Skip to main content
Ben Erez shows how to use AI to simulate candidates in big tech product sense interviews, helping PMs see what good performance looks like and practice their own approaches. Ben explains the context: “When you’re interviewing at series A, B, C companies, it’s about fit, behavioral questions, and chemistry. As you get to bigger companies, the people interviewing you probably won’t even work with you. I joined Facebook as the 2000th PM - what they’re looking for is standardized signals that eliminate bias and look for specific evaluation criteria.” Here’s Ben’s demonstration using Claude as a candidate answering “You are a PM at Uber. Design a product for kids”: First, Claude starts with assumptions: “I’m assuming we’re talking about children under 18, that this would be a new product line for Uber, and that safety is the primary concern.” This models what Ben coaches - always state assumptions upfront. Second, Claude outlines a structured approach: “I want to describe the product experience, break down the target audience, identify key problems, pick one, brainstorm solutions, pick one, and describe a V1.” Ben responds with terse “Yes” answers - exactly how real interviewers respond. Third, Claude works through each section methodically:
  • Sets a mission statement connecting to Uber’s broader mission
  • Identifies ecosystem players and chooses to focus on parents
  • Creates three mutually exclusive segments with scoring
  • Maps user journey to identify problems
  • Develops solutions and evaluates on impact/effort
  • Describes V1 implementation with risk mitigation
Fourth, Claude self-evaluates its performance, scoring each section 1-5 and breaking down time allocation. Ben notes this helps candidates internalize what good looks like. Ben emphasizes: “So many people who have been heads down doing really good work in their companies are really rusty. They don’t have experience interviewing at big tech companies. All they need is to see examples - what does it look like when someone actually does this?” The setup requires feeding Claude detailed project instructions - Ben gave Claude his distilled interview advice as project knowledge so it could demonstrate the exact signals interviewers look for. âžĄïž Use AI to see what “good” looks like in structured interviews. Watch Claude demonstrate proper signal delivery, then practice the same approach yourself until the frameworks become second nature. Check out Ben’s course (not sponsored, he’s just awesome).

​
Let’s build together

If you’re reading this, you’re in the top 5% of people leaning into AI agents for real work (even if it doesn’t feel that way, trust me). To go deeper, join me for a live, hands-on workshop. To get your team on board, book an AI build sprint.

​
Get notified about future posts

Subscribe