You Can’t Hire Your Way Out of an AI Strategy

I read a job description last week that I haven’t been able to stop thinking about. It was for an AI Product Manager, and I want to walk you through what it asked for.

The successful candidate would run at least ten customer interviews a week and keep an active cadence with channel partners. They’d identify, size and sequence opportunities for agentic AI to displace vendor software and manual workflows. They’d write detailed PRDs covering agent scope, autonomy levels, human-in-the-loop checkpoints, eval approaches and platform dependencies. They’d build and stress-test pricing models with RevOps for every proposed agent. They’d present business cases to the CTO and executive team, orchestrate engineering squads across product lines, and then partner with sales enablement, product marketing and channel partners to launch the whole thing.

I read it twice, because the first time I assumed I’d misunderstood and it was describing a team. It wasn’t. It was describing one person.

Organisations are looking for unicorns, and unicorns don’t exist

Product management has always been a broad role. That breadth is part of what I love about it. But breadth and depth are not the same thing, and we keep pretending they are.

None of us can be equally deep in everything, and that isn’t a personal shortcoming. Depth costs time and attention, and both are finite. The research on this is remarkably consistent, and it shows up in fields as different as academic research and medicine: people who focus on a narrower area go deeper and have more impact than people spread across many domains. So each of us skews towards a strength. One product manager is brilliant at field research. Another is the person you want writing a product spec. Someone else can build a commercial model that survives a CFO.

You can be genuinely excellent at one or two of the things in that job ad. Asking for expert-level depth in all of them isn’t asking someone to stretch. It’s asking them to defeat a basic constraint on how expertise gets built at all. And it isn’t free to ask. The search will drag on for months, the salary will be eye-watering, and when you do finally hire someone, they’ll be quietly mediocre at most of the job, because that’s the only way anyone could do all of it.

The job ad is a symptom

Here’s the part I think will ruffle some feathers. I don’t think that job description was written by a careless hiring manager. I think it was written by an organisation that hasn’t done the hard work of deciding how AI fits into the way its teams operate, and has tried to solve that by putting the whole problem into one role.

The data on this is pretty damning. McKinsey’s 2026 research found 88% of organisations are experimenting with AI, 81% report no meaningful bottom-line impact, and only 1% describe their rollouts as mature. When leaders are asked what’s holding them back, they don’t point to technology. They point to change management and siloed ways of working.

So we’ve bolted AI onto an operating model that was designed for a different era, and when that doesn’t produce results, we reach for a hire. If we can just find the right person, the thinking goes, they’ll join up discovery, engineering, commercials and go-to-market on their own, and the transformation will happen around them. That’s not a role. That’s an org design you haven’t done yet, dressed up as a job ad.

BCG’s research found the organisations actually capturing value from AI are three times more likely to have redesigned their workflows end to end, rather than running isolated pilots on top of the old structure. No single hire does that for you.

And we’re screening for the wrong things

The other thing that bothers me about ads like this is what they emphasise. They’re heavy on AI fluency and light on everything else, which mirrors a mistake I see in a lot of product people right now. They’re pouring all their development energy into AI skills, as if that’s the only capability that matters.

It’s one technical skill among many. Employers themselves say so when you ask them directly. The World Economic Forum’s Future of Jobs Report puts analytical thinking at the top of the list of skills employers see as critical, with AI and big data sitting eleventh. Korn Ferry found 73% of talent leaders name critical thinking and problem-solving as the single most-needed skill for 2026. The person who stands out isn’t the one who can use the tools. It’s the one who can look at what AI produces, spot where it’s wrong, and decide when to trust it and when to override it. That’s judgement, and it’s built the old-fashioned way, through practice, feedback and a lot of decisions that didn’t go to plan.

To be fair to the hiring managers, the title itself isn’t helping. AI Product Manager postings have nearly tripled in two years, and it’s already one of the most inconsistently scoped roles on the market. What it means depends almost entirely on who’s hiring. In my view, the role does demand real technical fluency, including an understanding of architecture and production trade-offs, but the accountability should stay with the product outcome, not the codebase. When a role description can’t make that distinction, it’s usually because the organisation hasn’t made it either.

Design the team, not the person

If you’re about to write one of these job ads, I’d ask you to stop and try something harder first. Take that list of responsibilities and ask which of them genuinely need to sit with one person, and which belong to the team. Work out where AI takes on part of the load, and where a human needs to own the call. Decide how your product, engineering, commercial and GTM people will actually coordinate now that the boundaries between their disciplines are getting blurrier. Then hire for the gaps that are left.

It’s slower and less satisfying than posting a role and waiting for a hero to apply. But the hero isn’t coming, and the organisations that pull ahead over the next few years will be the ones that stopped waiting for them.

If you’ve written one of these ads, or applied for one, I’d love to hear how it went.