any approval preview all go to your direct message. A message in the channel that does not mention her is never read. Tracy reads the direct messages sent to her and the channel messages that @mention her, plus the `/tracy` commands you type. She posts to the person she is talking to. In a channel she adds one short line under your message that mentions her. Apart from that line, the only other place she posts is an alerts channel, and only if you set one up. If she does not answer a mention, check she was added to that channel. If you added Tracy to Slack before she answered mentions, click Add to Slack again and Allow, which grants the newer app_mentions:read permission. Whoever manages the Slack app must also subscribe it to the app_mention event, under Event Subscriptions and Subscribe to bot events, and then reinstall it. ## Bring in teammates A teammate in the same workspace does not add the app again. They send Tracy a direct message, and she replies with a Connect Tracy button. The link is good for 10 minutes. They open it themselves, sign in to their own Tracy account and click Link my account. The guide asks them not to pass the link on, because it is for the person who asked for it. Typing `/tracy connect` gives a fresh link if it expired. ## Optional alerts channel Open your account and use the Slack alerts panel to pick one channel. Tracy posts there when someone opens an invite, starts, finishes or leaves an interview, and when a result is ready. It is separate from talking to Tracy, so skip it if you only want the conversation. ## What this suits, and what it does not It suits a recruiter or hiring manager who lives in Slack and wants quick answers: who has finished, how many credits are left, send these three people the interview. If you recruit for client companies, [AI interview tools for staffing agencies](/blog/ai-interview-tools-for-staffing-agencies) compares the wider options. It does not replace the interview itself, and it does not show a full report in the chat. If you would rather set an interview up by describing the role in a chat window, [ChatGPT does that](/blog/tracy-chatgpt-integration), and [Workable users can invite shortlisted candidates](/blog/tracy-workable-integration) from the same kind of chat. For what happens after the invite, from the candidate's side and the host's, read [how AI interviews work](/blog/how-do-ai-interviews-work). ## What it costs Tracy's first 3 results are free. After that one credit covers each started 20 minutes of an interview that produced a result, at $1 per credit. A session with no answers is never charged. Ask Tracy in Slack how many free results and credits you have left, and see the [AI interview cost guide](/blog/ai-interview-cost) for how that compares with other tools. ## Common questions ### Does the interview itself happen in Slack? No. Tracy in Slack is for the person running the interviews: checking them, drafting one and inviting people. Interviews are two-way video interviews, and each person joins from a link in the browser. ### Can Tracy read my whole Slack workspace? No. Tracy reads the direct messages sent to her and the channel messages that @mention her. In a channel she does not read other conversations. ### What does my team see when I ask Tracy in a channel? One short line under your message. The answer, the Connect Tracy button and any approval preview go to your direct message. ### Will Tracy send invites without asking? No. She shows a preview first, and nothing is created or sent until you press Approve. ### How does a teammate connect? They send Tracy a direct message and press Connect Tracy. The link is good for 10 minutes, and they sign in to their own Tracy account and click Link my account. ### Can I read a full interview result in Slack? The Slack guide lists checking where results stand and optional alerts when a result is ready. It does not list reading a full report in Slack. --- **[Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=tracy-slack-integration&utm_content=try-free) with your first 3 results on us, or [book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=tracy-slack-integration&utm_content=demo) to see it run.** ## Sources All sources retrieved 4 Nov 2026. - Tracy AI Interviewer, [Add Tracy AI Interviewer to Slack](https://tracyinterview.ai/help/slack): what she can do, the permissions Slack shows, steps, channel behaviour, teammate linking, alerts, what she reads and posts. - Tracy AI Interviewer, [Pricing](https://tracyinterview.ai/pricing): first 3 results free, $1 per credit, one credit per started 20 minutes, no charge for a session with no answers. - Tracy AI Interviewer, [Home](https://tracyinterview.ai/): each person starts in the browser, at any hour; CV check adds 3 to 5 questions from each CV; transcript anchored to the recording; integrity signals with their evidence; teams can also connect through the API. - Tracy AI Interviewer, [Product](https://tracyinterview.ai/product): no install for the person being interviewed, ready in under 10 seconds at any hour, recorded on both sides with a transcript timed to the recording, CV check, three ways to start an interview through the API or MCP. - Tracy AI Interviewer, [Connect](https://tracyinterview.ai/connect): the MCP server connection. # AI Interviews With Workable: Invite Candidates to Tracy From a Chat > Tracy AI Interviewer runs beside Workable's official connector in Claude or ChatGPT. Setup steps, what Workable shares with Tracy, and what stays in Tracy. - URL: https://tracyinterview.ai/blog/tracy-workable-integration - Published: 2026-11-05 - Updated: 2026-11-11 - Author: Tracy team > **In short** > - Invite your Workable shortlist to a