Why LinkedIn is turning PMs into AI-powered "full stack builders”

Tomer Cohen, Chief Product Officer at LinkedIn, discusses how LinkedIn turned its feed from a promotional space into a place for professional knowledge exchange, and how he led the company's shift to an AI-first product mindset. The conversation covers his "we might be wrong but we're not confused" philosophy, setting ambitious goals, carving out test cohorts for bold bets, and how product leaders should understand their algorithms and data.

01Key learnings

• Prioritize clarity over being right: align the team on the problem and direction even if the bet might be wrong, since confusion is what truly stalls progress.
• Test whether a disagreement is real or just a misunderstanding by asking people to restate your point in their own words before debating.
• Make sure stated priorities match resourcing; if your top initiative lacks your best engineers, the priority is not real.
• Start goal-setting from the potential of the opportunity (e.g., how many professionals could benefit) rather than current metrics, then work backwards.
• For turnaround or minus-one-to-one products, carve out a randomized cohort of users to prove a new experience without being blocked by entrenched metrics and processes.
• Treat AI as the matchmaking engine of a marketplace and measure feed success by high-value downstream engagement rather than raw clicks.
• Product leaders should own AI fundamentals: the algorithm's objective, features, data collection strategy, and fine-tuning, rather than delegating them to engineers.
• In AI-first building, reframe the goal as controlling the ingredients rather than dictating the full experience, and let teams explore broadly before converging on a few top-down bets.

02Frameworks mentioned

Summary and takeaways written by PM Atlas; quotes are short excerpts. © the original author.