Skip to content

Capabilities · Interviews

Depth interviews, on the moderator’s terms.

The moderator is the method. This platform extends what one moderator can deliver — live support during the interview, transcripts and coding done once, and AI-moderated capacity under the moderator’s own guide when they choose it. A dial, from assisted to fully automated. The moderator sets it.

Intelligence from people, in their own words — with the moderator in charge.

One guide, and a dial you set

Three positions, chosen per study and changeable per interview:

  • You moderate; the platform supports. Live guide tracking, what’s still open, a suggested probe you are free to ignore. Nothing speaks to your respondent but you.
  • You moderate; remaining sessions run under your guide and your review. Overflow, screener-depth, or a second wave — run asynchronously, in parallel, by voice or text. You review every transcript, code under your frame, and own the findings.
  • Run under your direction, with your review of every session. For the studies where the volume is the point. Same guide, same codebook, same standards, and the comparison to your own interviews reported blind.
Interview session — live, assisted
MYou mentioned cost came into it. What actually happens in the room when it does?
RHonestly it’s the pharmacist who raises it, not me. By then I’ve already written it.
MAnd when you swap — who decides what you swap to?

Session view with the assisted-moderation drawer (guide progress, open objectives, suggested probe). Representative

Support that stays out of your way

  • Guide tracking. Where the session stands, which objectives are still open, what the last answer left unresolved.
  • Suggested probes, never spoken. A suggestion appears to you; only you decide whether to use it.
  • Recording and transcription with word-level timing; recordings kept only when the client has asked for them, discarded by default.
  • Live view for a colleague or client observer, read-only.

Transcripts and coding, done once

  • Transcripts with synchronized audio; select a run of words and capture exactly that clip.
  • Theme coding under your codebook; a respondents-by-themes grid where every cell cites its passages.
  • Quotes verified against the transcript or withheld. Absence stated: “not asked,” “asked, no response,” “session ended” are distinguished.
  • Findings that carry their evidence into the deck.

Capacity you direct

  • The AI arm runs your guide under your direction: opens, probes each theme to the depth you set, follows unresolved answers, respects refusals, closes.
  • Respondents answer by voice or text; a session survives connection and service failures — the answer is re-sent, not lost.
  • What it could not get is reported against your guide, not smoothed over.
  • Validated on your guide. Run both arms on one guide, coded under one frame, blinded; read five computed figures and expert ratings. You decide where the AI arm earns its place — and the platform reports honestly when it doesn’t.

For moderators who bring us their studies

  • Your guide is used as provided. We program it; we do not rewrite it.
  • You keep the client. White-label deliverables; we appear only if you want us in the room.
  • You choose the dial. Assisted only, if that is your practice. Nothing automated is switched on without your approval.
  • Fixed fee, firm quote, so the number in your proposal stays the number.

Common questions

Who moderates?

You do. The AI arm runs your guide at your direction, for the interviews you assign it. The assisted position involves no AI moderation at all.

Who owns the findings?

You do. Transcripts, codes, and findings are scoped to your study and your client, never pooled.

Can I see how the AI arm compares to my own interviews?

Yes — blind, on the same guide, with the comparison reported either way.

What happens to recordings?

Discarded by default; kept only where your client has asked, under their retention setting.

Can respondents tell the difference?

The disclosure says who is moderating. Consent is explicit and stored verbatim.

Bring us the guide. We program it as written, you set the dial, and the findings are yours.