Tracy AI Interviewer video interview from Claude or ChatGPT. Tracy can check each CV, asks tailored questions with follow-ups, and leaves a recording, a timed transcript and integrity signals to review. > - Tracy connects through MCP, teams can also connect through the API, and you approve a preview before anything is emailed. > - Workable shares each candidate's name and email. Results stay in Tracy, and you ask the chat to read them from there. > - Setup is Tracy plus Workable's official connector, and one test question. Your first 3 results are free. Fact-checked on 4 Nov 2026 against Workable's help pages and Tracy's help pages, all listed under Sources. With Tracy AI Interviewer connected to Workable, you can start an interview for every shortlisted candidate from the chat you already use. Tracy can read each candidate's CV and tailor questions to it and to the role, ask follow-ups on what the person says and wait for the end of their answer, and leave you a recording, a transcript timed to the recording and integrity signals with evidence for a person to review. The link between the two tools is an MCP connection, and teams can also connect to Tracy through the API. If you searched for a Workable AI interview integration, you may have expected an app in Workable's marketplace that adds an interview stage to your pipeline. Tracy's [Workable guide](https://tracyinterview.ai/help/workable) describes a chat that talks to both products. You keep your candidates in Workable, add Tracy and Workable's connector to Claude or ChatGPT, and ask for the candidates and the interview in the same conversation. Workable calls its connector the Workable MCP server. ## What you can do with it - Ask which jobs are open in Workable and list the candidates shortlisted for a job. - Invite those candidates to an interview you have already set up in Tracy. - Ask which of the candidates you invited have finished, and have the chat read and summarise their results from Tracy. The interview is a two-way video interview. Each candidate joins from a link, and the person being interviewed installs nothing. To see what the conversation looks like before you invite anyone, [how AI interviews work](/blog/how-do-ai-interviews-work) walks through both sides. What a candidate can read before they join is [How to prepare for an AI interview](/blog/how-to-prepare-for-an-ai-interview?utm_source=blog&utm_medium=seo&utm_campaign=tracy-workable-integration&utm_content=internal-link). ## What you need - A Tracy AI Interviewer account. Your first 3 results are free and no card is needed. - A Workable account. Workable says its MCP server is included on all plans at no additional cost. Some HR tools need the matching Workable products, and the chat can only see and do what your own Workable user can. - Claude or ChatGPT. Workable's help has a page for each, and both use its official connector. - An interview already set up in Tracy. If you have none, ask the chat to draft one first; [setting an interview up from ChatGPT](/blog/tracy-chatgpt-integration) shows how. In a managed workspace, an administrator's connector policy may block a connector. Workable's help says to ask the administrator to enable it, then connect with your own Workable login. ## Set it up in four steps 1. **Add Tracy to your AI app.** Follow the guide for [Claude](https://tracyinterview.ai/help/claude) or [ChatGPT](https://tracyinterview.ai/help/chatgpt). Tracy is listed there as "Tracy at Qode". 2. **Add Workable to the same app, from the official directory.** These are Workable's own steps, taken from Workable's current help pages, which are the source for every Workable step on this page. - In Claude, open Settings > Customize > Connectors, search for Workable and select Connect. Sign in to Workable, review the permissions and authorize. Workable says no separate API key is needed. - In ChatGPT on the web, open Settings > Plugins, search for Workable and toggle on the official Workable MCP Server connector. Sign in to Workable, review the permissions and authorize. Workable says the official connector handles the server configuration and sign-in. In a chat, type @ and select Workable. 3. **Test Workable alone.** Start a new chat and ask Workable's own test prompt: "Use Workable to list the accounts I can access. Call get_accounts first, then use my chosen account's subdomain as the account argument on subsequent calls. Show the open jobs in that account." If you can reach more than one Workable account, say which one to use. 4. **Ask for both in one chat.** Start a new chat with Tracy at Qode and Workable both turned on, and use a prompt like the ones below. If your app shows different names from the ones above, follow what it shows. ## Prompts to try These come from Tracy's guide. Each one assumes the interview exists in Tracy. > List the candidates shortlisted for our Customer Success Manager job in Workable. Then invite them all to my customer success interview in Tracy AI Interviewer with invite_participants, and show me the preview before anything is sent. > Take the newest candidate in Workable for our Customer Success Manager job and invite them to my customer success interview in Tracy AI Interviewer. Show me the preview first. > Which of the candidates I invited from Workable have finished my customer success interview? Read their results from Tracy AI Interviewer and summarise each one. ## What moves between Workable and Tracy | What | Where it lives | How it works | | --- | --- | --- | | Candidate name and email | Workable | Workable shares both with the chat. Tracy needs them to send an invite. | | The invite | Tracy | Tracy shows who it would invite and waits for your approval