AI surveys

AI surveys

Writing good survey questions is hard — the wording biases the answers, and you’re too close to your own idea to be neutral. The survey tool generates questions for you, delivers them to respondents, analyzes the answers, and feeds the results back into your PMF score.

1. Generate

Tell the tool about your idea, and it drafts a complete survey:

• Question generation — questions are written from your idea’s context, so they’re specific to your product rather than generic.
• Built-in templates — choose a starting structure: problem validation, willingness to pay, feature prioritization, NPS/satisfaction, competitive positioning, and others.
• Bias protection — the generator applies rules designed to avoid leading questions, and it can flag bias warnings so you can see where wording might steer respondents.

Review the generated questions before saving — you can edit, remove, or add questions freely. The survey is yours; the AI just writes the first draft.

2. Deliver

Each survey needs respondents. Choose a delivery method:

• Built-in delivery — publish the survey inside the platform and share a link with your audience panel or anyone else. Responses are collected directly in the app. This is the zero-setup option.
• External survey platform — publish the survey to a professional survey platform for a wider audience or more advanced response collection. This requires a connection to the platform; see the Integrations overview for setup.

Whichever route you choose, responses come back into the same place, so you analyze everything in one view. Duplicate respondents are filtered out automatically.

3. Analyze

As responses arrive, the analytics view turns them into answers:

• NPS — the Net Promoter Score from your NPS-style questions.
• Sentiment — overall tone of open-ended answers.
• Word clouds — the words respondents actually use, so you can see their language rather than yours.
• Cross-tabulation — break responses down by segment or question.

You can also export the raw responses as CSV whenever you need to work with the data elsewhere — see Exporting your work.

4. The PMF auto-bridge

The most powerful part is what happens next. The PMF bridge maps question answers to the dimensions of your PMF score — must-have, disappointment, willingness to pay, referral likelihood, market size, and competitive moat.

Once a survey collects enough responses across enough mapped dimensions, the bridge updates your PMF score automatically. Your score stops being your opinion and becomes a summary of what respondents told you.

This matters more than it sounds:

• Before the bridge — “I think people would pay $29/month.”
• After the bridge — “82% of 40 respondents said it’s a must-have, and 61% would pay $29/month.”

Tip

Design surveys with the bridge in mind. If you want your PMF score to reflect real data, include questions that map to the dimensions you care about — the generator suggests these when you choose a PMF-oriented template.

A practical flow

1. Generate a survey around your riskiest assumption (pricing, must-have, referral).
2. Publish it to your audience panel or an external platform.
3. Watch responses arrive and read the analytics.
4. Let the bridge update your PMF score — then check whether the evidence matches your assumptions.

If responses contradict your assumptions, that’s not failure — that’s the survey working. Revisit your positioning or your go/no-go decision with the new evidence.