Before CHROs ask the enterprise to lead humans and AI together, HR must become the demonstration project.
It is nearly 30 years since Dave Ulrich’s Human Resource Champions handed our profession its modern ambition: be a strategic partner, not an administrative function. And for nearly 30 years, we have made the same request of every executive committee we have sat on. Free us from the administration, and we will be strategic. Give us better data, and we will be true business partners. Give us time, and we will develop people instead of processing them.
Quietly, over the past three years, that request began to be granted by technology.
So I want to pose an uncomfortable question to you – one I first had to pose to myself, as a group HR executive: the freedom we asked for is arriving. What exactly are we doing with it?
I have occupied the CHRO seat twice, at a multinational pharmaceutical company and in a JSE-listed group, and in both seats I lived the old constraint honestly. The HR functions I inherited were not unstrategic because their people lacked ambition. They were unstrategic because the operating model consumed our capacity: the reporting cycles, the legacy systems, the manual analysis – or the lack of any analysis at all – the training needs assessments outdated on arrival, the engagement survey whose results reached line managers months after the sentiment had moved on. We were the profession of people, spending most of our hours on process. Some called us process ninjas. It was not entirely a compliment.
AI has begun to remove that constraint not merely by accelerating individual tasks, but by changing what a small, capable HR team can build and deliver. I know, because I have stopped talking about it and started building.
And this is where the leadership question begins, because leading humans and AI together starts inside our own function. HR cannot ask the rest of the organisation to redesign work, interrogate machine-generated insight and preserve human accountability while our own operating model remains unchanged. The division of labour is clearer than the anxiety suggests: the technology’s job is to assemble, screen, compress and signal; the human’s job is to define the criteria, interrogate the output, coach, decide and carry the accountability. Before this becomes an enterprise-wide leadership capability, it is a credibility test for the CHRO. Let me tell you what that test has looked like in my own work.
The machine reads the paper. My people read the person
The moment this became real for me was not a technology demonstration. It was watching my graduate programme change shape. For two years running, I have shifted my teams away from screening and towards scouting a graduate programme whose actual job is to identify future leaders, not to process applications. And we did it without waiting for the long-awaited applicant tracking system.
Consider what that problem used to look like. Thousands of CVs arriving in a mailbox. No applicant tracking system, and no budget to buy one. Candidates whose first human contact with us was a brief interview or an assessment centre, and a stranger evaluating a stranger. My team’s capacity was consumed by the volume, and the volume added no insight.
Today, the screening happens in seconds. I can see who our top graduates are across multiple data points at once – not one CV at a time, not one recruiter’s tired judgement at 5pm on a Friday. This is not the outsourcing of selection to an algorithm: people still define the criteria, review the evidence and make the decision. The technology compresses the administrative work; it does not inherit the accountability. The part nobody warned me about is that automation did not make our recruitment less human. It made it more human. The hours the machine returned to us are spent on campus, getting to know young people as people, long before any interview. The machine reads the paper; my people read the person. That is the trade our profession has been begging for since Ulrich first drew the model.
The dashboard I fought a decade for then built myself
The second story carries scar tissue. Years ago, I set out to build a simple thing: a view that showed me, and each of my HR business partners, where the organisation’s hot spots were. I printed reams of paper. I built spreadsheets upon spreadsheets. I argued with the CIO – not because they were wrong, but because an HR dashboard was never going to beat revenue systems onto the IT priority list. A business analyst eventually helped us deliver a tool, but its capability remained limited. That was the reality of the data-driven CHRO in the old world: your ambition queued behind everyone else’s.
I have now built that dashboard myself – a prototype developed which can be used with approved tools and appropriately governed data. It did not need a project code, a steering committee or a position in the IT queue to demonstrate what is possible; scaling it responsibly still requires technology, data, security and risk governance, as it should. It shows the state of the organisation cut by the HR business partner so the right conversation happens with the right partner about the right problem, while it is still a conversation and not yet a crisis. What changed is not the questions. It is the speed and confidence with which we answer them.
