Presenting Hands-On AI Product and Workflow Automation Experience as a PM

You are interviewing for a product manager role at Perplexity. Walk the interviewer through your experience with AI: projects you have worked on and anything you have automated with AI tools. ```hint Show the decisions, not the buzzwords Choose one project and be ready to explain why AI was the right tool, how you judged whether its output was good enough, and what you changed when it was not. ``` ### Clarifying Questions - Do you want products I shipped to users as a PM, or also automations I built for my own or my team's work? - Should I go deep on one project or give a quick tour of several? - How technical should I get: product decisions only, or also how the system worked, such as the model, retrieval and evaluation? ### What a Strong Answer Covers - A concrete project with a clear problem, its users, and the reason AI fit better than rules or manual work. - The candidate's own role and decisions, kept distinct from the team's. - How output quality was defined and measured, and how wrong or made-up answers were caught and handled. - Practical constraints such as latency, cost per request, privacy and fallbacks, and the trade-offs made among them. - Measured results, the main lesson, and a link to the quality problems an answer engine like Perplexity faces. ### Follow-up Questions - What was the most common way your system failed, and how did you find out? - If you rebuilt that project today with current models, what would you change? - How would you evaluate the quality of a Perplexity answer that comes with cited sources?