AI Follow-up Questions From the Respondent's Side¶

Everything about configuring AI Follow-up Questions - trigger conditions, custom instructions, the follow-up limit - happens on the builder's side (see AI Follow-up Questions: What Triggers One and What Doesn't). A respondent never sees any of that. What they experience is much simpler, and understanding it helps you write better trigger conditions and instructions.

It Looks Like the Next Message, Not a New Feature¶

When a follow-up triggers, it appears as the next message in the chat - visually and functionally identical to any other question in the survey. There's no label, badge, or indication that this particular question was generated based on their specific answer rather than written in advance. From the respondent's perspective, it simply feels like the conversation is following up on what they just said.

Why This Matters for How You Write Instructions¶

Since a follow-up should feel like a natural continuation, your custom AI instructions (see AI Follow-up Questions: What Triggers One and What Doesn't) are worth writing with tone in mind, not just content - a follow-up that reads as an interrogation rather than genuine curiosity can feel jarring even if it's asking about something relevant.

Only Available in Chat Interface¶

A respondent taking a survey in Traditional Form never encounters an AI Follow-up Question, regardless of how the question is configured - the feature is specific to the conversational, one-question-at-a-time nature of Chat Interface. See Chat vs. Traditional: Choosing How Respondents Take Your Survey for why the format matters here.

Answering a Follow-up¶

A respondent answers a follow-up the same way they'd answer any open-ended question - it's captured as its own response and can be run through Text Analytics like any other open-ended answer. There's no special handling a respondent needs to be aware of.

Not Every Answer Triggers One¶

Since follow-ups only fire when a configured condition is met, most respondents in a well-tuned survey experience the standard flow without ever seeing a follow-up - only answers that meet the trigger actually get probed further. A respondent who never sees one hasn't missed anything; the survey simply didn't have a condition to react to in their case.

FAQ¶

Does a respondent know in advance that follow-up questions might happen?
There's no built-in indicator warning respondents a follow-up might occur - it's designed to feel like a natural part of the conversation rather than a disclosed mechanic.

Can a respondent tell how many follow-ups they've received relative to the Max Follow-up limit?
No - the limit is a builder-side setting with no respondent-facing counter.

Is there a delay while a follow-up question is generated?
Generation happens live during the respondent's session - check your survey's actual behavior in preview to see the timing, since this can vary.

Does a respondent's answer to a follow-up affect what later questions they see?
Only if Display Logic is specifically set up to reference that follow-up's answer - by default, a follow-up doesn't change the rest of the survey path on its own.

For the builder-side setup, see AI Follow-up Questions: What They Are and Where They Work.