# Run your UX interviews from the ChatGPT chat you already use

> ChatGPT for UX research drafts the plan and reads the notes. Add Tracy to run the user interviews from the same chat, with a preview to approve first.

- URL: https://tracyinterview.ai/blog/chatgpt-for-ux-research
- Published: 2026-11-18
- Author: Annie

> **In short** ChatGPT for UX research is good at the prep and the first read: plans, guides, screeners and note summaries. It cannot interview your users. Add Tracy AI Interviewer to the chat and each interview is previewed, approved and then run on video, with follow-ups in the participant's own field. Every result carries a recording and transcript, and the researcher decides what the findings mean.

## What ChatGPT for UX research does well

ChatGPT drafts, restructures and summarises, which covers most of the desk work around a study. This page follows one study, a check of the checkout on a home-goods store, from prompt to readout. Good prompts for the desk work:

- "Turn this question into a research plan with one decision it informs."
- "Write eight discussion-guide questions about returning to checkout after a failed payment. Flag any that lead the participant."
- "Write a screener for shoppers who bought home goods online in the last three months."
- "Here are my notes. List what each participant said about delivery cost, with a quote for each."

All four work on text you supply. None of them gives you an answer from a real shopper. ChatGPT can plan the checkout study and read the notes afterwards, but the step in the middle, asking a shopper what stopped them, is the one it cannot take. Tracy AI Interviewer, added as a plugin, takes it.

## How do you add Tracy to ChatGPT?

Use ChatGPT on the web. OpenAI's help article lists custom MCP servers for ChatGPT Business and Enterprise/Edu on the web, and says you do not need to turn on Developer Mode first in the current Plugins experience. Its pricing page, read 15 Nov, lists Developer mode (beta) on Plus and Pro and not on Go. Check what your plan and workspace allow. If you do not see the option to add a custom MCP server, check whether Developer mode is on for your workspace.

1. Open **Tracy AI Interviewer** in ChatGPT's Plugins and click **Install plugin**.
2. Finish the Tracy sign-in and click **Allow**.
3. Start a new conversation, type **@** and select Tracy AI Interviewer.
4. Ask for the interview.

The connector asks for one permission, named tracy, and acts only on your own Tracy account. OpenAI says write actions require confirmation by default. The full walkthrough is in the [ChatGPT setup guide](/blog/tracy-chatgpt-integration), and [a short explainer on MCP servers](/blog/what-is-an-mcp-server) covers the connection.

## What does ChatGPT show before an interview exists?

A preview. Ask in plain words:

> @Tracy Set up a video interview for shoppers who abandoned checkout in the last month. I want to learn what they were buying, what stopped them and what they did next. Five questions. Show the preview and do not invite anyone.

ChatGPT answers with the opening line, each question and the field it reports under. The screenshot below shows that preview for a practice interview. You change what you want in the same chat, then approve it. Approving creates the interview with nobody invited. An invitation is previewed as well, and is not sent until you approve it. A new interview starts with its public link on (anyone with the link can join, up to 20 people for 14 days), and you can turn the link off.

![ChatGPT chat: @Tracy asked to draft a video practice interview; preview shows opening line, three questions, report fields.](/blog/images/chatgpt-for-ux-research-1.webp)

*Screenshot from the Tracy AI Interviewer ChatGPT guide: the @Tracy request and the preview with the opening line and three questions, each with the report field it feeds. Source: Tracy AI Interviewer ChatGPT guide, read 9 Nov 2026 (re-read 14 Nov).*

ChatGPT lays out the interview as a preview, with each question and the report field it feeds.

## What Tracy adds to ChatGPT for UX research

1. Reach every participant you invite. Each person starts on click, at any hour, with no scheduling and no researcher on the call. Your sample stops depending on calendars.
2. Ask the next question. Tracy follows the answer in the participant's own field and waits for it. "It was confusing" gets a question about which step.
3. Keep the evidence. Each result has a recording, a timestamped transcript and a written assessment per question with its reasoning, so a quote in your readout traces to a moment.

The same run works at volume: 100K+ interviews delivered, with a reported 99% completion rate.

![Four icon steps: draft in ChatGPT, approve the preview, Tracy interviews on video, read the results.](/blog/images/chatgpt-for-ux-research-2.png)

*ChatGPT drafts the plan, you approve the preview, Tracy interviews on video, and ChatGPT helps you read the results. Illustration of the workflow.*

ChatGPT drafts and reads, you approve the preview, and Tracy runs the interview.

## What "it was confusing" leaves open

A shopper says the checkout was confusing. That is a note, not a finding. The useful part is the next question: which step lost them, and what did they do next. A researcher in a live session asks it. Across ten shoppers, each 20-minute session is about 3 hours of calls in total, and the follow-up has to be asked well every time.

