How the PM role is changing, and what good looks like now

The Product Manager role has been through multiple reinventions since it was codified in the eighties. It's absorbed design thinking, agile, lean startup, growth hacking, and a dozen other movements. Each time, the job expanded, the expectations shifted, and the definition of 'good' changed.

What's happening now is different in character from those previous shifts. It's not a new methodology being absorbed into an existing role. It's a change to the fundamental conditions in which the job operates. Understanding it, as a PM, a hiring manager, or someone placing PMs, requires being specific about what's actually changing and what isn't.

What's actually changing

The speed of product iteration has accelerated significantly. In teams using AI tools well, the gap between idea and working prototype has collapsed. This changes the economics of exploration, you can now test more ideas, faster, with less investment. For PMs, this is a gift and a trap. It's easier to build things. It's not easier to know what's worth building.

The information advantage that PMs once held over their teams is disappearing. A well-prompted AI assistant can produce competitive analysis, user research synthesis, PRD drafts, and market sizing in minutes. This used to be some of what PMs were hired to do. It isn't anymore, or at least, doing it yourself is no longer the point. The point is knowing what to ask for, judging the quality of what comes back, and making decisions with it.

The bar for technical literacy has risen without becoming a bar for technical expertise. PMs don't need to write code. They increasingly need to understand how AI systems produce outputs, what the failure modes are, how to evaluate whether a model is doing what it's supposed to do, and how to have a real conversation with an ML engineer about tradeoffs. This isn't a specialist skill set. It's baseline fluency, and the gap between PMs who have it and PMs who don't is already widening.

What isn't changing

The core of the job is still the same: understand what a group of people need, figure out how to build something that serves that need, and get a team aligned on doing it well. No tool changes this. No methodology replaces it.

Good judgement is still the scarcest and most valuable thing a PM brings. The ability to decide what not to build is still harder than the ability to describe what to build. The skill of holding a clear view of what matters while managing a team full of competing priorities and opinions is not automated away by anything currently on the market.

The relationship skills, with engineering, with design, with leadership, with users, remain determinative. A PM who is technically sharp and analytically rigorous but can't work effectively with people is still a PM who will underdeliver. This hasn't changed.

What 'good' looks like now

A strong PM in 2025 is not defined by their tool stack. It's defined by how they think.

They have a clear relationship with uncertainty. They can make decisions with incomplete information, name their assumptions explicitly, and update their view when the evidence changes. They don't confuse confidence with certainty.

They are genuinely curious about what's changing and honest about what they don't yet know. The PMs who are adapting well to the current environment are not the ones who knew about AI early. They're the ones who are willing to change how they work and are honest about what that process looks like.

They produce outcomes, not activity. In an environment where the tools make it easier than ever to generate deliverables, PRDs, research reports, competitive analyses, the PMs who stand out are the ones who are ruthless about whether the output is actually useful. They write less and think more. They ask whether the document needs to exist before they write it.

They understand their environment. The best PMs know what kind of PM they are, what conditions bring out their best work, and where they're likely to struggle. This self-knowledge is what makes them honest in the interview process and effective when they arrive.

What this means for hiring

It means the signals that mattered five years ago, impressiveness of the company on the resume, seniority of the title, breadth of experience, are less reliable than they used to be. A PM with a great title at a well-known company who hasn't updated how they work in two years is already behind. A PM at a less recognisable company who has been experimenting with new tools, updating their process, and thinking carefully about what good looks like now may be significantly ahead.

It means the interview has to probe for this directly. Not 'do you use AI tools?' but 'what has changed about how you work in the last two years?' Not 'are you familiar with evaluation pipelines?' but 'describe a time your confidence in a product decision was wrong, and how you found out.' The answers reveal something that the resume can't.

And it means that the people assessing candidates need to understand the role well enough to recognise the signals. That's always been true. It's just more consequential now.