# ChatGPT prompts for hiring managers, from scorecard to interview

> 16 ChatGPT prompts for hiring managers: scorecard, questions and feedback, then hand the scorecard to Tracy to interview every applicant.

- URL: https://tracyinterview.ai/blog/chatgpt-prompts-for-hiring-managers
- Published: 2026-11-17
- Author: Annie

> **In short** ChatGPT drafts a hiring scorecard, the questions and the feedback quickly, so you can say exactly what the role needs. Tracy AI Interviewer takes that scorecard, interviews every applicant on its criteria with follow-ups in their own field, and returns a recording, a timestamped transcript and what each person said to each question. You mark each criterion and decide.

If you are a hiring manager, your recruiter will send you a stack of CVs and expect an opinion by Friday. These ChatGPT prompts for hiring managers help you do the part only you can do: say exactly what the role needs, ask the same questions of everyone and write feedback someone else can act on. One sample output is shown, and the middle group takes your scorecard and gives it to [Tracy AI Interviewer](/blog/tracy-chatgpt-integration) to run as an interview.

## What do ChatGPT prompts for hiring managers do well?

They save the blank page. ChatGPT turns your answers into a scorecard, questions and feedback in minutes, from the facts you give it. Give it the facts a recruiter would ask for: the problem this hire solves, the team, what the last person did well, and what you would regret. Say "ask me questions before you answer" and reply to each. Keep names, ratings and anything private about people out of the chat unless your company allows it.

Four criteria on your scorecard can turn into four questions, each with follow-ups, asked of every applicant. That step is where Tracy comes in.

![Four inputs: the problem the hire solves, what the last person did well, what you would regret, who they work with.](/blog/images/chatgpt-prompts-for-hiring-managers-1.png)

*What a hiring manager gives ChatGPT first so the scorecard is specific. Illustration.*

Four facts to give ChatGPT before it writes a scorecard.

## What should a hiring manager ask ChatGPT first? (prompts 1 to 3)

Start with the problem the hire solves, then sort the skills and list what you would regret.

### 1. State the problem first
> I am hiring a [role] for a team of [size]. Ask me 8 questions, one at a time, to find out the problem this person must solve in their first 6 months. Then write it in three sentences.
### 2. Must-have or nice-to-have
> Here are the skills I listed for [role]: [paste]. Sort them into must-have and nice-to-have, and for each must-have say what I would see in an interview if the person really had it.
### 3. What I would regret
> Based on the role above, list 5 ways this hire could go wrong in 6 months, and a question that would show each one early.

## Can ChatGPT write a hiring scorecard? (prompts 4 to 6)

Yes, as a first draft. Prompt 4 asks for evidence per criterion and no scores; you edit it and decide how to mark each one.

### 4. Scorecard from the problem
> Write a scorecard for a [role]. For each of 4 criteria give a one-line definition, what evidence in an interview would show it, and a sign it is missing. Do not give weights or scores.
An example scorecard, written for illustration, for a fictional senior backend engineer.
**Sample scorecard: senior backend engineer**
| Criterion | Evidence that shows it | Sign it is missing |
| --- | --- | --- |
| Designs a service API | Explains a versioning choice and who it affected | Describes endpoints without trade-offs |
| Debugs in production | Walks through a real incident: signals, steps, fix | Says "I would check the logs" and stops |
| Owns reliability | Names a past alert they changed and why | Blames the platform team |
| Explains to non-engineers | Gives a clear account of a technical decision to a product manager | Uses jargon without checking understanding |
### 5. Tighten the wording
> Rewrite each "evidence" line in the scorecard so that two interviewers would mark it the same way. Remove adjectives like "strong" and "good".
### 6. Check the scorecard for bias risk
> Read this scorecard and flag any line that rewards background, accent or style over what the person can do. Suggest a replacement that checks the skill instead.

## Questions (7 to 9)

### 7. One question per criterion
> For each criterion in the scorecard, write one question that asks for a specific past situation, and two follow-ups that ask for the missing detail.
### 8. Follow-ups for a thin answer
> The candidate answered "[question]" with "[answer]". Write three follow-up questions that ask what they did themselves, what went wrong and what they would change.
### 9. A work sample brief
> Write a 45-minute work sample for [role] that checks criterion 2 and criterion 3 only. Include what a good result looks like and what to ignore.

## How do you hand the scorecard to Tracy? (prompts 10 to 12)

Once Tracy is connected in ChatGPT (type @ and select Tracy AI Interviewer), prompt 10 turns your scorecard into an interview. ChatGPT shows the draft as a preview and creates nothing until you approve it.
### 10. Scorecard to interview
> @Tracy AI Interviewer set up a screening interview for [role] candidates. Cover these four criteria, using the questions I wrote below, and keep my wording exactly: [paste]. Show me the draft first.
### 11. Choose the style
> Change the draft so the interview is about 15 minutes and goes deeper on criterion 2. Show me the preview again.
### 12. Read the evidence by criterion
> For each person who finished, list what they said under each of my four criteria with timestamps, and note anything the interview did not cover. Do not score or rank anyone; I will mark the scorecard myself.