Tracy asks it in each interview, so you read ten answers to the same follow-up instead of booking ten calls. A theme in the readout then has a quote behind it and a moment in a recording.

![A shopper says checkout was confusing; Tracy asks which step lost them and what they did next.](/blog/images/chatgpt-for-ux-research-3.png)

*Tracy turns "it was confusing" into the step that lost the shopper. Illustration with sample content. Sample content, not a real shopper.*

A general complaint becomes a specific step once Tracy asks what happened next.

## Do ChatGPT and Claude differ for UX research?

The research does not change, and the setup differs. Both can hold the Tracy connector, and both ask before they use it. In Claude you add a custom connector under Customize, Connectors, and Anthropic lists it on Free through Enterprise, with one connector allowed on Free. In ChatGPT you install the plugin and call it with @, and your workspace settings decide whether you can add it. Choose Claude when your team already works there or wants to try one custom connector on a Free plan. Choose ChatGPT when your team already works there and your workspace lets you add a custom MCP server (OpenAI names Business and Enterprise/Edu; its pricing page lists Developer mode on Plus and Pro). Check your plan first. The [Claude version of this page](/blog/claude-for-ux-research) covers the same study there.

## What should you check in the results?

Results arrive as people finish, so you can read two or three before you decide whether to change a question. Read them as a first pass and verify against the source. Each result in Tracy has a recording, a timestamped transcript and what the person said to each question, so open the recording at a timestamp before a quote goes in a readout. Tracy reports what people said. It does not score or rank participants, and you decide what the answers mean. For interviews you already hold, [coding transcripts in ChatGPT](/blog/chatgpt-for-qualitative-research) covers where its output needs a check.

Pricing follows the web app: $1 per credit, one credit per started 20 minutes, no seats, and a session with no answers is never charged. Ten shoppers at 20 minutes each is $10 on the [pricing page](/pricing) terms.

## Which path fits you?

- Setting up one study: [Tracy in ChatGPT, step by step](/blog/tracy-chatgpt-integration).
- Drafting a discovery guide first: [ten discovery questions with follow-ups](/blog/customer-discovery-interview-questions).
- Researching a market, not a product: [ChatGPT for market research](/blog/chatgpt-for-market-research).
- Prefer Claude: [Claude user interviews](/blog/claude-user-interviews).
- The whole flow: the [customer research use case](/use-cases/customer-research).

## Common questions

### Can ChatGPT do UX research?

It can plan, draft guides and screeners, and summarise notes you give it. It cannot interview a user on its own. With Tracy added as a plugin, it sets up the interview and Tracy runs it.

### What are good ChatGPT prompts for UX research?

Prompts that name the decision and ask for evidence. Ask for a plan tied to one decision, a guide with leading questions flagged, a screener, and a theme summary that gives a quote for each theme. Check the quotes against the recording.

### Claude vs ChatGPT for UX research: which is better?

Neither changes the research. Both can hold the Tracy connector and both ask before using it. Choose the app your team already works in; the setup differs and your plan or workspace decides what you can add.

### Does ChatGPT send the invitations without asking?

No. The interview is previewed first, and so is the invitation. Nothing is sent until you approve it.

### Can ChatGPT analyse interviews I already have?

Uploads depend on your plan: OpenAI's pricing page lists file uploads as limited on Free and included in Go, Plus and Pro. Check the original transcripts before you rely on any theme.

### Can ChatGPT find participants?

No. Tracy invites the email addresses you supply and does not look up or collect addresses. You bring the participants.

---

**Run your first UX interview from ChatGPT** Install the plugin, describe the study and approve the preview. You read what each person said. The first 3 results are free. [Try it free](https://tracyinterview.ai/connect?utm_source=blog&utm_medium=seo&utm_campaign=chatgpt-for-ux-research&utm_content=try-free) [Book a demo](https://cal.com/tracy-ai/tracy-demo-web?utm_source=blog&utm_medium=seo&utm_campaign=chatgpt-for-ux-research&utm_content=demo)

## Sources

- [Add Tracy AI Interviewer to ChatGPT (public guide)](https://tracyinterview.ai/help/chatgpt), retrieved 14 Nov 2026.
- [Tracy AI Interviewer Connect page](https://tracyinterview.ai/connect), retrieved 10 Nov 2026.
- [Tracy AI Interviewer pricing](https://tracyinterview.ai/pricing), retrieved 11 Nov 2026.
- [OpenAI: Developer mode](https://developers.openai.com/api/docs/guides/developer-mode), retrieved 14 Nov 2026.
- [ChatGPT pricing](https://chatgpt.com/pricing), retrieved 15 Nov 2026.
- [OpenAI: Connecting custom MCP servers in ChatGPT](https://help.openai.com/en/articles/12584461-connecting-custom-mcp-servers-in-chatgpt), retrieved 14 Nov 2026.
