No quote leaves your synthesis until you have found it in the transcr…
AI checks the transcripts, does the first coding pass and pushes back on your themes. You read, group and write the insights. In between, every quote gets found again word for word.
Why
A made-up quote looks exactly like a real one. In a team I was coaching, a user's sentence distorted by AI made it into a research readout. It was caught and corrected, but only afterwards. Nothing exotic about it: across five thematic analysis datasets, a general-purpose assistant's outputs often contained fabricated quotes (58% on average), while 79% of human analysts' quotes were accurate. AI still pulls its weight here: checking, a first coding pass, alternative groupings, jargon hunting. Deciding what the data means stays with you. This kit is about product synthesis, the kind that…
When
Set up this kit once the interviews are done and a readout is coming. Audience: product designers, PMs and POs who synthesise with or without a dedicated researcher. ✅ You know it's your kit when You paste transcripts into an assistant and copy its quotes as they come Your themes look like the keywords of the questions you asked Nobody remembers anymore which participant said the line the…
Tools included
- Claude
- Notebooklm
- Miro
- Dovetail