And most recently, I have built what I call an enterprise navigator – a strategic workforce planning tool. The first-pass analysis that once consumed weeks now begins at the click of a button: where the inefficiencies sit, and what they mean for the bottom line. The workforce judgement cannot and should not be automated away. But the conversation moves beyond headcount, the crude lever we reach for under pressure, to the work itself: which processes should be retired, which matured, which augmented with AI. Strategic workforce planning has been on every HR conference agenda for twenty years; few of us could ever do it at depth, because the analysis was unaffordable. Now the analysis is affordable and the judgement is finally what we get paid for.
None of these began with a transformation programme, a consultancy engagement or a major new budget line. They began with a business question, governed experimentation and the willingness to build. That is the point.
Four freedoms we now hold
The freedom to actually partner: The HRBP role has always carried a quiet fiction. We called it partnering, but much of it was reporting. When the analysis assembles itself, the HRBP’s hours shift from producing the numbers to standing in the business conversation with a point of view. The question in the room changes from what the data says to what we should do, and only one of those requires a human in the chair.
The freedom to give feedback that arrives in time: We kept the annual feedback ritual for decades because continuous, evidence-based feedback was operationally impossible at scale. It no longer is. With an instant insights report and real signals from real work, feedback becomes what it should always have been: a frequent, ordinary, useful conversation. The technology surfaces the evidence; the manager still owns the conversation.
The freedom to develop each person, not the average person: Our learning portfolios were built for a fictional median employee: generic programmes, sheep-dip compliance, catalogues nobody finishes. AI-generated learning pathways sequence development around an individual’s actual gaps, actual trajectory and actual pace. Targeted training was always the right answer; it was simply unaffordable. Now the generic version is a choice, not a constraint.
The freedom to measure what we always claimed mattered: For my whole career, HR has argued that culture, capability and leadership quality drive performance, while measuring almost none of it. We can now gather richer and more frequent signals about all of them than was previously possible. One caveat, and it is a design decision, not a footnote: the power to gather signals about everything is not a mandate to surveil everyone. The line between insight and surveillance is ours to draw deliberately, out loud, before someone else draws it by default.
The end of the excuse economy
Here is the provocation, and I direct it at us, not at our critics. If our function looks the same in two years – same rituals, same cycle times, same generic programmes, same lagging reports, then the constraint was never the tools. Every explanation our profession has offered for underdelivery has rested on capacity: too much administration, not enough data, no time, no budget. AI has weakened those excuses dramatically. What it increasingly exposes is the remaining variable: our own ambition, judgement and skill.
That should excite us more than it frightens us. Professions are rarely handed a moment where the gap between promise and capability closes this suddenly. But the standard has moved. A CHRO presenting last quarter’s headcount deck to an exco that knows analysis is now instant is not presenting data. They are presenting evidence of a function that has not noticed its constraint is gone.
We cannot credibly build organisations where leaders work fluently with AI if HR itself has not made that shift first. Our function is the demonstration project. The way we automate our analysis, redesign our feedback, personalise our development and govern our measurement is the organisation’s first lesson in leading humans and AI together. Leaders will treat AI the way they watch us treat it.
What I would put on the CHRO agenda this year
Not a transformation programme. Four honest questions, asked of your own function first. Which of our rituals exist because they were once operationally necessary and are now merely familiar? Where does automated analysis free capacity, and have we redirected it into partnering, or quietly refilled it with process? Which parts of our learning portfolio would survive if every employee could see the personalised pathway they could be on? And what have we chosen to measure, what have we chosen not to, and can we defend both choices to our own employees out loud?
Global research is blunt about the stakes. The World Economic Forum expects 39 percent of workers’ core skills to change by 2030. And McKinsey’s global AI research found that, of the organisational attributes it tested, the redesign of workflows had the largest effect on whether organisations saw bottom-line impact from generative AI. I can think of no better place to prove that redesign is possible than the function that owns work itself.
For 30 years we asked to be freed up. It is happening, faster than most of us have noticed. The era of explaining what HR would do if only it could is ending. I no longer describe what my function would deliver if it had the tools. I build the tools, and then I deliver – visibly, imperfectly, first. The machines now do what machines must do. Which means, at last, we are free to do what only humans can and to be the HR we have always wanted to be.
