Strategy question
Suppose user research shows strong demand for one-click agent execution, but design wants a highly autonomous experience while engineering believes current reliability only supports a confirmation-based copilot. How would you decide what version to ship first, and what evidence would you require to expand autonomy over time?
- Perplexity
- Strategy
- Hard
Practice this question out loud. An AI interviewer asks it, follows up like a real interviewer would, and scores your answer. Type or speak.
Start a mock interview on this question · Mock interview from a job description
What this question tests
Whether you can resolve a real design-versus-engineering tension with evidence rather than picking a side by authority, and define a credible path to expand autonomy later.
How to approach it
- Restate the conflict plainly: strong demand exists for one-click execution, but current reliability may not support it safely without confirmation steps.
- Ship the confirmation-based copilot first: reliability is a hard constraint, since a wrong autonomous action erodes trust faster than a slower confirmed one builds it.
- Instrument the confirmation flow to measure how often users approve without meaningfully checking, which becomes an early proxy for how much autonomy users would actually tolerate.
- Define explicit evidence to expand autonomy: task-level accuracy above a set threshold, low correction rate over a rolling window, and user research showing comfort with reduced confirmation.
- Communicate to design and engineering that this is a staged rollout, not a final decision, so both sides see their concern addressed on a timeline.
What a strong answer includes
- Chooses the lower-risk path first with a clear reliability argument rather than trying to please both sides at once.
- Defines a specific, measurable bar for expanding autonomy instead of leaving it as a vague future goal.
- Notices that rubber-stamped confirmations are themselves useful data about user tolerance for more autonomy.
Common mistakes
- Picks a side to appease one function without giving a path forward for the other.
- No concrete threshold for when autonomy expands, leaving the roadmap open-ended.
Likely follow-up questions
- What would you do if users start rubber-stamping confirmations rather than reviewing them.
- How would you message the staged rollout to users who wanted full autonomy from day one.
More strategy questions
- Perplexity made Comet free in 2026. How would you monetize an AI browser without hurting the experience?Perplexity · Strategy · Hard
- How would you handle publisher relationships when Perplexity summarizes their content?Perplexity · Strategy · Hard
- How should Perplexity compete with ChatGPT's built-in web browsing?Perplexity · Strategy · Hard
- How would you grow Comet Enterprise adoption inside large companies?Perplexity · Strategy · Hard
- You believe Perplexity’s core agentic experience could unlock value for an enterprise knowledge-work use case, such as competitive research or internal document synthesis. How would you decide whether that opportunity deserves investment now, and what tradeoffs would you make between enterprise-specific requirements and keeping the everyday core product fast, intuitive, and cohesive?Perplexity · Strategy · Hard
- Users often start a task on mobile and expect to continue it later on web, but today the experience feels fragmented. How would you define the cross-device product direction, decide what state and context must persist, and prioritize the first release across iOS, Android, web, and backend teams?Perplexity · Strategy · Hard
More questions from Perplexity
Learn the skill behind it
Chapters of the AI PM course that teach what this question tests.
- Chapter 4: Discovery and strategy for AI products
- Chapter 9: Prove it paid off: outcomes, economics, and pricing
- Chapter 14: Get the job: the AI PM interview loop