Why Half of Product Managers Are in Crisis (and How to Get Out)

Nikhyl Singhal worked at Meta, Google, and some of the most mature product organizations in the world. When he writes that half of all product managers are in existential crisis, it’s not rhetorical exaggeration. It’s a diagnosis from someone who watched the profession scale—and is now seeing the correction kick in.

What’s happening isn’t simply “AI will replace PMs.” That would be easy to process. The real problem is more structural: the profession expanded without ever settling the question of what actually differentiates a PM from a well-paid project coordinator.

The growth that created the problem

Product management became the coveted tech career over the last decade. High salaries, “mini-CEO” status, the promise of influence without needing to code. The market responded to demand: bootcamps, certifications, frameworks, books, podcasts.

The result was an explosion of professionals who articulate well, speak fluent product, know the rituals—but don’t necessarily deliver measurable business results.

The problem isn’t that these professionals are incompetent. It’s that the role definition became vague enough to accommodate very different profiles under the same title. And when pressure hits—whether from layoffs or automation—the first to be questioned are those whose impact is hardest to demonstrate.

The two types of PM that exist today

Singhal makes a distinction I find precise: there are PMs who are owners of outcomes and PMs who are process facilitators.

Facilitator PM

• Runs ceremonies and aligns teams
• Translates requirements across departments
• Ensures the backlog is prioritized
• Measures success by features shipped

Outcomes-Driven PM

• Defines which problem to solve and why
• Makes trade-offs based on impact
• Owns responsibility for business metrics
• Measures success by results generated

The facilitator PM isn’t useless—in some organizations, that role is genuinely necessary. The problem is when this profile presents itself (and is paid) as if it were strategic.

AI accelerates this distinction. Automation tools can handle much of the facilitation work: summarizing meetings, organizing backlogs, generating initial specs, synthesizing user feedback. What they can’t do is decide what’s worth building and why.

Why the crisis is structural, not technological

It would be convenient to blame AI. But the PM crisis existed before ChatGPT. The technology just accelerated a correction that was already forming:

1. Many companies hired too many PMs too early. Growth-stage startups built product teams based on benchmarks from mature companies. The result: bloated structures for the company’s actual stage.

2. The role was sold as strategic but operated as tactical. The “mini-CEO” narrative created expectations of influence that rarely materialized. Many PMs discovered they were, in practice, backlog managers with access to leadership meetings.

3. Success metrics never got defined. In engineering, you deliver code that works. In sales, you hit quota. In product, what do you deliver? That question never had a standardized answer—each company invented its own.

What separates those who survive from those who become obsolete

Singhal’s analysis points to a set of competencies that are becoming non-negotiable. They’re not new skills—they’re always differentiated the best PMs, but they’re now table stakes:

• Ability to define problems worth solving (not just prioritize solutions)
• Data fluency sufficient to question analysis, not just consume it
• Skill to make decisions with incomplete information and defend the reasoning
• Real understanding of how the business makes money
• Ability to say no—and sustain the position

What stands out about this list is that nothing is about tools or frameworks. Everything is about judgment, clarity of thought, and accountability for results.

The AI paradox for PMs

Here’s what few are talking about: AI can be the biggest threat or the biggest lever for product managers—depending on how you position yourself.

If your value is organizing information, synthesizing inputs, and producing documentation, you’re competing with tools that do it faster and cheaper. The math doesn’t work.

But if your value is deciding what to build, AI becomes an amplification tool. You can test hypotheses faster, analyze more data, prototype more scenarios. The bottleneck shifts from execution capacity to quality of judgment.

What to do if you’re a PM today

I’ll be direct: if you’re reading this and recognize yourself in the “facilitator” profile, the time to adjust is now—not when the next round of layoffs arrives.

First, get clarity on your impact. Can you point to a business metric that improved because of a decision you made? Not a feature you shipped—a metric that moved. If you can’t, that’s the gap to close.

Second, get close to the money. Understand how your company makes revenue, where the margins are, what moves the needle. PMs who don’t understand the business model become executors of someone else’s roadmap.

Third, use AI to amplify, not to look busy. Generating more specs, more PRDs, more analyses isn’t progress. Making better decisions, faster, is.

Fourth, accept that the title might change. Maybe what you describe as a PM role today gets called something else in three years. What matters isn’t the name—it’s the competency to define what’s worth building and ensure it gets built right.

The correction coming

The product market will go through a correction. Not apocalyptic, but real. Smaller teams, higher expectations, less room for ambiguity about who delivers what.

For those well-positioned—with clarity of impact, business fluency, and decision-making ability—this is an opportunity. Less noise, more space for those who actually solve things.

For those who built careers on rituals and facilitation, the moment is for honesty. The market that allowed that is shifting. Adapting now is choice. Adapting later is reaction.