Q61. What adoption problems typically occur in AI products and how do you navigate them?
Type: Case | Difficulty: Medium
What they're testing: Whether you know AI-specific adoption failure modes vs generic engagement problems.
Framework hint:
First-use failure — one bad output kills the account
Trust decay as edge cases emerge
Prompt anxiety
Novelty vs habit — 20% use once, 2% integrate
Better onboarding, scaffolding, quality gates
Q63. Users are afraid the AI will make mistakes. How do you design for trust?
Type: Case | Difficulty: Medium
What they're testing: Trust design as a first-class product problem, not a marketing message.
Framework hint:
Show confidence differently by level
Reversibility of every AI action
Explainability — "AI suggested this because..."
User control: accept, edit, reject
Never claim 100% accuracy
Q64. Your AI feature launched with 30% adoption week 1, dropped to 8% by week 6. What's happening and what do you do?
Type: Case | Difficulty: Hard
What they're testing: Diagnosing AI-specific drop-off patterns. Watch for candidates who default to generic engagement analysis.
Framework hint:
Novelty effect — 20-30% try once
Cohort analysis for power users
First-use quality correlation with churn
Trust erosion over time
Fix onboarding, first output, failure salience
Q74. Your AI feature caused a public incident (wrong output at scale). Walk me through the first 48 hours.
Type: Case | Difficulty: Hard
What they're testing: Incident response for AI at scale. Reveals how many times they've done it.
Framework hint:
Hour 0-2: kill switch, blast radius, user comm
Hour 2-12: root cause, internal + exec brief
Hour 12-24: public statement, outreach, remediation
Hour 24-48: fix rollout, post-mortem, monitoring
Long-term: evals, guardrails, review changes
🔓 16 more questions in this segment alone — including enterprise trust rollout, regulatory patchwork, and hostile-user adoption stories. Get the full 100 on Gumroad →