A respondent gives you three words on an open end. An AI reads those three words and asks one follow-up, right there in the survey you already field — same questionnaire, same panel, same response record.
Paste a snippet, or connect your survey by webhook. You choose which question the follow-up hangs off — nothing else in the instrument changes.
It reads what the respondent actually said or typed, together with the study context you set: industry, language, tone, and what you are trying to learn.
The respondent sees the follow-up in the flow they were already in and answers it. Both answers land on the same response record, joined by session ID.
It is not an AI-moderated interview. The AI does not run your study, does not decide what the study asks first, and does not hold a multi-turn conversation. It reads one answer and asks one follow-up on it; your questionnaire still decides everything else. If you need adaptive probing across a whole session, an AI-moderated platform is the right tool and we would rather say so.
Three ways to get depth out of an open end, and what each one costs you operationally.
| A fixed open end | Voice Capture + AI follow-ups | AI-moderated platform | |
|---|---|---|---|
| Who asks the next question | Nobody — the questionnaire you wrote | Your questionnaire, plus one AI follow-up on the answer | The AI moderator, throughout |
| Where the study runs | Your survey platform | Your survey platform | The vendor's platform |
| Your questionnaire | Unchanged | Unchanged — one snippet on the question you choose | Re-authored as a conversation guide |
| How adaptive it is | Not at all | One adaptive probe per targeted answer | Adaptive for the whole session |
| Fits a running tracker | Yes | Yes, introduced at a wave boundary | Not without breaking the time series |
| Cost shape | Included in your platform | Credits from the same balance as voice | Ask each vendor |
It embeds the way the microphone does: a script tag in Alchemer or QuestionPro, or a custom webhook from another platform — talk to us about that one. Follow-ups work on typed answers as well as spoken ones — the AI reads the text either way.
Open the add-onA follow-up draws a credit from the same pack that pays for voice transcriptions. The free trial's 250 credits cover both, and there is no card to enter.
See pricing250 credits included · no credit card · works with the survey you already run
No. It runs inside it. Your routing, quotas, screener and panel wiring are untouched — the follow-up is asked in the flow the respondent is already in, and lands on the response record you already have.
No, and the difference matters when you are choosing. An AI-moderated platform conducts the whole session and decides what to ask next, on its own platform. This asks one follow-up on one answer to a question you wrote, on yours.
Alchemer and QuestionPro today, self-serve. Other platforms can be wired up with a custom webhook — talk to us. Qualtrics is on the way.
Both. The AI reads the transcription of a spoken answer or the text of a typed one — the follow-up is generated from what the respondent said either way.
One credit, from the same balance as your voice transcriptions. Credits are bought once and never expire, and the 250 free credits on signup cover follow-ups too.
We use cookies to improve your experience and analyze site traffic. Learn more