AI is a tool, not a fundamental change to the product management discipline. The core competencies—understanding the user, defining the 'why', and driving outcomes—remain the same. Fluency with AI is becoming a baseline expectation, not a specialized role.
Product managers should be able to use any AI system because the fundamental interaction principles are the same, regardless of the specific tool. The skill is in the thinking and interaction pattern, not fluency with a single branded product like GPT.
When joining a new company or industry, it's impossible to absorb a decade of accumulated knowledge in weeks. Product managers must be patient and reject the '90-day plan' pressure, as deep learning realistically takes months.
Data and metrics are essential but incomplete; they lack insight into user motivation. To truly understand the 'why' behind user behavior, PMs must engage in qualitative research to uncover users' feelings, thoughts, and wants, which dashboards cannot capture.
A common mistake in new product development is worrying about feature parity (table stakes). The initial focus must be on building the fundamental, non-negotiable core of the product (the table). Without it, nothing else matters. The goal is to get feedback as fast as possible.
While engineers focus on what's feasible, a PM's job is to be the dreamer, pushing for the ideal user experience even when told it's too hard. This 'unreasonable' conviction forces the team to find creative solutions and prevents settling for mediocrity.
While consumer products can win on ease of use, enterprises primarily evaluate products based on their ability to increase profit or lower costs. Ease of use must be framed as a secondary benefit that drives those primary goals, such as saving employee time.
Complex AI development uses a pool of specialized agents. Like ants building a hill, some are workers, some are managers, and some review and discard bad code. This collaborative, layered system produces emergent results without a single orchestrator.
A powerful evaluation technique is to ask an AI agent to analyze its own poor output. The agent can review its context and process, explain why it made a mistake, and even suggest how to update its own instructions to prevent future errors.
While it's valuable for PMs to learn to build, they shouldn't forget their primary value. Their superpower isn't writing production code, but deep product sense and user empathy. Embrace new skills, but hone your core strengths to avoid becoming a lesser engineer.
Your first users are often tech enthusiasts happy to use a new thing because it's cool. The real test is scaling to users who don't care about your product or vision; they just want a stable, effective solution. This requires a different mindset and a higher quality bar.