As of 13 August 2026, AI can write user interview questions for your product research.

What to actually do

1. Hand it to a person

   The route this page recommends

   A person who owns the outcome does this end to end, worth it when the failure is dear.

2. Use a tool built for this

   Second choice

3. Do it yourself

   The distant third

   A chat interface, no skill needed, and roughly 30 minutes until you can act on the result.

   How to actually do it

   1. Open a blank document and write the product, research objective, decision to inform, target participants, interview format, available time and known constraints.
   2. Paste those details into the prompt's bracketed slots, then run the prompt in ChatGPT, Claude or Gemini.
   3. Read the generated research aim and compare it with the decision you actually need to make, correcting any invented assumptions in the brief.
   4. Check each main question against the participant's past behaviour, removing questions that lead to a preferred answer, combine two topics or ask for unsupported predictions.
   5. Paste the final guide into your interview document and mark the essential questions that must fit within the available interview time.
   6. Run one practice interview with a colleague or suitable tester, then rewrite any question they misunderstand before inviting research participants.

   Prompt

   Act as a careful user-researcher helping me prepare a semi-structured interview guide.
   
   Product or service: [describe it]
   Research objective: [what I need to learn]
   Decision this research will inform: [what may change based on the findings]
   Target participants: [who they are, including relevant experience]
   Interview format and length: [for example, remote or in person, and available time]
   What participants can see or use: [concept, prototype, live product or none]
   Known constraints: [topics to avoid, recruitment limits, commercial or ethical constraints]
   
   Create a practical interview guide that:
   1. Starts with a short welcome, consent check and neutral warm-up.
   2. Uses open questions about participants' real past behaviour before asking about opinions or reactions.
   3. Avoids leading, double-barrelled, hypothetical and jargon-heavy questions.
   4. Separates the main questions from optional follow-up probes.
   5. Orders the questions so the most important learning is gathered early.
   6. Includes a short note beside each question explaining which research objective it serves.
   7. Flags any assumption or important gap in my brief instead of inventing details.
   8. Ends with a neutral closing question and a request for permission to follow up.
   
   Return the result as:
   - a one-paragraph research aim;
   - a numbered interview guide with estimated time beside each section;
   - optional probes under the relevant main question;
   - a list of questions you rejected or rewrote because they were leading or otherwise weak;
   - five checks I should make before using the guide.
   Do not claim that the questions will produce representative findings, and do not collect unnecessary personal or sensitive information.

   Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.