![Four steps: write the scorecard, Tracy interviews, read the evidence, you mark and decide.](/blog/images/chatgpt-prompts-for-hiring-managers-2.png)

*Prompts 10 to 12: your scorecard becomes a Tracy AI Interviewer interview and you mark each criterion from the evidence. Illustration.*

The scorecard stays yours; Tracy supplies evidence per question.

## What does a scorecard need before it can decide anything?

A scorecard with no evidence is a row of empty boxes. Take the sample above: for "Debugs in production" the evidence that shows it is a real incident walked through, and the sign it is missing is "I would check the logs" and a stop. Someone has to ask the follow-up that tells the two apart.

Four criteria, one question each and two follow-ups each (prompt 7) make 12 asks per candidate. For 100 applicants at 20 minutes, the interviews fill about 33 hours of live calls.

Tracy asks the four questions of every applicant, with follow-ups where an answer stays thin. Each box on your scorecard can then point to a quote, and you mark it yourself.

![Sample interview: an applicant says they would check the logs, and Tracy asks which signal told them where to look.](/blog/images/chatgpt-prompts-for-hiring-managers-3.png)

*Tracy asks which signal told the applicant where to look, so criterion 2 has a quote to point to. Illustration with sample content.*

A vague answer gets a follow-up before you mark the box.

## What Tracy adds to your scorecard

You set the skills Tracy covers for the role, and the interview starts from your questions or from the job description.

01

### Reach every applicant

Each applicant starts on click, at any hour, with no scheduling and no recruiter on the call.

02

### Ask the next question

Tracy follows the answer in the applicant's own field and waits for it.

03

### Keep the evidence

Each finished interview comes with the recording, a timestamped transcript and a written assessment per question with its reasoning, so each box on your scorecard can point to a quote.

You mark the criteria and make the call. See how the pieces fit in the [interview scorecard template](/blog/interview-scorecard-template) and [what a structured interview is](/blog/what-is-a-structured-interview).

## Feedback (13 to 14)

### 13. Feedback from my notes
> Turn my interview notes into feedback for the recruiter: for each criterion, what the candidate said, how it compares with the scorecard, and what is still unknown. No recommendation. Notes: [paste]
### 14. Rewrite vague feedback
> My feedback says "not a culture fit." Ask me what I actually saw, then rewrite it as specific evidence a recruiter could check.

## Decision (15 to 16)

### 15. Compare two finalists on the scorecard
> Here is my evidence for [A] and [B] against each criterion: [paste]. Lay out where each is better supported and where I have no evidence. Do not choose for me.
### 16. Message after the decision
> Write a short message to the recruiter with my decision on [name], the two reasons from the scorecard and the next step I want. I will edit it.

## Which path fits you

- A reusable scorecard layout: [interview scorecard template](/blog/interview-scorecard-template).
- Build the whole setup in Claude: [Claude hiring manager skill](/blog/claude-hiring-manager-skill).
- Capturing what each interviewer saw: the [interview feedback form template](/blog/interview-feedback-form-template).
- Hiring many people for the same role: [hiring at volume](/use-cases/volume-hiring).

## Common questions

### What are good ChatGPT prompts for hiring managers?

A scorecard from the problem, one question per criterion with follow-ups, and feedback from your notes. The 16 above follow a hire from intake to decision.

### Can ChatGPT write a scorecard?

Yes, as a first draft. Prompt 4 asks for a definition, evidence and a missing sign per criterion with no scores. You edit it and decide how to mark each one.

### Can ChatGPT run the interview?

Not by itself. Prompt 10 passes your scorecard to Tracy AI Interviewer, which interviews candidates on video and returns the evidence. You approve the preview, and ChatGPT sets it up.

### Will Tracy score my candidates?

No. Tracy reports what each person said, with a recording, a transcript and the answer to each question. You mark your scorecard.

### How do I keep feedback fair?

Write it against the same criteria for everyone, quote what the person said and leave out words like "fit". Prompt 14 helps.

---

**Turn your scorecard into an interview** Connect Tracy in ChatGPT, run prompt 10 on one role and read the evidence per criterion. The first 3 results are free. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=chatgpt-prompts-for-hiring-managers&utm_content=try-free) [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=chatgpt-prompts-for-hiring-managers&utm_content=demo)

## Sources

- [Add Tracy AI Interviewer to ChatGPT (public guide)](https://tracyinterview.ai/help/chatgpt), retrieved 14 Nov 2026.
- [Tracy AI Interviewer home page](https://tracyinterview.ai/), retrieved 11 Nov 2026.
