The CFO Had a Ledger. The CHRO Had a Photograph.

Part I

The Question

The board meeting had gone well for ninety minutes, which is roughly how long it takes for a board meeting to go badly.

The CFO had presented first. Eleven slides. Every number was traceable to a ledger, every ledger was reconciled, and every reconciliation was signed. When a director challenged the depreciation schedule, the CFO did not defend it. He simply opened a system and showed where the number came from. The director nodded and moved on. Nobody thanked him, and nobody needed to. That is what being governed looks like: nobody notices.

Then it was the CHRO's turn.

She was good. Genuinely good, with fifteen years behind her, three transformations delivered, and a team that would follow her into a fire. Her deck was beautiful. Engagement was up four points. Attrition was down. There was a slide about culture with a photograph of the offsite on it.

The chair let her finish. Then he asked the question.

"How mature is our HR function?"

She said, "We're on a journey."

There was a pause of the kind that gets remembered.

"On a journey," the chair repeated, not unkindly, which made it worse. "Where are you on it?"

"We've made significant progress in..."

"I'm not asking whether you've made progress. I'm asking where you are. He can tell me where he is to two decimal places," and here he nodded toward the CFO. "People are our largest cost, larger than everything on his slides. You are telling me that the largest number in this company is managed by a function that cannot say how well it is managed."

She said the only honest thing available to her. "I can get you a number."

"Get me a number I can defend," said the chair. "You have two quarters."

Part II

The Inventory

She went back and did what every competent executive does first. She asked her own team for the facts.

This is where it fell apart, and it fell apart faster than she expected.

Here is the part she did not say out loud, because saying it out loud would have been unbearable. None of this was incompetence. Every one of these people was excellent. They had simply been asked, for their entire careers, to run the largest cost line in the company on memory. They had done it heroically, and heroism does not survive an audit.

Part III

The Heresy

She hired a consultant, because that is what you do.

The consultant was expensive and said the word maturity nineteen times in the first hour. He proposed a transformation programme: eighteen months, three workstreams, a change narrative, a journey.

"Can it fail?" she asked.

He blinked. "I'm sorry?"

"The journey. Can it fail? Is there a point at which we look at it and say that it did not work?"

He explained that transformation was iterative, that maturity was a direction of travel, and that the important thing was momentum and buy-in.

"So no," she said. "It cannot fail. That is why we like it."

She did not hire him.

Instead she went down two floors and asked the CFO how Finance had become governed.

He looked genuinely puzzled by the question. "We didn't adopt a mindset," he said. "We implemented a ledger. Everything else is a consequence of the ledger."

"That's it?"

"That's it. You build the components in order and at the end you have a system. If you build them out of order, you have a very expensive filing cabinet."

She asked what HR's ledger was.

"I have absolutely no idea," said the CFO. "That is your job. But if it were mine, I would start by finding out whether you all use the same words. We spent six months on the chart of accounts before we automated anything. Everybody hated it, and it was the only reason any of the rest worked."

Part IV

The Build

So she built. Six things, in order, and the order was the whole point.

One taxonomy, in three levels: Groups, Families, Processes. It was not borrowed wholesale from a framework and imposed, and it was not invented from scratch either. She took a reference skeleton and ran a brutal scoping exercise, cutting the families that did not apply and adding the processes unique to the company. It took eleven weeks. It was the most boring work anyone had done in years. Two people told her privately that it was beneath her.

Then she published one number: the share of HR's actual footprint that was named, owned, and inside the taxonomy. She watched what happened to the conversation when the claim that HR was governed became a percentage instead of an assertion.

Later she would say this. Every failure she saw at other companies came from skipping this step. Skip it and every later step fragments. Assessments measure things nobody can compare. Maps document processes nobody can find. Skip nothing else before this one.

Four groups scored the same processes on the same scales: the HR lead, the HR team, the CEO and executive sponsors, and the employees HR served. They scored importance, capacity, satisfaction, and time.

