Write a Customer Interview Script That Validates a Problem with Past-Behavior Questions
Get a 30-minute problem-discovery interview script with past-behavior questions, two follow-up probes each, banned-question rewrites and evidence signals.
You getOpening script, table and closing script with a referral ask
Prompt
Act as a product researcher who runs problem-discovery interviews for early-stage products.
Write a 30-minute customer interview script that checks whether a problem is real, frequent and painful enough to solve. Before writing, ask me these in one message and wait for my answers:
1. What problem do you think exists, and for whom?
2. What do you believe people do about it today?
3. Who will you interview (role, company size or life situation)?
4. What decision will this round of interviews inform?
Rules:
- Every main question asks about past behavior ("Tell me about the last time..."). No hypotheticals ("Would you...", "If we built...") and no leading questions that name the problem or its pain before the interviewee does.
- Give 2 follow-up probes per question that dig for specifics: dates, time spent, money spent, tools used, who else was involved.
- Never mention a solution or pitch an idea.
- Mark which answers count as strong evidence and which are just polite interest.
- List 5 tempting questions to avoid, each with a past-behavior rewrite.
Output format:
1. Opening script (under 80 words)
2. Table: # | Main question | Probe 1 | Probe 2 | What we learn
3. Questions to avoid → rewrites
4. Closing script with a referral ask
5. Signals: strong evidence vs. false positives
End with one line naming the riskiest assumption these interviews cannot test.Example output
Tips and common edits
Why this works
- Banning hypotheticals and leading questions keeps answers tied to what people actually did, not polite guesses.
- Two probes per question push for dates, money and tools, the details that separate real pain from mild annoyance.
- The strong-vs-false-positive list helps you score interviews honestly instead of hearing what you hoped.
Use before building or scoping a feature, when you have a problem hypothesis and a list of people to talk to. Paste into ChatGPT, Claude or Gemini and answer its four questions about the problem and audience.
Common edits
- Ask for a 15-minute version with only the four most important questions.
- Add a column for note-taking cues or a tag scheme to synthesize across interviews.
Prompt details
- By
- Prompts-GPT editors
- Published
Did this prompt work?
Keep exploring
More like this
Synthesize User Interview Notes into Counted Themes and Insights
Turns raw interview notes into a ranked theme table showing how many participants raised each theme, with observations kept separate from interpretation.
Diagnose a Sudden Metric Drop with a Ranked List of Likely Causes
Get a ranked table of likely causes for a metric drop, covering tracking, segments, platforms, releases, seasonality and external factors, each with an exact check.
Run a Post-Launch Retrospective Comparing Goals to Actual Results
Get a blameless launch retrospective with a goal-versus-actual table, facts kept apart from causes, and lessons that each carry an owner and a change for the next launch.
Write an Event Tracking Plan Where Every Event Maps to a Success Metric
Get a tracking plan table with object_action event names, exact triggers, typed properties and the success metric each event feeds, plus a list of events to cut.
Define a North Star and Success Metrics for a New Feature with Guardrails
Get a North Star metric, up to three primary metrics each paired with a guardrail, exact formulas from your existing data, targets, and ship or roll-back rules.
Draft a Now/Next/Later Quarterly Roadmap Tied to One Outcome Metric
You get a one-quarter roadmap grouped into Now, Next and Later, with every item linked to your outcome metric, a confidence level, a leading signal and parked items.