Are product managers cooked? (April 2025)

You’ve seen the article titles, you’ve read the tweets - devs/designers are cooked, AI will take over, etc, etc. But what about product managers?

This post goes over how AI is fundamentally changing (or sunsetting?) product management (as of April 2025) - the Rorschach test of the tech industry (separate post about that soon). Engineers question if PMs add value beyond slides and meetings. Executives sometimes view them as glorified project managers with inflated titles.

Indeed AI is already here

One-fourth of startups in YCombinator's most recent cohort have approximately 95% of their codebases generated by AI. The vibe coding revolution is here and it’s allowing a fairly technical person to ship. The 0 to 1 journey seems to be getting easier by the minute. Or is it?

Same goes for design. We see products like Bolt and v0 creating solid UIs directly out of the box. And we have multiple desgin-specific apps like Galileo UI ramping up.

So surely now with all the advanced LLM models and the uprise of agentic tooling AI can automate product managers. We can have our PRDs neatly written, our stories stacked in Jira and whatever else those folks are doing, right?

So let’s get rid of PMs

Many folks are starting to embracing this change - from creating a single use case to trying to automate as much possible.

The list of PM tasks now getting the AI treatment is extensive:

• Roadmapping and prioritization - roadmap.sh

• Analyzing customer feedback - Dovetail

• Creating user personas - userpersona.dev

• Market and competitor research - Perplexity

• Writing (endless) documentation - pick any model

And that's just the standard stuff. Lovable even lets you vibe-code your wireframes and then no-code them afterwards. So now you can have a prototype instead of your weird Miro drawings (mostly talking about me).

Going even deeper, AI is capable of analyzing how users interact with your product at scale, identifying patterns, bottlenecks, and trends that would take humans weeks to spot. It can suggest A/B tests based on these insights, and then analyze the results.

OK, cool. No PMs. Then what?

AI seems ready to automate (or at least dramatically enhance) most of a PM's traditional workload. But building products isn't just about engineering though, right? There are multiple non-engineering elements.

Here's the thing—despite all this automation, the most crucial aspects of product management still remain deeply human (as of today). Building relationships with your team, truly understanding user needs aren't just responsibilities—they're human-centered skills that AI can't fully replicate (yet).

Empathy isn't something you can prompt into existence as of April 2025. The ability to read between the lines in a customer interview (my take on user needs), to notice the subtle shift in someone's expression when a feature doesn't work as expected, to build trust with your cracked dev who doesn’t seem to like you at first — these skills remain uniquely human (still). And last but not least - do not underestimate your gut.

Enhanced product management

So rather than sunsetting the product management role, I think AI is extending it. I see a repeating pattern here with an earlier experience of mine - the ROI of my obsession with no-code tools 5-6 years ago has been tremendous (spinning up landing pages, prototyping, embedding stuff, etc.) , and I think the situation with AI is very similar. That's why I'm relentlessly trying out new models, tools, and weird apps - I'm confident the payoff will be just as great.

Product leaders are starting to acknowledge this shift. Yana Welinder’s podcast with Peter Yang and his article on the different areas he use AI for his work such examples.

We're also seeing this shift at companies like Spotify, where AI usage is a requirement and with Figma’s Dylan Field who have been vocal about AI amplifying rather than replacing.

The path forward.

Every hour you save from writing documentation or creating wireframes is an hour you can spend on what really matters: deep user understanding, creative problem-solving, and building the relationships that make great products possible.

I recently had a chat with two developer friends at a fairly-sized fintech. When I showed them Replit, they were blown away. But I was even more surprised that they were seeing these capabilities for the first time through me. It felt like they were falling behind - and that's when you are cooked.

Product management isn't cooked (yet) — it's evolving. I believe the same goes to engineering or design. The future belongs to the Generalist - PMs who leverage AI to ship products themselves or engineers who go-to-market with useful products. So it feels like the lines are getting blurrier and blurrier.

So go ahead, let AI handle your next PRD. Then use the time you saved to talk to an actual human about what they really need - because that human connection remains the most crucial skill in your toolkit.

The key takeaways here are as follows: experiment with AI, use the new tools, try to automate yourself, educate others and trust your gut.

Do share this post if you found it useful! And subscribe to product3.