Q5. How do you decide when NOT to use AI, even if it's technically possible?
Type: Case | Difficulty: Medium
What they're testing: Judgment about AI's limits. Great PMs know when the answer is don't use AI here.
Framework hint:
Deterministic requirement — use rules
High-stakes reversibility
Small segment doesn't justify cost
Simple UX solves 80%
Explainability legally required
Q11. A PM proposes adding AI summarization to our email product. What questions do you ask before saying yes?
Type: Case | Difficulty: Easy
What they're testing: Whether you challenge AI feature requests through user-value first, or approve because AI = good.
Framework hint:
User problem validated?
Frequency of long emails?
Failure cost of missed key point?
AI better than existing solutions?
Cost/latency vs value?
Q14. You have 3 use cases where AI could help — a chatbot, a recommender, and a summarizer. Which do you build first?
Type: Case | Difficulty: Medium
What they're testing: Applied prioritization with concrete options. Watch for candidates who pick based on cool not on impact/feasibility.
Framework hint:
User pain × frequency × feasibility × data
Recommender wins on data flywheel
Summarizer lowest risk
Chatbot highest visibility but hardest
Justify with framework, ask for more data
Q17. Your CEO wants to add AI to the product. How do you push back or scope it?
Type: Case | Difficulty: Medium
What they're testing: Stakeholder management + product judgment under exec pressure to ship AI theatre.
Framework hint:
Never say no flat
Understand underlying driver
Propose 2-3 ranked opportunities
One pilot with clear success + kill criteria
Set realistic timeline expectations
🔓 16 more questions in this segment alone — including agentic workflow design, sunsetting AI features, and 10x vs 10% opportunity framing. Get the full 100 on Gumroad →