AI-native PM interview trend
- Jessica Verrilli shared Bret Taylor's PM interview redesign: Plan, Build (two hours with AI tools), Review. - The process replaces traditional coding tests and focuses on product thinking, technical judgment and AI usage. - This signals interviewers are testing product sense and AI tooling fluency rather than only raw coding ability (x.com).
A product manager interview is starting to look more like a work session with ChatGPT than a whiteboard test. Jessica Verrilli said Bret Taylor now uses a three-part loop — Plan, Build, Review — for PM candidates. (x.com) Verrilli said the middle section gives candidates two hours to build with artificial intelligence tools, not a traditional coding exercise done without help. She described the redesign as a way to test how people think about products, make technical calls, and use AI in the process. (x.com) Taylor is the co-founder and chief executive of Sierra, the customer-service AI startup he launched in 2023 with former Google executive Clay Bavor. He also serves as chair of OpenAI’s board and previously ran products at Google, Meta and Salesforce. (sierra.ai, cnbc.com) The format departs from the standard PM loop, which usually splits candidates across product sense, strategy, analytics, behavioral and technical rounds, with take-homes used only sometimes. In that older model, technical screens often test how well a candidate explains systems or works through structured cases, not how they use live AI tools under a deadline. (thita.ai, tryexponent.com) That change arrives as Taylor has been publicly arguing that software work is moving into an “autopilot era” shaped by AI coding assistants and agents. In recent interviews, he has described developers less as pure code writers and more as operators who guide, verify and ship with machine help. (developers.slashdot.org, sierra.ai, theverge.com) Interview prep companies have already started adapting to that shift. Newer AI PM guides now emphasize model behavior, trade-offs, evaluation and uncertainty, while practice tools market AI-assisted mock interviews as part of standard preparation. (institutepm.com, ideaplan.io, fearcutter.com) Supporters of AI-assisted interviews say they reflect actual work, because many product and engineering teams now build with copilots, code generators and agent tools every day. Critics say the format can blur who did the work unless interviewers score planning, prompting, verification and judgment separately from raw output. (formation.dev, coreydaley.dev) For PM candidates, the signal is straightforward: knowing how to frame a problem, steer an AI tool, and audit what it produces is becoming part of the interview itself. The Plan, Build, Review loop turns that workflow into the test. (x.com)
