Your AI assistant can now work inside Userfeel
You've just wrapped up a 12-session usability test. Before, that meant hours of watching recordings, clipping highlights and writing up findings. Now you can just ask: "What were the biggest problems?"
Connect Claude, Claude Code or another AI assistant to your Userfeel account, and it can read your test results, annotate sessions and draft new tests for you.
The short version
- What: AI assistants can now work directly with your Userfeel data.
- How: through MCP (Model Context Protocol), the open standard that lets AI assistants use outside tools safely.
- Setup: about two minutes, from Dashboard → AI access.
- Control: you decide what each assistant can see and do, and nothing gets launched or spent without your click.
Ask it anything about your tests
You talk to your assistant in plain language, just like you'd ask a colleague. Here's what that looks like.
Get answers fast
- "Summarize the top five usability problems from last week's checkout test, with a tester quote for each."
- "Which testers struggled to find the pricing page? Give me the session and timestamp for each."
Skip the busywork
- "Clip every moment where a tester hesitates on the sign-up form and add a note explaining why."
- "Rate each session in the onboarding study for task success and tag the ones I should watch first."
Plan what's next
- "Draft a follow-up test for the new navigation, reusing the tasks that worked best last time."
- "The redesign test has enough sessions. Pause it, set up the comparison test and ask me to launch it."
The result: hours of watching, tagging and writing up become a quick conversation. You spend your time on decisions, not admin.
You choose how much it can do
Every assistant gets one of three access levels. Start with the lowest one that gets the job done.
Look only (recommended)
Reads your projects, sessions, transcripts and clips, but changes nothing. Perfect for summaries and quick answers.
Look and edit
Everything above, plus it can add clips, notes and ratings, and write draft tests. Great for tagging sessions and planning your next study.
Edit and launch
Everything above, plus it can pause tests and ask you to launch them. Ideal for running a study end to end, with you approving every spend.
You also pick which projects it can see (all of them, or just the ones you choose) and how long its access lasts (from 7 days to 1 year, or until you revoke it).
Set it up in 5 quick steps
Go to Dashboard → AI access and click + Connect another assistant. Then:
- Name it. Something only you see, like "Research summary bot".
- Pick its access. Look only, Look and edit, or Edit and launch.
- Choose projects. All of them, or only the ones you pick.
- Set an expiry. 30 days is a good default.
- Connect. Create the token and add it to your assistant.
That's it. Ask your first question.
Safe by design
- Your credits, your call. Launching a test or spending credits always needs your confirmation in Userfeel.
- Least access by default. Look only is recommended, and you can limit an assistant to a single study.
- Tokens expire. Pick the shortest period that fits, and revoke access any time. You can also check each assistant's activity.
- Watch for tricky text. Tester transcripts can contain text written to mislead an AI. Use read-only access for assistants that can also browse the web or run commands.
Try it today
AI access is available now on Teams and Enterprise plans, along with our other AI features. Start small: connect a Look only assistant to one project and ask it to summarize your latest study.