HR Got A Big Promotion

Why the Chief People Officer just became one of the hardest jobs in the C-suite

A short list, but not a complete list, of things the head of HR at a large company is now expected to have a credible opinion on: which jobs in the organisation will still exist in 2030; how much surveillance is acceptable in their productivity software; what to do when an AI hiring tool stops shortlisting diversity in engineering roles; whether the manager development program they spent four million on last year is still relevant; how to retrain 8,000 customer service staff without telling them what for, because nobody actually knows yet.

What's striking is how few of the old questions have actually gone away. Hiring still happens. Compensation cycles still run. Compliance still has to be filed. The engagement survey still goes out, and the same managers still ignore the results. The new responsibilities haven't replaced the old ones. They've been bolted on top, and the average CHRO is now expected to handle both with roughly the same headcount and budget they had five years ago.

This is part of what makes the role one of the hardest in the modern C-suite, and a big part of why most leadership teams are still pricing, and paying, it wrong. When the CFO absorbs new mandates (climate accounting, AI investment governance, geopolitical risk modelling) they usually get new tools and new headcount to do it. When the CTO picks up a new responsibility, there's a budget line attached. The CHRO, or CPO, has watched the scope of their role roughly double in five years and is largely expected to figure it out with what's already to hand. Fifteen years ago, all of this would have been unthinkable for someone whose remit was largely payroll, policy, and leave.

We're watching one of the more interesting shifts in the C-suite of the last forty years, and most companies still don't have a clean name for what's happening.

How HR got here

It's worth a quick look at how HR ended up in this position, because the trajectory is important. The function started life in the mid-twentieth century as "personnel," a back-office team that processed paperwork, ran payroll, and refereed grievances. Through the 80s and 90s it picked up compliance work as employment law thickened: discrimination claims, workplace safety, benefits regulation. The early 2000s brought what got called talent management. Performance reviews, competency frameworks, the first attempts to link people decisions to actual business outcomes. Most of those frameworks have aged poorly. The leadership competencies that looked sharp in 2008 read like museum exhibits now.

Then came "strategic HR," a phrase every CHRO has been deploying for the last fifteen years. The intention was right. Move HR from administrator to advisor. But the practical reality, in most companies, stayed administrative. The function ran the engagement survey, owned the annual review cycle, and kept the leadership programs ticking over. Real strategy got made by finance, technology, and operations.

So when people now say HR is changing, the more accurate statement is that HR is finally being asked to do the job it claimed to be doing two decades ago. The difference is that the stakes have changed enough that the claim is being tested for the first time, and not every CHRO is going to come out the other side.

What AI is actually doing to the function

The obvious thing AI is doing to HR is eating the transactional layer. Resume screening, interview scheduling, candidate sourcing, onboarding workflows, benefits questions, the eight thousandth policy clarification on parental leave. All of it is being absorbed by software, fast. Recruiting coordinators who lived in spreadsheet hell two years ago now spend their days reviewing what the AI shortlist got wrong. HR tech vendors that sold modules in 2022 are selling agents in 2026. The pace is unusual even by software standards.

The pattern is the same one that has played out in law, consulting, and software engineering. The transactional middle gets compressed. The genuinely judgemental top end gets more strategic. The bottom of the market gets eaten. People assume AI is hollowing out the People function. The more accurate read is that it's hollowing out the function's middle layer, while making the top of the function more important than it has ever been.

It's worth pausing here for a second, because the framing above can sound smoother than reality is. Most enterprise AI deployments are running behind schedule. Productivity gains have been real but uneven, with serious gaps between teams that have adopted the tools well and teams that have adopted them badly. Some of the most expensive rollouts of the last two years have quietly underperformed and never made it into the case studies. Manager skepticism is real, employee resistance is real, and there are functions where the honest answer is that the technology isn't yet ready for the work it's being asked to do. The CPO of a serious company has to be candid about all of this, because pretending the transformation is going better than it is destroys credibility faster than almost anything else they can do.

