Interview transcripts contain IA evidence that a general summary will flatten. When a participant says "I gave up and just searched for it," that is a navigation finding. When three participants call the same thing by a name that appears nowhere in your navigation, that is a labeling finding. This prompt asks for those specific reads, finding behavior, mental models, navigation friction, vocabulary, expectation gaps, rather than an overall summary, and it requires a participant count and a supporting quote beside every theme.
Paste in excerpts rather than your notes about the excerpts. The participants' own words are the point, and summarising strips out exactly the vocabulary you are trying to capture. Giving the participant count matters too, because it is what lets you catch the most common failure in AI-assisted synthesis: a theme presented as widespread that traces back to one talkative person. When you read the output, check each theme against its own quote and count. If the quote does not plainly say what the theme claims, the theme is an interpretation, and interpretations belong in a separate column from evidence.
A theme without a count and a quote is an opinion with a citation style. The count is what tells you whether to act on it.
Fill in your own details below; the prompt updates as you type. When it's ready, copy it into Claude or whatever AI tool you use.
Casts the AI as: You're a UX researcher extracting information architecture insights from interview data.
Names 5 specific outputs to produce, so the response comes back structured rather than a general summary.
Research context, Number of participants, User type and Current IA tell the AI what your specific situation is, not a generic one.
Thematic analysis with quoted evidence and IA recommendations