AI won't kill the PM role. Same job, better tools.
Generative AI is a powerful new tool. The core of the job remains exactly what it always was.
Please like ❤️ and restack 🔁 this article so others can find it, too. Thank you!
You have probably heard it by now: AI is coming for the PM role. Some say it will shrink the job. Others say it will eliminate it entirely. And if you are honest with yourself, there is a part of you that wonders whether they might be right.
Let’s settle this. Because the answer is not complicated. It is just buried under a lot of hype.
What actually changed, and what didn’t
Generative AI is a new tool. A genuinely powerful one. It can help you synthesize research faster, draft documents in minutes, and surface patterns you might have missed. Smart PMs are already using it, and they should be.
But what didn’t change is the thinking. The carpenter who gets a better saw still needs to understand what the customer wants built. The saw makes the work faster. It does not replace the judgment behind it. The same logic applies to every PM who picks up an AI tool today.
What is the PM job, actually?
Strip away the jargon and the frameworks. The job looks the same whether you are building a traditional SaaS product or an AI-powered one:
Understand your customer deeply. Not just their stated needs, but the problems they cannot always articulate. That requires real conversations, careful observation, and genuine curiosity. No model does this for you.
Create value for that customer. Value means enabling something they could not do before, or solving a problem that was costing them time, money, or frustration. Vague value propositions don’t ship.
Create value for your company. A product that delights users but destroys margins is not a success. The business side is just as much your job as the customer side.
Make it usable. Great ideas that nobody can figure out how to use are just expensive experiments.
Build a business model that works. How does this product make money? Who pays, when, and why? These questions do not answer themselves.
Define your go-to-market strategy. Who are you selling to, through which channels, with what message?
Know your ICP. Trying to build for everyone is the fastest way to build for no one.
Every single one of these applies to AI products. None of them disappear because your product uses a large language model under the hood. The questions are identical. The stakes are identical. The work is identical.
Why the “AI will kill PM” argument falls apart
This claim comes up constantly. On LinkedIn, on Reddit, in Slack channels. It sounds plausible at first. And then you think it through.
The job still needs to be done. Someone still needs to talk to customers, make sense of what they hear, decide what to build next, align stakeholders who disagree, and take responsibility when the call turns out to be wrong. AI does not do any of that. It assists. It accelerates. It surfaces information. But it does not decide, and it does not own the outcome.
Saying AI will eliminate the PM role is like saying GPS eliminated the need for drivers. Navigation got easier. The driving still requires a human making real decisions in real time.
What does a PM actually look like in an AI world?
More powerful. Not obsolete.
AI handles the repetitive and the mechanical. The summarizing, the drafting, the first pass at analysis. That frees up time and mental energy for the work that actually moves products forward.
What fills that space? The things AI cannot replicate. Taste: knowing what good looks like, even when the data is ambiguous. Differentiation: making choices that are not obvious, that competitors have not made. Courage: committing to a direction when reasonable people disagree and the outcome is uncertain. Judgement: knowing which signal to trust and which to ignore.
These are not soft skills. They are the job. They always were.
The PMs who will struggle are not the ones replaced by AI. They are the ones who never understood what the role actually required in the first place.
What I read
As usual, I will list some of the best articles I read on the Internet. I will keep a list of the best articles (currently >900) at https://www.digital-product-management.com. These are today’s picks:
The Leadership Leap We Don’t Talk About: The Leadership Domains framework distinguishes between functional and organizational leadership work.
Why your team keeps shipping products nobody buys: Aligning go-to-market strategy with product discovery prevents the gap between shipping features and achieving market success.
Displacement of Cognitive Labor and What Comes After: Artificial intelligence fundamentally shifts human identity, productivity ratios, and the nature of intellectual work through iterative collaboration.
Benefits
Save 20 USD when subscribing to Proton: Secure, end-to-end encrypted email, calendar, drive and more.
Save 10% on the first month when subscribing to Krater.ai: Every AI You Need.
One Subscription.
As Cory Doctorow said: It does not matter whether you can be replaced. It matters whether your company higher-ups _believe_ it and wish it. One of your readers, Uwe Mierisch, recently had an interesting article about the longer-term consequences of letting go a team that was deemed superfluous. https://uwemierisch.substack.com/p/todo-management / https://uwemierisch.substack.com/p/todo-management-de
I have read similar things about software developers / Tech people in general. My assumption right now is that the predictions do not tell who will be replaced (effectively, without regret and rollback) but whom the author does not like or respect.