The traditional workflow (Idea -> PRD -> Alignment) is outdated. Now, PMs first create a functional AI prototype. This visual, interactive artifact is then brought to engineers and scientists for debate, accelerating alignment and making the development process more creative and collaborative from the start.
The "AI PM" title is a temporary distinction that will become redundant. The expert view is that within a few years, all products will have smart functionality. As a result, every Product Manager will de facto be an AI PM, and the specialized title will become obsolete, just like "Internet PM" did.
If your company lacks access to modern AI tools, don't see it as a blocker; view it as a leadership opportunity. Create a concise 'one-sheeter' outlining specific use cases, estimated hours saved, and productivity gains. Presenting a clear business case can turn hesitant leadership into champions for modernization.
Hiding the use of AI to create product artifacts is a mistake born from insecurity. Google AI PM Marily Nika advises PMs to be transparent, even sharing their custom PRD generators. This normalizes AI usage and reframes the PM as an efficiency leader, as those who don't adopt these tools will be left behind.
Notebook LM is a powerful tool for interview preparation. A Google AI PM uploaded a four-hour investor video and the target job description, then asked the AI what she needed to know. It distilled the content into 15 key points, enabling her to master the material and excel in the interview the next day.
To break into AI product management, avoid giant leaps. Instead, move adjacently by leveraging your unique background. For example, a professional with experience in hearing aids is a perfect fit for a PM role on Apple's AirPods hearing aid feature. Your domain expertise is a powerful, non-obvious differentiator.
A common red flag in AI PM interviews is when candidates, particularly those from a machine learning background, jump directly to technical solutions. They fail by neglecting core PM craft: defining the user ('the who'), the problem ('the why'), and the metrics for success, which must come before any discussion of algorithms.