Uh-oh. Are you lower-value human capital?

Let me tell you about the worst thing to happen to the English language since someone decided "synergy" was a real word.

A banker named Bill Winters — chief executive of Standard Chartered, compensated so magnificently that his own shareholders once staged something close to a revolt about it — recently looked across his organization and apparently concluded that quite a lot of the people in it were, and I am quoting directly here, lower-value human capital.

Not staff. Not colleagues. Not even "personnel," which was already bad enough. Lower-value human capital.

I've heard some spectacular corporate gibberish over the years. I once sat through a forty-minute presentation in which a man in a very nice suit explained, with genuine passion, that his company was going to "leverage its core competencies to deliver transformative stakeholder outcomes." Nobody left. Nobody laughed. We all just nodded and waited for the snacks. But "lower-value human capital" is something else entirely. It is the language of a man who has spent so long surrounded by spreadsheets that he has forgotten, in some fundamental neurological sense, that the numbers on them used to be people.

He tried to walk it back on LinkedIn. Twice. This tells you everything. A man who posts on LinkedIn twice to explain himself has not had a sudden attack of self-awareness. He has had a sudden attack of communications advice.

Mortgage industry

AI displacement risk by role

Scores are indicative estimates based on task routinisability and current AI capabilities, not empirically measured data.

Now. The reason this is personally uncomfortable — and here I should probably confess that I am not writing this purely as a detached observer — is that a second executive has come along to provide the intellectual scaffolding for what Winters accidentally blurted out. Matthew Prince, the chief executive of Cloudflare, laid off more than 20% of his workforce while his company was growing at 30% and generating record revenues, and then wrote about it approvingly in the Wall Street Journal. He has a framework, borrowed from Peter Drucker's 1954 book The Practice of Management, which divides all workers into three categories: Builders, Sellers, and Measurers.

Builders create things. Sellers sell things. Measurers do everything else — finance, compliance, legal, operations, internal audit, middle management. And the Measurers, Prince explains cheerfully, are being replaced by AI, because AI can measure an organization with a level of precision and tirelessness that no human can match. He is not apologetic about this. He seems, if anything, rather pleased.

I sat with this framework for a while. Then I applied it to myself.

I am not a Builder. I have attempted, over the years, to construct a raised garden bed, a set of flat-pack shelves, and what was supposed to be a garden gate. All three projects ended in a way that can best be described as "structural learning experiences." I have also, on three occasions, attempted to hang a picture, and on each occasion the wall has come off worse.

I am not obviously a Seller either. I don't really persuade people to part with money in any conventional sense. What I do — and I recognize this sounds faintly absurd when laid out plainly — is sit down, think about things quite hard, and then write articles to share.

Is that a Builder? Is it a Seller? Or is it, in the cold taxonomy of the new corporate religion, a Measurer? Am I, at fifty-something, a woman whose function can be replicated by a sufficiently well-prompted language model?

The question is not entirely rhetorical. These AI systems are, by any reasonable assessment, getting very good at this sort of thing. They read widely. They write fluently. They do not have off days, they do not require a functioning furnace at home to be able to concentrate, and they have never once gone to pieces because someone left passive-aggressive feedback in the comments.

Prince's consolation is that Builders and Sellers are safe. Humans, he argues, still want to deal with other humans — people who build trust, understand context, and show up personally when it all goes wrong. That is a reasonable point, although it does imply that the future of human employment is essentially limited to plumbers and relationship managers, which is not quite the renaissance the economists promised us when they said we'd all be freed from drudgery to do more creative work.

What I find genuinely troubling — and I don't use that word loosely, because I am a woman who finds very little genuinely troubling — is the gap between what these executives are saying and what they appear to mean. "Every opportunity to reposition," says Winters. "AI-native builders and sellers," says Prince. What they are describing, if you strip away the optimism, is a future in which the large and unglamorous middle of the workforce — the people who have spent twenty years being reliably competent at things that turn out to be measurable — will need to reinvent themselves entirely, on their own time and at their own expense, into something a machine cannot yet do.

That is quite a significant ask of someone who is fifty-three, has a mortgage, and was genuinely excellent at reconciling the quarterly accounts.

I did, in a fit of either journalism or anxiety — I am no longer sure which — ask Claude the question. And Claude didn't exactly put my mind at rest — rather than telling me there is nothing to worry about, it asked me to self-reflect.

The harder, more honest questions to ask yourself:

• Could a well-prompted AI produce 80% of what you produce in a day, at acceptable quality?
• Is your value mostly in processing and routing information, or in judgment, relationships, and accountability?
• Does your organization know specifically what would break if you left tomorrow — or would they mostly just redistribute your calendar?
• Are you the person clients or colleagues ask for by name, or are you interchangeable with your job title?
• Do you make decisions, or do you prepare the materials that help someone else make decisions?

Claude did, at least, give me some career advice on how to make myself — or at least appear — more indispensable. And maybe this could also help you, if, like me, you are at least mildly concerned about what kind of threat AI poses to your job security.

What makes someone harder to replace, based on current evidence:

• Deep contextual knowledge of a specific organization, client, or domain that isn't written down anywhere
• The ability to handle genuinely novel situations with no established playbook
• Trust — specifically, relationships where the human connection is the product
• Creativity that produces things people didn't know they wanted
• Physical presence (care work, skilled trades, anything requiring hands)

At least Claude gave me more career advice than the man who started this whole conversation.

Which is, when you think about it, exactly the problem his employees face.