Back to Skills Library

You've got 30 minutes with one customer and a list of assumptions to test - what do you actually ask them?

Discovery interview coming up - what questions actually surface the real problem?

Recommended AI Agent Skills

Guides a PM through preparing discovery interview questions tied to a stated research goal, avoiding confirmation-biased questions.

User ResearchInterview PrepProduct Management
2026-02-126.9k

Maintains a weekly customer-touchpoint cadence and builds an Opportunity Solution Tree connecting business outcomes to real customer problems.

Product DiscoveryOpportunity Solution TreeCadence
2026-02-232.1k

Applies Steve Blank's Customer Development methodology to validate hypotheses before building anything.

Customer DevelopmentValidationLean Startup
2026-02-13149

Action framework

  1. Write your research goal as one sentence before drafting any question - anything that doesn't trace back to it gets cut.
  2. Ask about past behavior, not future intentions - "Tell me about the last time you..." beats "Would you use a feature that...".
  3. Avoid questions that can be answered with a nod - "Is pricing a problem?" gets a polite yes; "Walk me through your last renewal decision" gets a real answer.
  4. Never pitch the solution mid-interview - if the customer asks what you're building, deflect back to their problem until the interview ends.
  5. After each answer, ask one follow-up "why" or "what happened next" before moving on - the first answer is rarely the real one.

Before

"Would you use a feature that let you export this to Excel?"

After

"Walk me through the last time you needed this data outside the app - what did you do, and what took the longest?"

Explore course: AI PM