Perplexity AI·Product Manager·Onsite - Product Sense / Strategy·Senior

Senior

Apr 2026

Summary

Interviewed for a PM role at Perplexity AI. One product strategy question about rolling out a new feature. Short post, not much context to go on.

Questions Asked (1)

How would you approach rolling out a new feature for Perplexity?

Product StrategyGo-to-Market (GTM)A/B Testing & Experimentation

Author's notes

Classic GTM meets product sense mashup.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by clarifying the feature's goal and target user segment, then outline a phased rollout that balances speed with learning. Emphasize a hypothesis-driven approach with clear success metrics, and describe how you'd iterate based on data and user feedback.

Pro tip: Tie your rollout plan to Perplexity's core value of accurate, cited answers—show how you'd measure trust and answer quality, not just engagement. Also, mention a rollback plan to demonstrate risk management.

1. Define Success and Hypotheses

Clarify the feature's objective, target user segment, and key hypotheses. Define success metrics (e.g., answer quality, user retention, query volume) and guardrail metrics (e.g., latency, citation accuracy).

2. Design the Experiment

Choose an A/B testing or phased rollout design. Determine sample size, duration, and randomization unit. Ensure the experiment isolates the feature's impact and includes a control group.

3. Execute a Phased Rollout

Start with a small internal or beta group, then expand to a percentage of users. Monitor metrics in real-time and be ready to pause or roll back if guardrails are breached.

4. Analyze and Iterate

After sufficient data, analyze results against hypotheses. If successful, plan a broader launch; if not, iterate on the feature or pivot. Document learnings for future experiments.

5. Scale and Communicate

Once validated, roll out to all users with a clear communication plan. Highlight the feature's value to users and internal stakeholders, and set up ongoing monitoring for long-term impact.

Key Points to Mention

Hypothesis-driven approach with clear success metrics
A/B testing methodology and statistical significance
Phased rollout to mitigate risk and gather feedback
Guardrail metrics to ensure quality and trust (e.g., citation accuracy, latency)
Iterative learning and willingness to pivot based on data
Cross-functional collaboration with engineering, design, and data teams

AI-generated suggestions, not part of the candidate's original notes. May be inaccurate — verify before relying on them.