Once HR teams aren't drowning in admin, the questions executives bring to them get sharper. Not "how is the engagement survey looking" but "which of our roles are going to look completely different in eighteen months, and what should we be doing now." Not "are we hitting our diversity targets" but "are our managers ready to supervise teams that include software agents alongside humans." Not "what's our retention rate" but "where is burnout actually building, and is it building because of the work or because of the AI we just rolled out."

These aren't HR-as-administration questions. They're strategy questions that happen to live inside the people function. The more interesting CHROs have been preparing for them for a while now, often without much support, because their organisations were still treating HR as the department that runs the engagement survey.

From running HR to designing the workforce

There's no good single phrase for what the modern Chief People Officer actually does, but "workforce architecture" is closer than most. The old job was running an HR function. The new one is designing how the company actually works, and then redesigning it every couple of years as the technology underneath shifts.

Researchers and consultants who study this stuff have started using the phrase HRM 5.0 to describe where it's heading: human-centred organisations operating in AI-native environments. The label is awkward, and most CHROs would never say it out loud in a board meeting. But it captures the essential shift, which is that the People function is being redefined around how humans and machines work together rather than around how humans get hired, paid, and reviewed.

The CPO of a large company is now expected to have a defensible answer to which work humans should be doing versus which should go to AI. That sounds abstract until you have to apply it to a 400-person engineering department, where the answer changes the team structure, the hiring plan, the budget, and the career paths of everyone involved. They're expected to redesign roles so that people are doing the parts of their jobs AI can't do well, rather than the parts AI happens to be bad at this particular quarter. None of this is the kind of thing that fits neatly on a job description, and yet it's where the role is going.

This is where the research field that has started calling itself human-AI teaming lives, and the questions are harder than they first appear. How does a manager supervise a team where some members are software that runs continuously, doesn't get tired, and occasionally produces confidently wrong work? How does trust form between an experienced human contributor and an AI agent that handles part of their workflow, when the agent's reasoning isn't always inspectable? What happens to professional identity when half the work that used to define a junior role is now done by an agent, and what does the career path look like for the humans on the other side of that change? These aren't future questions. They're already showing up in engineering teams, customer service operations, and legal departments. The CPO is the executive who has to own the answers, not because they're a technologist, but because the answers have to hold up in front of the humans who have to live with them.

A couple of practical examples of where the leading work is going. Several large professional services firms are running what they call "talent operating systems": internal platforms that map every employee's skills, growth trajectory, and likely next role, with the CPO using them to forecast capability gaps a year or two out. A handful of banks have started running quarterly "workforce stress tests," modelling what happens to teams, productivity, and culture under different AI deployment scenarios before committing to anything. These aren't pilots dressed up in jargon. They're the early version of what the People function looks like when it grows up.

The CPO ends up answering questions that used to fall awkwardly between functions and never quite get owned. AI ethics in hiring lands here because nobody else is structurally responsible for it. Surveillance limits in productivity software land here because nobody else has the moral authority to draw a line. Whether the company is still a place worth working at after a third major reorganisation in two years lands here because, for better or worse, that's what people leadership has always been about, even when nobody was paying attention.

Talent intelligence is something most HR teams aren't ready for

For the last decade, the People function has been one of the least quantitative parts of most large companies. Finance had its dashboards, operations had its KPIs, marketing had its attribution models. HR had an annual engagement survey, a turnover percentage, and a hiring funnel that nobody quite trusted. The decisions that mattered most, who to promote and who to retain and which teams were actually in trouble, came down to manager judgement and gut feel, dressed up in whatever talent calibration framework happened to be in fashion that year.

That gap is closing, and it's closing in a way that's making a lot of senior HR people uncomfortable. There is now a whole category of tooling, broadly called talent intelligence, that uses workforce data to predict things companies have always wanted to predict but couldn't: which employees are most likely to leave in the next six months, which teams are running hot enough to start losing people, where leadership gaps will open up two years out, which skills are decaying inside the organisation as the broader market shifts. Some of this is genuinely useful. Some of it is overpromised vendor demoware. The CPO is the executive who has to know the difference and pick accordingly.