before anything is emailed. | | The interview | Tracy | The candidate joins from a link and installs nothing. | | Results | Tracy | Results stay in Tracy. You ask the chat to read them from there. | | Stages, comments and the candidate record | Workable | Tracy's guide describes no step that changes them. Workable's connector can, if you ask the chat. | ## Does anything go back into Workable? Tracy's guide says results stay in Tracy AI Interviewer and gives no step that sends them to Workable. Workable's help lists tools to add comments with visibility controls or ratings, move candidates between pipeline stages and update a candidate's profile fields. It describes the connector as able to both read and write. So a chat that has read a result from Tracy could, if you asked, add a comment to the candidate in Workable. The Workable connector would do that, not a step in Tracy's guide, and Workable tells you to review the record and the requested change before you approve anything that updates Workable. ## When this is not the right setup Pick a different route if you need any of these: - **Results that arrive on the candidate's Workable record by themselves.** Tracy's guide has no step for it, so a reviewer who lives in Workable has to open Tracy or ask the chat. - **A pipeline stage that sends the invite on its own.** Here a person asks in a chat each time and approves the preview. - **A team that does not use Claude or ChatGPT.** The setup depends on one of them. If you want a first round that reads like an interview, with follow-up questions asked of each candidate, and you are comfortable with a person starting each batch from a chat, this setup keeps your candidates where they are. If you are weighing it against an interviewer built into your applicant tracking system, [the AI interviewer in your ATS versus a dedicated one](/blog/ats-ai-interviewer-vs-dedicated) sets out what each covers. If your team works mostly in Slack, [Tracy has an app there](/blog/tracy-slack-integration) too. ## What it costs Tracy's first 3 results are free. After that one credit covers each started 20 minutes of an interview that produced a result, at $1 per credit. A session with no answers is never charged. The [AI interview cost guide](/blog/ai-interview-cost) puts that next to other tools' published prices. ## Common questions ### Is Tracy a Workable marketplace app? Tracy's guide connects Tracy and Workable's own MCP server through Claude or ChatGPT. It does not describe a Workable marketplace app. You add both to the same account and ask for both in one chat. ### Do interview results go back into Workable? Tracy's guide says results stay in Tracy AI Interviewer and describes no step that sends them to Workable. Workable's connector can add comments and move candidates, so you could ask the chat to do that yourself. Review the change before you approve it. ### What does Workable share with Tracy? Each candidate's name and email, which is what Tracy needs to send an invite. ### Does it cost extra to connect Workable? Workable says its MCP server is included on all plans at no additional cost. Tracy's first 3 results are free, then it is $1 per credit. ### Will Tracy email candidates without my approval? No. Tracy shows you who it would invite and waits for your approval before anything is emailed. ### Do I need a custom connector for Workable? Workable's own steps do not use one. In Claude you pick Workable from the official connector directory, and in ChatGPT you switch on the official Workable connector in Plugins. --- **[Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=tracy-workable-integration&utm_content=try-free) with your first 3 results on us, or [book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=tracy-workable-integration&utm_content=demo) to see it run.** ## Sources All sources retrieved 4 Nov 2026. - Tracy AI Interviewer, [Use Tracy AI Interviewer with Workable](https://tracyinterview.ai/help/workable): chat-based setup, the prompts, name and email shared, preview before invites, results stay in Tracy. - Workable, [Connecting Workable to Claude](https://help.workable.com/hc/en-us/articles/44379447823639): official connector directory, steps, permissions follow the signed-in user, all subscription tiers, review before an update. - Workable, [Connecting Workable to ChatGPT](https://help.workable.com/hc/en-us/articles/44379407715351): Settings > Plugins steps, the official connector handles configuration and sign-in, @ to select Workable. - Workable, [Using the Workable MCP server to connect to AI assistant](https://help.workable.com/hc/en-us/articles/39415735712407): what the connector can read and write, including comments, ratings, stage moves and profile fields. - Workable, [Workable MCP server](https://resources.workable.com/workable-mcp-server): included on all plans at no additional cost. - Tracy AI Interviewer, [Add Tracy AI Interviewer to Claude](https://tracyinterview.ai/help/claude): Tracy listed as Tracy at Qode. - Tracy AI Interviewer, [Add Tracy AI Interviewer to ChatGPT](https://tracyinterview.ai/help/chatgpt): the listing name Tracy at Qode. - Tracy AI Interviewer, [Pricing](https://tracyinterview.ai/pricing): first 3 results free, $1 per credit, one credit per started 20 minutes, no charge for a session with no answers, no card needed. - Tracy AI Interviewer, [Connect](https://tracyinterview.ai/connect): the MCP server connection. - Tracy AI Interviewer, [Product](https://tracyinterview.ai/product): two-way video interview, teams can connect through the API or MCP, no install for the person being interviewed, recording with a transcript timed to the recording, integrity signals for a person to review.