Hiring data on AI product management roles shows a pattern that cuts against how most people assume AI is changing software teams. Postings are running at a steady ~714 new US roles a week — not spiking, not correcting, just a persistent baseline — and 47% of them target manager-level hires, against only 2% junior. Demand has spread well past AI-native companies: technology and financial services post the most roles, but enterprise-scale organisations with 10,000-plus employees now account for 41% of the market, meaning traditional, non-tech companies are building this function in-house rather than treating AI product work as a specialist agency function.
The read: fewer generalists, more owners
The skew toward experienced hires rather than junior generalists tracks with a broader shift agentic coding tools have caused: when a capable AI agent can execute a well-scoped task, the scarce skill isn’t execution anymore, it’s judgement — knowing what to build, what to cut, and how to scope a problem so an AI-assisted team can actually deliver it. Companies aren’t hiring fewer people to manage software; they’re hiring more senior ones, because the job has shifted from coordinating a large team’s output to directing a smaller team working faster.
So what
If your own organisation is restructuring around fewer, more senior owners rather than a large team, that has a direct implication for how you commission external development too: the value in a build partner increasingly sits in judgement about scope and architecture, not headcount thrown at the problem. That’s the model we work to — senior involvement across a build rather than a large team billing hours. See our services or start a project to talk through scope with someone who’ll actually own the decisions on it.