A related shift is happening on the inside of large companies. More of them are now building what get called internal talent marketplaces: platforms that match employees to projects, gigs, and short-term assignments based on skills and stated interests, rather than relying on managers to know who's available and what they can do. Done well, this becomes a serious mobility tool. The organisation deploys capability where it's needed without going outside, and employees get a way to grow without having to leave. Done badly, it becomes another unused HR portal that nobody trusts and managers route around. Which version a company ends up with depends almost entirely on whether the CPO governs it well and gets the rest of the leadership team to actually use it.

The deeper version of this shift, which is starting to show up in research from MIT Sloan and a few of the better corporate research groups, is something closer to organisational simulation. The idea is that you can model the workforce the way climate scientists model weather. Not predict a specific event, but generate a defensible range of outcomes under different scenarios. What happens if we deploy AI agents across customer service in Q3? What does it do to attrition, manager workload, the leadership pipeline two years out? The available tools aren't there yet. The frontier work, in companies and in academic research, is heading in that direction fast, and "workforce digital twin" is the phrase that has started attaching itself to it.

The CPO who treats all of this as a technology problem to delegate will lose. The CPO who treats it as a strategic capability to own, choosing the questions the data should be asked, deciding which predictions are worth acting on, defending the line on what counts as ethical use of employee information, ends up walking into board meetings with the kind of evidence that, until now, only the CFO has had. That changes the politics of the role considerably. It also changes who gets hired into it.

The reskilling problem

"Continuous reskilling" is one of those phrases that has been on HR conference slides for fifteen years, and it has rarely meant anything in practice. Most corporate learning programs are theatre. Employees know they're theatre. The compliance modules get clicked through on the bus, the Coursera licence goes unused, completion metrics look great in the engagement deck, and a year later the same skill gaps are sitting there.

The CPO of the next decade has to actually solve this, because the half-life of a job has gotten short and the cost of failing is no longer abstract. When the marketing team's role is meaningfully different every eighteen months, the old model (an annual training day, a couple of online learning licences, the leadership program for high potentials) isn't a serious response. It's a polite gesture.

The companies getting this right share a few features. They protect real learning time on the calendar. Not "encouraged" time, scheduled time, the kind that doesn't get sacrificed when a deadline slips. They build internal mobility programs so people can move sideways into adjacent roles without it being treated as a demotion or a failure. They have honest conversations about which roles are sunsetting and what the company will do for the people in them, instead of pretending the restructure that everyone can see coming isn't actually coming. They put serious budget into AI literacy for everyone, not just the engineering teams who already get it. They measure capability change rather than course completion.

This is hard work, and most of the resistance isn't ideological. It's structural. Managers don't want to lose their best people to internal moves, because their best people are the reason the team makes its targets. Finance teams can't see the ROI on a six-month upskilling program until the program is over and the results are baked in. Employees are exhausted from the last three initiatives and have low patience for a fourth. The CPO has to push through all of that, repeatedly, while the rest of the organisation is busy reorganising around the latest model release.

Get this wrong and you slowly hollow out your workforce while telling everyone you're investing in their growth. The good people leave first. They always do. They have options. Get it right and you end up with a workforce that adapts faster than your competitors', which over a five-year horizon is the kind of advantage that's almost impossible to copy.

Building leaders for jobs that don't exist yet

Almost every leadership development program designed in the last twenty years has assumed a relatively stable corporate hierarchy. You join as an analyst, you become a senior analyst, you make manager, you make senior manager, you make director. Each step came with a known set of competencies. Companies invested in those competencies because they had a reasonable bet that the leaders they were developing would still be running something recognisable in fifteen years.

