The Shifting Mandate of the Chief People Officer in 2026

The role of the Chief People Officer changed fundamentally sometime in the last eighteen months, and most of those in the seat face an entirely new wave of challenges.

I say this having spent the last quarter in rooms with a lot of them. Group Talent Directors at listed businesses. CPOs at PE-backed scale-ups. Heads of HR at AI labs where the median engineer earns what a public company CEO earned ten years ago. The job they signed up for is not the job they once thought they had.

Four themes arise again & again. None are trends. They are structural problems with no defined or time-proven answer, and most CPOs are wrestling with them in private.

Every organisation of consequence has, historically, understood itself through the people it develops. The milkround. The graduate scheme. The trainee programme. The associate class. Call it what you want. The idea was the same: bring in capable people at the bottom, give them real work slightly beyond them, let the good ones surface, and in ten to fifteen years some of them are your senior operators. It wasn't always efficient. It was how the company's lifeblood simply replenished.

That machine has stopped.

I spoke last month to the CPO of a mid-cap professional services firm, about 3,500 people. They told me billable hours at the associate level had fallen roughly 40% year on year, because the work is being done by an LLM layered on their document management system. Nothing unusual there you may think? The partners are delighted. However, the pipeline of future partners has quietly collapsed. They asked me, only half joking, whether I'd be identifying their firm's senior leadership in 2035, because they couldn't see where those people would come from in an organic sense.

This isn't one firm. I've heard versions of the same story from a Group Talent Director at a FTSE 250 industrials business, from a Head of Talent at a global bank, and from the CPO of a software company growing 60% a year who told me plainly that their engineering org has no junior rung anymore. They're only hiring senior ICs. The work that used to season a junior into a mid-level engineer simply isn't there.

Being an employer of choice in 2026 cannot mean what it meant in 2019. It cannot be ping pong tables, a learning stipend, and a good parental leave policy. The promise now is sharper and much harder to deliver: we will still develop you. We will give you the messy, ambiguous, human work that grows people, even when the efficient thing is to give it to a model. We will take the short-term margin hit to keep the lifeblood flowing, because otherwise we have no company in a decade.

A post-training researcher at a frontier lab earned, last year, more in total compensation than most FTSE 100 CEOs. I know because we shortlist those people. I also know what the bands look like inside a typical PLC's comp framework, because I've sat across from the people who designed them.

The two numbers are not in the same universe.

The HR profession trained itself on a set of assumptions: internal equity, banded fairness, pay ratios, broad market data from Mercer or Willis Towers Watson. That apparatus assumed a reasonably smooth distribution of compensation across skills. It assumed the labour market for any given role behaved, broadly, like a labour market.

For a growing slice of talent, it doesn't. It behaves like a capital market. Specific individuals carry asymmetric value. Total compensation for a handful of frontier AI researchers or elite software engineers is now set the way footballer transfer fees are set — by what one or two bidders will pay, not by a salary survey.

My conversations have made me believe that CPOs are now capital allocators across a bifurcated talent market, and most governance structures haven't caught up. The ones who crack this will pull away. The ones who don't will keep losing their two or three most valuable people and replacing them with averages.

What we now call the CPO job could be seen as two jobs duct-taped together.

Both are legitimate. Both are full time. You cannot do both to a high standard.

I watched this play out at a European manufacturing-meets-technology group earlier this year. Their CPO, a genuinely impressive operator, spent roughly 70% of their time on employee relations and policy work across seven jurisdictions. The strategic workforce redesign, the agent integration, the capability mapping — all of it got done in the margins or by consultants. They left after eighteen months.

This is where I hear the least confident answers. If thirty, forty, fifty percent of a team's output comes from agents, what does culture mean?

A Head of People at a mid-sized tech business put it to me over coffee in Dublin a few weeks ago. They said their engineering teams are writing less code than they used to, and spending more time reviewing, prompting, and judging output from models. The nature of the craft has changed. The things that used to bond engineers together — the shared late nights, the war stories about gnarly bugs, the apprenticeship from senior to junior — are thinner on the ground. A culture they had spent five years building is quietly dissolving, and they don't yet know what to put in its place.

None of this is easy. All of it is possible. The gap between the CPOs doing this work and the ones still running the 2019 playbook is widening every quarter, and it is increasingly visible from the outside. Candidates notice. Boards notice. Investors are starting to notice.

The lifeblood question is the one that will haunt us longest. The choices made about junior talent in the next two or three years will determine which firms have senior leadership teams worth leading in the 2030s.

That's the bet. Most people aren't making it consciously. The ones who are will be the ones worth working for in ten years.