AI is changing who you should hire. Here’s how to get it right

“We need someone who’s done this before.”

Translation: we need someone who can absorb a strategic pivot, upskill personally for AI, manage a workforce whose skills and expectations are shifting, maintain execution velocity, and make faster and better decisions—with the same budget, the same headcount, and no additional runway.

That’s not a job description. That’s a superhero spec.

And the person most organizations reach for to fill it—the candidate with deep sector experience, the safe hire, the one who’s “done this before”—is often exactly wrong for what the role now requires.

The logic behind the experience filter is not irrational. Sector knowledge compresses ramp time. It signals credibility with peers. It reduces the number of things that can go wrong in the first ninety days. When the environment was stable and execution was the output, it was a reasonable proxy for readiness.

That environment is gone.

AI has compressed execution timelines and put judgment at the center of competitive advantage. The work that once required a team now requires one person with the right capabilities. And the capabilities that matter most—operating without a playbook, making decisions under uncertainty, building alignment across functions—are not what the sector-experience filter selects for. It selects for pattern reproduction. In roles that now require pattern disruption, that’s not risk reduction. It’s risk amplification.

New criteria

A recent Strategy Science study, summarized by HEC Paris, found that within-industry breadth combined with cross-functional experience predicts stronger strategic foresight than narrow same-sector depth—particularly under conditions of uncertainty. The implication is uncomfortable: the profile most organizations default to in hiring may be the profile least suited to the moment they’re hiring for.

The middle management layer is where this mismatch is most expensive.

These are the people being asked to do something genuinely unprecedented: translate executive vision into execution reality, interpret and validate AI outputs, manage a workforce in transition, and make judgment calls faster—simultaneously, at the same level of quality, with no increase in resources. Every one of those demands has escalated in the last two years. None of them has been removed.

According to Gartner research cited by HRDive, 75% of business managers are overwhelmed by growing responsibilities, and 82% of HR leaders say managers are not currently equipped to lead change. AI is not relieving this pressure. It is adding a new layer: managers must now decipher AI initiatives, test tools, validate outputs, and explain limitations upward—while managing fewer junior staff to absorb the work.

This is the job that exists. It was built incrementally, requirement by requirement, until it became something no single person was designed to do. And the response—find someone who has done this before in our industry—does not solve the problem. It fills the role with someone selected for the conditions that no longer apply.

The cost of the wrong hire

The economics make this harder to ignore than most leadership teams have allowed themselves to.

The visible cost of bringing in a judgment-first hire without deep sector background is real: structured onboarding, longer ramp time, investment in building context deliberately. Organizations weigh that cost and reach for the familiar.

What they are not weighing with the same rigor is the cost of the wrong hire. Research from the Recruitment and Employment Confederation, cited by Gatenby Sanderson, estimates that a mid-level manager earning around £42,000 can cost a business more than £132,000 once recruitment, training, wasted salary, and lost productivity are included. That figure does not capture decision drag—the slower decisions, the missed pivots, the team that stalled waiting for direction that never came with sufficient clarity.

Organizations are making some of their most consequential talent decisions without serious cost data on either side of the equation. The familiar choice feels cheaper. It often isn’t.

The supply side

The math is also running out on the supply side.

According to ATD, middle manager hiring has fallen 43% since 2022—more than three times the drop in entry-level hiring. The experienced cohort that has historically filled these roles is aging toward retirement. The replacement cohort is smaller and carries less of the accumulated sector depth that organizations currently require as a baseline. At the same time, Deloitte’s 2025 human capital research finds that the work itself is shifting—AI is automating administrative and coordination tasks, increasing the need for managers who can coach, interpret ambiguity, and build alignment across boundaries.

Organizations are trying to solve a new management problem with a labor-market assumption that is breaking down. The experienced sector hire will become harder to find, more expensive to attract, and less suited to the actual job—in that order, and faster than most hiring plans currently reflect.

None of this is an argument for discarding experience. There are roles where deep sector knowledge is genuinely non-negotiable—where regulatory context, technical domain, or client relationships make it irreplaceable. The problem is that organizations apply the sector-experience filter uniformly, across roles where it matters and roles where it has simply become the default. Most have never made that distinction explicitly.

The organizations making progress on this are not overhauling their entire talent strategy. They are running contained experiments: small teams, high-performing, curious, change-ready. They are measuring what happens when the hiring criteria shift. They are designing for learning before they design for scale.

That is how you find out whether the model works before the hiring math forces the issue.

The question worth sitting with: in your organization, which management roles genuinely require sector depth—and which are using it as a shortcut?

Have you ever asked?

View the full article