A track record tells you how someone performed in a receding world. Hiring well now means reading for the one that is arriving...
Not every prediction about AI and the corner office is worth taking seriously.
The genre runs ahead of itself. Boards seating AI agents next to the directors. The CEO as curator of machine-generated insight. The org chart dissolving into something that assembles itself around problems. Some of it is already happening, in pockets. Most of it is the sound of an industry talking itself into a future it cannot yet see. For a technology this good at prediction, the predictions about it are remarkably bad.
Underneath the noise there is a real shift, and it lands on the part of the business where I spend my time.
I have spent my career running C-level executive search across technology, commercial, and financial functions, for private equity and venture-backed, founder-led B2B and B2B2C technology businesses. A good share of that work is for European and Nordic CEOs, investors, and founders hiring senior US leaders, which means I often run a search for a role that does not yet exist within the company or that they have tried and failed. I have also done the other version of this work, building and running talent functions inside the kind of companies I now recruit for. From either seat, the question is the same. What are we actually hiring this person to do, and what evidence can we verify that they can do it?
For most of that time, the answer to the second half was a track record. You hired the CFO who had taken a company public, the operator whose numbers were a matter of record, the commercial leader who had carried a number through a particular order of magnitude. On my team we call it execution pedigree, the things a search has always read well. Scale, M&A, value creation, the go-to-market motion, functional and domain depth. It was most of the read, and a good one. The past was the best available evidence of the future.
That assumption is weakening across every C-level role, and the CFO is where it loosens first.
A recent piece in Harvard Business Review by Tomas Chamorro-Premuzic put a clean frame around why. His argument is that AI is commoditizing expertise at the top of the organization. For a century, leaders rose because they knew more than the people around them, and they proved it with the right degrees, the right experience, and a record against the conventional measures. A great deal of that knowledge is now available on demand. A model runs the scenarios, optimizes the supply chain, and reads the market faster than anyone in the building. When hard skills become easier to copy, the difference between one leader and another shifts to the parts that resist automation. Judgment under ambiguity, how fast someone updates their thinking when the evidence changes, and the self-awareness to know the limits of their own view. Chamorro-Premuzic is the chief science officer at Russell Reynolds, so this is a man describing the same market I work in, from the research side of one of its largest firms.
He makes it concrete with the CFO. Set how the role is described now against how it read a few years ago. The competencies trending up are data analytics, machine learning, cloud computing, and scenario modeling. The ones becoming table stakes are technical accounting, auditing, and regulatory mechanics. Those older skills have not stopped mattering. They have become the floor. The CHRO is on the same path, pulled from administering people toward designing how people and machines work together. The repeatable work is automating. The rest is what you hire for.
The market data says the same thing in colder terms. Protiviti’s most recent global finance survey found AI use among finance teams at over 70%, more than double the year before. BCG’s work on the widening AI gap found that leaders who have built around the technology are pulling away from their peers in cost, margin, and return on capital, and that the gap is compounding. And Gallup, surveying senior executives at the start of this year, found only around one in eight saying AI had changed how the work actually gets done. Most organizations are not there yet, and that gap is the one a hire either closes or widens.
The track record is only half the read
Here is the practical problem. A track record is a record of the past, and the part of the job changing fastest is the part no past contains.
So we read for a second thing now, alongside execution pedigree. For lack of a better phrase, we call it AI skill and will. Not whether a candidate can talk about AI. Whether they have personally changed how they think and decide because of it. Those are different people, and the difference is easy to miss if you are only listening for the right words.
The tell is specificity. The leader who has actually rewired their own work names the tools they use and the decisions those tools changed. They can give you a before-and-after from their own week, not their team’s roadmap. They reason in probabilities when the ground is uncertain, rather than waiting for the ambiguity to resolve before they commit. Ask them about a strategic bet that went wrong, and they can name the precise moment they realized it, and the signal they had missed. The leader who has only sanctioned AI at the org level cannot do any of that. They speak about it in the third person. They cite the program, the governance, the metrics. They have handed exploration over to a chief AI officer and carried on deciding as they always have.
Both can look strong in a first conversation. Only one of them has changed.
What a title does not tell you
This is where the new boxes on the org chart get misread.
The chief AI officer has gone from rare to routine in about a year. By IBM’s count, roughly three-quarters of large organizations now have one, up from about a quarter the year before. That speed is itself the warning. The box gets added far faster than the capability behind it. A company that appoints a chief AI officer has not thereby acquired an AI strategy, any more than the chief innovation officers of the last cycle produced much innovation. The title names the problem. Whether anyone solves it depends on capability, incentive, and culture, none of which arrive with a business card.
The same trap shows up one level down, in the people being hired. The leader who sponsors the AI initiative is not the same as the leader who has been changed by it, and the first is much easier to mistake for the second. Sponsoring is safe and legible. It shows up in board decks. What is harder to assess, and what predicts how someone leads through a shift like this, is whether the person in front of you has done the work themselves.
Which means some of the signals a search has always relied on now need to be discounted. A named AI program on a resume is org-level housekeeping, not evidence of personal change. Headcount managed says less than it seems, because AI-era leadership is about the quality of the structure someone builds, and running four hundred people well beats running two thousand badly. Revenue posted in a strong market is not proof of the judgment you need at an inflection point, when the tailwind drops and the call has no precedent. None of these are worthless. They are weaker predictors than they used to be, and the danger is hiring on them precisely because they are the easy ones to see.
What does not move
The same commoditization has already reached the search business. A model maps a market and builds a long list in an afternoon, work I would have paid an analyst weeks for not long ago. Research is becoming a commodity, just as executive expertise is.
I have written before about where that stops. The judgment does not move. A model can tell me what a candidate has done. It cannot sit across from them and tell me whether the confidence is conviction or defensiveness, whether they will go quiet the first time a board pushes back, or whether the moment they claim to have changed their mind ever actually happened. It cannot judge whether a candidate fits the culture they are walking into, or whether they and the CEO will work well together once the first hard decision lands. That read is the job. It always was. The technology has only made it obvious by clearing away the parts that were never the point.
Hiring for the future tense
Where this leaves me, in practice, is hiring more for trajectory than for record. Reading both axes, and weighting the one the past cannot show you.
The questions I care most about are no longer only about what someone has done. They are about how a person behaves when they are wrong, whether they can hold a strong view and still revise it, and whether they can work past the edge of their own experience without pretending the edge is not there. Those qualities travel into a role whose contents will keep shifting.
None of this makes the search easier. It was always easier to hire the person whose past matched the page. The harder thing, and the more useful one, is to read a candidate for what they could become inside a business that will not sit still long enough for anyone’s experience to stay current.
The executives who do well in this period are not the ones racing the machines on knowledge they are going to lose anyway. They are the ones honest enough to see what the machines now do better, and secure enough to spend their own attention on the part that is still theirs. Judgment, the self-awareness to know which is which, and the willingness to keep learning in public. Those were always the things that separated the people who grew with a business from the people the business had to grow around.
The technology has only made the difference harder to ignore.
Sources: Tomas Chamorro-Premuzic, “How C-Suite and Board Roles Are Being Reshaped Around AI,” Harvard Business Review (2026), drawing on Russell Reynolds Associates data. Protiviti, Global Finance Trends Survey (2025). Boston Consulting Group, “How Leaders Build an AI-First Cost Advantage” (2026). Gallup, AI in the Workplace (2026). IBM Institute for Business Value, CEO study on AI and the C-suite (2026).
Corporate Dyak 