Why Product Management Feels Harder Right Now

AI is causing a real crisis for product managers, and it is not just the usual fear that a new tool will change how we work. It is a crisis of identity, because PMs are being pulled in multiple directions at once. In most other job families, the AI engagement model is pretty predictable. In customer support, it shows up as triage. In sales, it shows up as call coaching and outbound assistance. In marketing, it shows up as creative and asset generation. With PM, it is different. AI is touching the work in at least three ways simultaneously, and the overlap is what makes it feel so destabilizing.

The first pressure is asset work. PMs are responsible for a lot of writing and synthesis, and AI can now draft PRDs, briefs, updates, and research summaries quickly. That sounds like a simple productivity win, but it is not. The skill is not just generating text. The skill is keeping your thinking, judgment, and creativity intact while being expected to move faster because the tool exists.

The second pressure is that PMs are increasingly asked to build AI products. And AI products do not behave like traditional software. They are probabilistic. They work most of the time, until they do not, and the edge cases are part of the product. Executives still want timelines, and engineering teams are understandably cautious about committing to deadlines when outcomes are uncertain. PMs end up stuck between organizational urgency and technical reality.

The third pressure is that PM has always been a glue role. The person who keeps engineers out of meetings and turns messy inputs into alignment. But now that AI can move information around so easily, people start to question which parts of that glue work are still uniquely human, and which parts can be automated. At the same time, the boundaries are shifting in the other direction too, because PMs are now expected to prototype, to vibe-code, to ship small changes themselves, and to blur into design and engineering in ways that used to be unusual.

The hopeful part is that the core of the PM craft has not changed, even if the surface area has exploded. The way through is not to lean out. It is to lean in with discipline.

First, PMs have to get technically fluent in how LLMs and agents actually work. Not as a buzzword, but at the level where you can have real tradeoff conversations with engineers about structure, tooling versus prompting, reliability, evaluation, and the practical constraints that determine what ships.

Second, you need to work on things that matter. A lot of burnout right now comes from being asked to push through AI initiatives that do not move the needle, often driven by executive anxiety or social pressure rather than customer value. That is demoralizing, and it teaches the wrong lessons.

Third, you cannot outsource product intuition. AI can help you go faster and see more, but taste, judgment, and direction still belong to you. If you defer too much to the tool, you do not just weaken the product, you weaken the craft you are trying to build.

And finally, the alignment job has not gone away. AI can help you draft materials, but it cannot replace conviction, negotiation, and the human work of creating shared reality between leadership and engineering.

That is still the job. AI is a tool in the toolkit. The PM is still the craftsperson.