Adaptive Microsurveys uses AI to generate personalized follow-up questions and automatically route respondents to the right end action based on their initial response in the Microsurvey.
Instead of manually building branching logic for every possible response, you describe what you want to learn. Prism writes the initial question, generates real-time contextual follow-ups, and decides what happens next for each respondent.
Availability & Usage
🧪 Early access coming soon
📌 Powered by Prism
⚙️ Create from the Microsurveys page
How do Adaptive Microsurveys work?
Every Adaptive Microsurvey follows a fixed 3-step structure:
Question: The initial survey question. You choose the format: Rating, NPS, Multi-button, Dropdown, or Input. Add the main Question and tell Prism what you want to achieve.
Adaptive Follow-Up: Prism reads each respondent's answer and generates up to 3 contextual follow-up questions in real time. If a respondent gives short or unhelpful replies, Prism moves them to the End Action step early instead of pushing for more.
Adaptive End Actions: Prism routes each respondent to the best-matching end action you've configured (e.g., ask for a review, offer support, trigger an Experience). You can also have a single End Action for all respondents.
👉 Adaptive Follow-Ups are conversational. Each follow-up question is based on what the respondent just said, not on a pre-configured decision tree. Two respondents who give the same initial rating but different explanations will get different follow-up questions.
How do I create an Adaptive Microsurvey?
Go to Microsurveys in your Dashboard and click "Create Microsurvey"
Choose Adaptive Survey (powered by Prism) as the type
Once you select Adaptive Survey, the creation flow differs from that of a Manual one. Instead of picking a question type and writing copy yourself, you'll:
Describe what you want to learn — Tell Prism the topic or goal (e.g., "feature satisfaction with the new sidebar," "why users aren't completing onboarding")
Choose your question format — Select from Rating, NPS, Multi-button, Dropdown, or Input
Update your main Question
Click Build — Prism generates your initial question and adds an Adaptive Follow Up Step and an End Action Step.
After building, your Microsurvey has the 3-step Adaptive structure ready to configure.
What's the difference between Adaptive Survey and Manual?
| Adaptive | Manual |
Question text | AI-generated based on your goal | You write it yourself |
Follow-up logic | AI generates personalized follow-ups per respondent, in real time | You configure Steps manually for each answer |
End Actions | AI analyzes the conversation and routes to the best matching action | You assign fixed end actions per answer path |
Number of Follow Up Questions | Up to 3, determined dynamically by Prism | Up to 2 |
Best for | Exploratory feedback, understanding "why" behind ratings, conversational surveys | Precise control over every question and routing path |
💡 Use Adaptive Surveys when you want richer qualitative feedback without the overhead of building complex branching logic. Use Manual when you need exact control over every Step (specific wording, fixed answer options, deterministic routing).
How do I configure the Adaptive Follow Up?
The Adaptive Follow-Up step lets you guide Prism's behavior when generating follow-up questions. You can add instructions that shape the tone and direction of the conversation, for example, "keep it conversational," "ask about pricing directly," or "focus on usability issues."
These instructions influence how Prism formulates follow-up questions for every respondent. You can edit them at any time from the Adaptive Follow-Up step in the Dashboard.
ℹ Prism will generate up to 3 Follow Up questions per respondent. If a respondent gives minimal or unhelpful replies, Prism will skip the remaining Follow Ups and move them to the End Action step. This prevents respondent fatigue and keeps the survey feeling natural.
How do I set up Adaptive End Actions?
Adaptive End Actions determine what happens after the conversation is complete. Prism analyzes the full exchange — the initial answer plus all follow-ups — and routes each respondent to the end action that best matches their responses.
Default end action
Every Adaptive Microsurvey has a default end action. This is the fallback; respondents are routed here when their responses don't strongly match any other configured action or as the final Step of the survey. The default includes:
A body message (e.g., "Thank you for your feedback!")
A dismiss button
Optional additional actions (open URL, trigger Experience, etc.)
Adding custom end actions
You can add custom end actions for specific routing scenarios. Each custom end action has:
Field | Purpose |
Routing description | Tells Prism when to route here. Describe the type of respondent this action is for (e.g., "users who are unhappy with the sidebar," "users who mention pricing concerns"). |
Body copy | The message shown to the respondent when routed to this action. |
Button copy | The button text to close the Microsurvey or open your End Action. |
Additional Actions | What happens after: the same Additional Actions available in any Microsurvey. |
Chameleon also provides suggested end actions to get you started quickly. These are common routing scenarios (e.g., "Speak with support," "Ask for a review") that you can add with one click and customize.
💡 Want to reuse a routed group later as an audience in Chameleon? Add a Tag as an End Action. For example, tag respondents routed to your "happy users" action so you can target them with other Experiences.
How does Prism decide where to route respondents?
Prism reads the entire conversation to decide which end action best fits each respondent. The routing description you write for each end action is what Prism uses to make the match. For example, if you have three end actions configured:
Default: "Thank you for your feedback!"
Unhappy users: Routing description says "users who express frustration or dissatisfaction" → shows a "Speak with us" CTA
Happy users: Routing description says "users who are satisfied or enthusiastic" → asks for a review
A respondent who gives a low rating and explains they're frustrated with load times would be routed to the "Unhappy users" action. A respondent who gives a high rating and says they love the new feature would be routed to "Happy users." Someone whose responses are neutral or don't clearly match either would land on the Default.
💡 Write routing descriptions as clearly as possible. The more specific you are about what type of respondent should be routed to an action, the more accurate Prism's routing will be.
Can I edit an Adaptive Microsurvey after building it?
Yes, with some constraints:
What you want to learn — Editable. You can update your goal, and Prism will adjust.
Question text — Editable after building
Adaptive Follow-Up instructions — Editable. You can change the prompts that guide Prism's follow-up behavior at any time.
End actions — Fully editable. Add, remove, or modify end actions and their routing descriptions. Note that respondents matching an Action you remove will fall back to the default action.
How do I preview an Adaptive Microsurvey?
You can preview your Adaptive Microsurvey from the Builder. The preview is fully interactive: you can answer the initial question, experience Prism's follow-up questions in real time, and see how end action routing works, all without publishing.
💡 Use preview to test different types of responses and see how Prism adapts its follow-up questions. This is a good way to validate your follow-up instructions and routing descriptions before going live.
Tips for getting the most out of Adaptive Microsurveys
💡 Be specific about what you want to learn. "Feature satisfaction" is fine, but "satisfaction with the new reporting dashboard, specifically around load times and data accuracy," gives Prism much better context to generate relevant questions.
💡 Start with 2–3 End Actions. A default plus one or two custom actions is enough for most use cases. You can always add more as you see patterns in responses.
💡 Write routing descriptions like you'd brief a colleague. Instead of "unhappy users," try "users who mention bugs, slow performance, or missing features they expected." Specificity improves routing accuracy.
💡 Use Adaptive Follow-Up instructions to set tone. If your product voice is casual, tell Prism. If you want it to dig into pricing feedback specifically, say so. These instructions shape every follow-up conversation.