For two hundred years we made people fit the system. That era is quietly ending.

There is an assumption so old and so deep that we stopped noticing it was an assumption. It is this: the human being adapts to the system. The production line, the org chart, the standard shift, the one-size job description, the process that everyone bends themselves around. It built the modern economy, and it was, if we are honest, quietly brutal, because it treated people as more or less interchangeable and asked each of them to file down whatever parts did not fit the slot.

We did not choose that out of cruelty. We chose it out of arithmetic. Fitting the system to each individual person was a luxury you could afford for exactly one human, the chief executive, who got the assistant and the tailored everything. Everyone else got standard issue, because bespoke did not divide by a thousand employees. So we made people fit the system, and we called the leftover friction "just how work is."

I think that era is ending, and I do not mean that as a slogan. I mean the arithmetic has actually changed.

The layer we have been ignoring

We have spent a decade getting good at what I would call the Business OS: the strategy, the goals, the metrics, the process, and in the work we do, the memory and the learning loops. All of it useful. But a Business OS is an abstraction. It does not run on servers. It runs on people, and every person is already running their own operating system.

Call it the Human OS. It is made of biology, attention, energy, neurology, personal history, circumstance and aspiration, and it is utterly unique to each person. It is also, mostly, invisible to the business layer sitting on top of it. We manage the Business OS with great care and treat the Human OS it depends on as a black box we are faintly surprised by when it burns out.

Here is the shift, and it is the whole essay. AI is the first technology that makes it economically possible to reverse the two-hundred-year assumption. To adapt the system to the human, one human at a time, at scale. The tailored support, the individualised development, the ways of working that fit the actual person, all of it was once affordable for one and is becoming affordable for everyone. That is not a productivity upgrade. It is a change in what an organisation can be.

It is an ecosystem, not a machine

The temptation is to picture this as a stack: humans at the bottom, business on top, people serving the company. That is just the old machine again, with nicer words. The truer shape is an ecosystem, and the Business OS is not the top of it. The Business OS is where the ecosystem becomes visible.

Think of it as one instrument panel with three readings. Are we succeeding for our customers? Are our people thriving? Are we in balance? Customer experience, financial health and human thriving are not three competing agendas. They are three dials on one system, and the entire craft is balance.

Which leads to the least glamorous and most important idea here. If human thriving has no dial on the panel, it is not real to the business. It becomes a value on a wall, invisible the moment a trade-off is made, losing every argument to the numbers that are actually measured. You do not protect what you do not measure. So the first move is almost embarrassingly practical: give human thriving a place on the instrument panel, next to revenue and retention, so it can no longer be quietly sacrificed.

And here is the part a hard-nosed finance mind should like

None of this asks anyone to be softer at the expense of results. Quite the opposite. Human-thriving measures are leading indicators of the outcomes the business already cares about. Thriving people build better products and serve customers better. Depleted people ship worse work more slowly, and then they leave, taking what they knew with them. So thriving sits upstream of customer and shareholder results, on the same metric tree, not in a separate wellbeing silo that the real business ignores.

That reframes the humane thing and the commercial thing as the same thing. You are not choosing between looking after people and winning. In a world where advantage comes from how well your system fits the humans running it, looking after people well is how you win. The ethics and the economics point the same way, which is a rare and comfortable place to stand.

There is a nice symmetry in it too. A healthy body keeps itself alive through balance, holding many variables in a liveable range rather than maximising any one of them. A healthy company does exactly the same thing one level up. Push a single number to the exclusion of the rest and the system degrades, usually on a delay long enough that nobody connects the cause to the effect. The company that lasts is the one in balance, and now, for the first time, balance is something we can actually see.

Why now, and not five years ago

I could have written most of this a while ago and it would have been a nice idea with no way to act on it. What has changed is not the values. It is the arithmetic, and it has changed in two directions at once.

On one side, the old source of advantage is fading. When everyone can reach the same capable AI, having more people doing average work is not an edge, it is an overhead. Advantage now comes from the things the average cannot give you: real creativity, genuine innovation, and the speed of your learning loops, all of which run on human capacity that a cortisol-soaked, burnt-out organisation simply does not have.

On the other side, the thing that was always unaffordable, fitting the system to the individual, has become affordable, because the AI thread can carry that individualisation at scale. So the reason to do it has grown at the same moment the cost of doing it has collapsed. That is an unusual alignment, and it does not last forever. The companies that notice it early will pull away quietly, then not quietly.

Where I have to be honest

I am not going to pretend the same capability has no shadow. The exact thread that could personalise support could equally power surveillance and quiet manipulation. The line between a system that understands you well enough to help and one that understands you well enough to exploit is drawn by consent, governance and intent, not by the technology. When we build the Human OS layer, it has to be something the individual benefits from and has a say in, not something done to them. I would rather say that plainly at the start than have it be the thing we all pretend not to see.

The invitation, not the pitch

So the question I would leave with a leadership team is not "how do we use more AI." It is larger and more interesting. What would it mean to run a company that finally fits the people in it, that puts human thriving on the same panel as revenue, that treats the mix of what its people need as a real input rather than a black box, and that uses AI to make all of that visible and affordable for the first time?

That is a genuinely different kind of company, and it is buildable now. Not by being softer. By being better designed.

Adapt the system to the human, not the human to the system. It is the humane thing, the biological thing and the winning thing, and for once those are all the same thing.