That bet is getting harder to make. The director jobs of 2030 are likely to look meaningfully different from the director jobs of today, both in what they manage (mixed teams of humans and software, with workflows that change every couple of quarters) and in what they need to know (how to govern AI tools, how to make decisions with incomplete and partly-synthetic information, how to keep people steady during the next restructure). Some of these capabilities are barely teachable in the abstract. They have to be developed under conditions that resemble the real thing.

The CPO becomes responsible for a kind of leadership development that doesn't quite exist yet. Less classroom, more rotation through genuinely difficult assignments. Less competency framework, more deliberate exposure to ambiguity. More coaching from leaders who have actually navigated AI transitions inside their own functions, less generic executive education from people who haven't. Some of the more thoughtful succession work happening in companies right now isn't even framed as "leadership development." It's framed as identifying people who can hold their nerve when the playbook stops working, then putting them in situations that force the trait to develop.

This matters because the cost of getting it wrong isn't just leadership gaps. It's leadership gaps at exactly the moment a company is trying to navigate its hardest transition. Boards that have lived through acquisition integrations or restructuring waves know how this goes. The wrong leadership in a transformative period can take a decade off a company's trajectory. The CPO is the executive whose job it is to make sure the bench has the right people on it, in a market where the definition of "right" is shifting under everyone's feet.

The skill stack the job now demands

The job description for a future Chief People Officer is insanely broad, and that breadth is the actual problem. The role demands credible literacy across at least half a dozen domains that used to belong to other functions, and you can't fake any of them.

Data literacy is the most obvious. A modern CPO needs to read workforce analytics the way a CFO reads financial statements, and to recognise when the data is being asked the wrong question, which it usually is. AI literacy comes with that. Enough understanding of how the models work, where they fail, and what they actually cost to govern, that the CPO can push back credibly when the CTO says something convenient. Organisational design is essential because the company will be redesigned every couple of years for the foreseeable future, and someone needs to think clearly about the second-order effects rather than just the org chart. Behavioural psychology and applied neuroscience are no longer fringe interests. They're how you understand what's actually causing the burnout pattern, why the high performers are leaving, and what motivation looks like when nobody trusts the numbers anymore. Ethics is now a daily concern rather than a quarterly review topic, because algorithmic decisions about people will attract regulators, journalists, and lawsuits over the next few years. Strategic communication matters because change of this scale fails without it, and most companies underweight it badly.

The honest read on this is that it's a more demanding intellectual stack than most CFO roles. It is also, in many companies, paid significantly less, which says something about how seriously most boards have actually taken HR over the last twenty years. That's going to change, but probably not as fast as it should.

The hard part

There's an optimistic version of this story where AI gives HR a strategic seat at the table, the function gets the budget it always deserved, and everyone ends up doing more interesting work. Some of that will probably happen. The harder version is also worth telling, because most companies underestimate it.

Workforce anxiety is real and growing. People read the news. They know what AI is doing to roles like theirs. They don't trust the standard reassurances, and they shouldn't, because the reassurances are usually wrong. CPOs who pretend otherwise lose credibility very fast. CPOs who acknowledge the anxiety openly, name the uncertainty without trying to wallpaper over it, and show what the company is doing about it, build the kind of trust that becomes priceless when the next disruption hits.

Adaptation fatigue is just another way of describing of burnout, and it's becoming the more common diagnosis. People can absorb a lot of change. What they struggle with is constant change, especially the kind where the goalposts keep moving and the success criteria keep shifting underneath them. Researchers studying this have started using terms like cognitive overload and perpetual adaptation stress to describe what happens when an organisation runs three transformations on top of each other and expects everyone to stay sharp through all of them. The CPO who can spot this pattern early, and who is willing to push back on the next initiative when the workforce is already saturated, ends up protecting performance in a way that doesn't show up on a roadmap but matters more than most things that do.

