Launched this week
Inqueria: Interviews that Hold Up
AI that interviews your users like a trained researcher
34 followers
AI that interviews your users like a trained researcher
34 followers
Inqueria runs AI-moderated interviews that adapt to every answer, the way a trained researcher would. It redacts PII in-house before any model sees a transcript, then turns every conversation into themes with the exact quotes behind them. It checks each theme against every transcript and shows you the participants who disagree. Run 50 interviews overnight. Skip the month it takes to schedule five.
Dial
you asked for where this breaks, so here's my honest skepticism: the interviewer itself is a variable. two runs of the same study could phrase follow-ups slightly differently and nudge people toward different answers, and that's invisible in the output since you only see themes and quotes, not the exact wording of what was asked. do you log the actual follow-up questions the AI generated per session, so a skeptical reader could check whether the moderator was leading the witness?
@galdayan Fair challenge, and yes. Every question the AI asks, follow-ups included, is saved word for word in the session transcript. Each quote in a theme links back to its session, so you can read the exact question asked right before it. Each session also records which version of the question plan (the plan AI generated based on researcher's objective) it ran against. These transcripts are available to the researcher (the study owner/creator).
The interviewer runs under a hard rule not to lead or suggest answers, but a rule isn't proof, which is why the wording is all kept.
Where you're right: The app provides a client viewing read-only report that the study owner/creator share and they see themes and quotes, not the transcripts, so they'd need the researcher's export to check the moderator. And nothing automatically flags a leading follow-up yet. Showing the question next to every quote is a good idea, and I'm adding it to the list.