How AI reshapes the product role — Lenny's Podcast

What the video covers

Oji Udezue (former CPO at Calendly and Typeform, product leader at Twitter, Atlassian, and Microsoft) and Ezinne Udezue (former CPO at WP Engine, VP Product at Procore) join Lenny Rachitsky to discuss how AI is changing the product management role. Despite their seniority, both are taking beginner AI courses and learning from engineers half their age — a detail that sets the tone for a conversation grounded in humility rather than proclamations about the future.

Who it’s for

Senior PMs and product leaders who want a balanced perspective on how the PM role is evolving — one that acknowledges both what is changing and what remains the same. The conversation avoids the hype cycle and focuses on practical patterns from companies that are succeeding with AI adoption.

Key takeaways

1. The “shipyard” framework embraces controlled chaos. Instead of linear product development processes, the Udezues advocate for a model where multiple AI experiments run in parallel, with the expectation that most will fail. The shipyard metaphor captures this: many ships being built at once, at different stages, with the organization learning from each one. This requires a tolerance for messiness that traditional product organizations often lack.

2. PMs are now the bottleneck, not engineering. Engineers equipped with AI coding tools are moving so fast that PMs struggle to keep up with product definition, customer insight gathering, and decision-making. The shift is disorienting for PMs accustomed to waiting on engineering capacity. The implication is that PM skills need to accelerate — faster customer validation, faster prioritization, faster communication of intent.

3. “Sharp problems” cut through AI hype. The Udezues introduce a methodology for identifying problems that are specific and well-defined enough for AI to address effectively. Vague problem statements like “improve customer experience” lead to vague AI implementations. Sharp problems — like “reduce the time a support agent spends searching for a customer’s billing history from four minutes to thirty seconds” — produce measurable results.

4. AI at the core vs. AI at the edge determines competitive position. Companies building entirely new AI-centric codebases will eventually outperform those that bolt AI onto existing products. The distinction matters for PMs making architecture decisions: an AI-at-the-edge approach is faster to ship but creates technical debt, while an AI-at-the-core approach requires more upfront investment but produces compounding advantages.

5. Curiosity, humility, and agency are the three skills that matter most. Rather than prescribing specific technical skills, the Udezues argue that the disposition of the PM matters more. Curiosity drives learning. Humility enables PMs to learn from junior engineers who understand AI tools better. Agency ensures PMs do not wait for permission to experiment.

Worth watching if…

You are a product leader trying to figure out which organizational changes are necessary to keep up with the pace at which AI is enabling engineering teams to build.