AI surveillance is another problem and often a more serious one. The productivity tools that have been embedded in workplaces over the last few years track an extraordinary amount: keystrokes, screen time, meeting participation, message tone, focus minutes, calendar density. Most employees haven't fully clocked how much is being measured about them, but they will. When that conversation breaks open publicly, probably via a story about a specific badly-behaved tool at a specific named company, the firms that deployed these systems without thoughtful governance will face a backlash that makes the 2022 return-to-office wars look polite. The CPO is the executive who has to decide where the line is, and who has to be willing to defend that line against whoever in operations bought the tool that crosses it.

Algorithmic management is the harder version of the same problem. Software is now making real decisions about who gets hired, who gets promoted, who gets the bonus, and at some companies who gets let go. The systems are often opaque. The humans who run them frequently can't explain why a particular candidate was filtered out or why a particular employee was flagged. Regulators haven't caught up yet, but they will, and the companies that deployed these systems aggressively without governance will end up explaining themselves to people who have subpoena power. This responsibility has moved from "interesting future topic" to "current liability," and it's on the CPO's desk because no other executive has the authority and context to handle it well.

Generational fragmentation runs underneath all of this and rarely gets discussed cleanly. The expectations a 25-year-old has of a workplace are not the same as those of a 50-year-old, and the gap has been widening since at least the pandemic. Younger employees expect AI-fluent leadership and aren't impressed by training programs that treat AI as exotic. Older employees, often the ones with the deepest institutional knowledge, are reasonably wary about how fast their roles are being rewritten. Holding both groups in the same culture, and getting them to keep teaching each other, is not a soft skill. It's one of the hardest organisational design problems on the modern CPO's desk.

And underneath all of it is culture, which is probably the most difficult part because it's hard to measure and easy to lose. Companies that go through three or four big transformations in a decade can come out of it stronger and more capable. They can also come out hollowed out, with the high performers gone, the institutional knowledge walked out the door, and a brand that no longer means anything internally. Holding culture together through this kind of change is genuinely difficult work. It doesn't show up cleanly on dashboards. It often only becomes visible after it's already broken.

Why the role gets more important, not less

There's a pattern in big technology shifts that's worth keeping in mind. The early winners are usually the companies with the best technology. The medium-term winners are the companies that integrate the technology fastest. The long-term winners, almost without exception, are the companies that figured out how humans work alongside the technology better than anyone else. That pattern played out with electrification, with the PC, and with the internet, and there's no obvious reason it won't play out again with AI.

Every company is going to end up with similar models, similar tools, similar automation, eventually. What companies won't share is whether they can actually retrain people quickly, hold their culture together through change, develop leaders who can navigate ambiguity without burning out, and keep their best people from drifting toward the exit. Those things take years to build, weeks to lose, and they sit, almost entirely, on the desk of the Chief People Officer.

There's a way of describing all of this that's slightly clearer than any of the existing job titles. The future People function is the organisational resilience function. The thing that lets a company keep performing through change rather than just surviving it. Hiring, pay, compliance, engagement, all of those still happen, but they end up sitting underneath something larger. The CPO is the executive responsible for whether the organisation can absorb what's coming next without breaking, and whether the people inside it will still be capable, motivated, and willing to do the work on the other side of it.

The CPO of 2030 is not going to be a better-resourced version of the HR director of 2015. The role is doing something different now. Some combination of organisational designer, ethicist, behavioural economist, leadership advisor, and resilience strategist. The title may change at the executive level too. Chief Workforce Officer, Chief Human-AI Officer, Chief Adaptation Officer. Take your pick. The label probably matters less than the function, and the function is rapidly becoming non-negotiable for any serious company.

The companies that figure this out early will have an unfair advantage in the second half of the decade. Not because they automated faster than their competitors, but because their people held together while everyone else was breaking things and apologising for it later. Which has, when you think about it, more or less always been the case. It's just that nobody quite expected the function delivering it to be HR.

If you have an opinion on the future of HR, I'd love to hear it. The above article was built on some of the most recent research exploring the topic, but anecdotal evidence is very useful in understanding the realities of the individual HR leaders out there.