Two rules made it work. The first was that everyone scored the same processes on the same scales, because anything else is a spreadsheet full of opinions. The second was that no individual answer was ever exposed, only aggregates. That is not a courtesy. It is the entire reason the CEO and the employees answered honestly.

The results were worse than she feared and better than she hoped. HR's self-assessment and the CEO's assessment of HR diverged sharply on four families. That gap had existed for years, and it had silently governed every budget conversation she had ever lost. Nobody had ever measured it, so nobody had ever been able to argue about it with evidence, only with feelings, and the person with the most authority always wins an argument made of feelings.

She stopped losing those conversations. Not because she got better at them, but because she stopped having them.

Every process in the taxonomy received a swimlane map, a named owner, and a review date.

The maps were not the point. Two other things were. The RACI exercise, run honestly, produced a list of processes with no accountable owner, and that list rather than the neat grid was the real deliverable. Version control with approvals and scheduled reviews was the second, because a map without a review cycle decays into the exact tribal knowledge it replaced, only now wearing a suit.

Her experts hated authoring from a blank page and were merciless when correcting a rough draft. So she stopped asking them to author. Drafts were cheap and judgement was not, so she spent their time on judgement.

There were six levels, so that documents stopped competing with each other: Policies, Standards, Procedures, Guidelines, Work Instructions, and Forms. Every one was versioned, every approval was recorded, and every trail was intact. The question of who approved a document and when became a lookup rather than an excavation.

The instrument that mattered was coverage. Because every document anchored to the taxonomy, she could see which processes were documented and which were naked. The holes were where the risk was highest, as they always are. Most functions find that out during a dispute. She found it out during a review, which is the same discovery at two orders of magnitude less cost.

The strategy had a vision, the business priorities it served, and three to five strategic dimensions, driven down into objectives, pillars, initiatives, and key results on a single page. If it does not fit on a page, it will not survive the operating year.

Then came the part that actually mattered, and the part almost nobody does. Each quarter's version was fixed, immutable, and reviewed on the record. When the board asked what had changed since last quarter, the difference was there in writing.

Taxonomy adoption, assessment results, map coverage, policy coverage, and strategy governance all fed a single composite maturity index on a four-level model, with a quadrant view, a trend line, and weights she could explain in one sentence.

She set the weights before she saw the score. That is a small detail, and it is the difference between a measurement and a piece of marketing.

Part V

The Second Question

Two quarters later, the chair asked the same question.

"How mature is our HR function?"

She gave him a number. Then she gave him the four things it was built from, the two families dragging it down, the prioritised list of what she would fix first, and the trend line showing the direction.

A director challenged the weighting on one pillar. She did not defend it. She opened the model, showed him where the number came from, and showed him what the score became if it were weighted his way instead. It was 0.3 lower, with the same two families dragging.

He nodded and moved on.

Nobody thanked her, and nobody needed to.

Part VI

The Part She Did Not Expect

The last thing happened by accident.

Someone in her team pointed an AI assistant at the governed data and asked, in plain language, which process families had no accountable owner. Out of curiosity, she asked a peer at another company, one that was bigger, better funded, and ungoverned, to run the same question in their environment.

That is when she understood what she had actually built. Governance is not the alternative to AI in HR. It is the precondition. Point an assistant at a governed function and you compound the advantage. Point it at a folder of stale decks and you automate the guessing, faster, at scale, in a font that looks like authority.

She kept the access read-only, scoped to the permissions of whoever was asking, with survey data exposed as aggregates only. Those guarantees were possible only because the structure underneath existed. There is no way to bolt them onto a shared drive.

Part VII

The Moral, Stated Plainly

Because fables should not be coy.

Everyone told her there was no time for this. There never is. That is the symptom rather than the objection. The reason there is no time is that the function is ungoverned, and ungoverned time disappears into re-explaining processes, hunting policy versions, and re-litigating debates that a dataset would have settled in an afternoon.

She did not run a transformation. Each step paid for itself alone: the assessment justified its own cost, and so did the coverage view. She started where the pain was loudest and built outward, and the taxonomy made sure the pieces joined up later.

There are six components, and the order is not a preference. It is the mechanism.