We often talk about using AI for productivity—automating user story creation, summarizing research, etc. But the real shift isn’t just using AI as a tool; it’s integrating AI as a core, sometimes unpredictable, member of our product.
Think about it: when an LLM is a fundamental part of your user experience, you’re not just managing code anymore. You’re managing a learning entity. It has biases, it has ‘hallucinations,’ and its behavior can evolve in ways you didn’t explicitly program. This isn’t like a predictable API; it’s more like a new team member who is incredibly fast and knowledgeable but has zero real-world context and requires careful guidance.
This introduces a new layer to our roles. How do we write a PRD for a feature that’s non-deterministic? How do we build a roadmap when a core component is constantly learning and changing? We’ve spent years honing our skills in user research and cross-functional collaboration with engineers, designers, and marketers. Now, we need to learn how to collaborate with a non-human intelligence that is part of the product itself.
What’s the single most important skill product managers must develop to effectively ‘manage’ AI as a core, active component of their product?