Summarize Usability Test Sessions Across Many Participants

1. Go to maze.co and sign up, or sign in with your existing account.
2. Connect your Figma file, or upload the prototype you want tested.

What you should see: A project dashboard with an option to create a new usability test.

1. Start a new usability test and add your prototype.
2. Define the tasks you want participants to complete and the success criteria for each (reaching a specific screen, completing a checkout flow, finding a setting).
3. Add any follow-up questions you want participants to answer after each task.

What you should see: A test builder showing your task list, with a preview of what participants will experience. Troubleshooting: Keep the task list short. more than five or six tasks per session tends to produce fatigued, less useful responses.

1. If you're bringing your own participants, generate the shareable test link and send it directly.
2. If you're using Maze's participant panel option, set your targeting criteria and launch recruitment through the platform.

What you should see: A live test link and a dashboard tracking how many participants have started and completed the study.

1. Leave the test open for participants to complete on their own schedule, or run it live if you're moderating.
2. Check the dashboard periodically as responses come in.

What you should see: Task completion rates, time-on-task, and individual session recordings populating as participants finish.

1. Once you have enough completed sessions, open the results dashboard.
2. Look for the AI-assisted summary feature that surfaces patterns across open-text responses and flags friction points in the task data.
3. Review the generated findings alongside the raw task metrics (completion rate, time on task, drop-off points).

What you should see: A report highlighting where participants struggled most, grouped by task or theme, rather than a separate written note for every individual session. Troubleshooting: If the AI summary feels thin or generic, check that you have enough completed sessions. summarization features work better with a meaningful sample rather than two or three responses.

1. Rank the flagged issues by how many participants hit them and how severe the impact was (a full task failure versus mild confusion that self-corrected).
2. Cross-check the top two or three issues against the actual recordings before writing them into a design backlog, since a summary can occasionally overstate or miss context a video makes obvious.
3. Share the prioritized list with your team instead of raw session recordings, saving everyone the multi-hour watch-through.