The Work That Makes Us Human

Today I had an argument with Charlotte, and it reminded me how it feels to be alive.

I should explain.

Charlotte works for a leading restaurant company. Or at least she answers the telephones for them.

I was calling about a reservation I have made for my wife’s birthday. We have arranged a private room for about twenty friends, many of whom have not seen each other for a long time. It should be a great occasion and, as these things tend to do when you care about them, I had started thinking about some fairly small details.

What I wanted to know was how wide the long table was, and whether we could seat more than one person at each end.

That was it.

I wasn’t trying to change the booking or reorganise the evening. I really just wanted somebody at the restaurant to walk into the private room, have a look around and tell me what they could see.

My heart sank when Charlotte answered because I knew almost immediately that she was not going to make this happen.

I started saying, probably a little quickly, “Can you just put someone else on? Someone who can actually help me?”

Charlotte persisted, perfectly politely, asking me to be clearer about what I wanted. I thought I was being clear. I needed somebody to walk through a door, look with their own eyes and have a short conversation with me.

Charlotte couldn’t do that, because Charlotte is an artificial intelligence.

Eventually, through some combination of persistence and whatever escalation logic sits behind her software, I reached an actual person. The problem was solved very quickly, and I found myself saying something slightly grumpier than I intended about what a strange idea it was for a good restaurant to put a computer between a customer and the person who could actually help them.

I’ve since been thinking about the unnecessary complication Charlotte caused. She wasn’t missing any information. What she lacked was “presence of mind” because she was not present at all.

A person walking into that room would simply have looked at the table and made a judgement. They might have noticed the shape of it, whether there was an awkward corner, or that one arrangement would work better than another. They might also have asked why I cared about the table width.

And if I’d said that twenty friends were getting together, some of whom hadn’t seen each other for years, and we didn’t want to strand anyone alone on a table end, they might have suggested doing something differently.

That one unprompted question could improve the whole evening.

It isn’t really customer service at that point. It is the sale, the loyalty and the review hiding inside a moment of human curiosity.

Living people can turn a door handle and enter a room for themselves. We can hear irritation in somebody’s voice. We can notice that the question about chairs is really about whether Rita, Sue and Bob are going to sit beside each other or not.

And I increasingly wonder whether we are drawing the boundary between those things and technology in the wrong places.

The robot with the toolbox

I spend a considerable part of my working life thinking about software and building artificial intelligence, so I am hardly approaching this as a technological sceptic.

I use frontier AI tools constantly. They have changed the way my company works, changed how quickly I can work through a problem, and made possible things that even a few years ago would have required considerably more people and time.

For better or worse there is little debate over whether AI is extraordinary or not.

What I don’t believe is that the progress of these models leads in a straight line to a world in which huge parts of useful human work simply disappear because intelligence itself has become cheap.

The physical world is much less cooperative than that.

If and when a robot arrives at my house carrying a toolbox, works out why there is water coming through the ceiling, finds the pipe somebody boxed in ten years ago, understands whatever inventive solution the previous plumber came up with, avoids destroying the wooden floor, finds the right replacement part and fixes the problem safely, I will revise my view.

For now, I would still quite like a plumber.

The difficulty is not simply making a machine intelligent enough to understand what a boiler is. It is putting that machine into an unfamiliar physical environment where the plans may be wrong, the last person changed something nobody recorded, the customer is standing behind it asking questions, and there are real consequences if the answer is wrong.

That becomes especially important in what we might think of as a regulated experience.

Consider almost any school, hotel, hospital, care home, factory, rented flat or public building. Underneath the experience of using it sits an enormous physical infrastructure of pipes, wires, ventilation, drainage, heating, refrigeration, fire protection, lifts and increasingly batteries, solar generation and EV charging.

Much of the work on that infrastructure doesn’t simply need to be done. It needs to be tested, certified and signed off. Somebody carries financial liabilities for the result.

Somebody’s name goes on the piece of paper.

That mixture of an unpredictable physical environment, judgement, liability and the cost of being wrong is quite different from asking a model to produce the first draft of an email.

And yet a remarkable amount of our conversation about the future of work focuses on the email.

The bit of the economy I know

Perhaps I notice this because of the industry I work alongside.

Electric vehicle charging is my useful example.

From the outside, the transition looks overwhelmingly technological. Cars become computers, charging becomes software driven and the whole energy system becomes increasingly intelligent. Algorithms can work out when the car should charge according to electricity prices, grid conditions, renewable generation and what the household needs.

All of that is true.

But eventually somebody still has to drill a hole in a wall.

Someone has to visit the house, understand the electrical supply, work out where the charger can safely go, route the cable and deal with whatever strange decisions were made when the place was rewired in 1987. They have to talk to the customer, complete the installation, test it, make it compliant and occasionally come back because something doesn’t work quite as anybody expected.

The digital system can become extraordinarily sophisticated whilst the last metre into the real world remains intensely human.

I know some very good people in these industries, and a depressing number of the really skilled ones are either approaching retirement or quietly calculating how many more years they want to keep doing it.

If they own their business, there is usually another question sitting behind that calculation.

What happens to the company?

These are often businesses that grew in ways that would look almost comically unfashionable to the startup world. There was no accelerator, no pitch deck and certainly no product-led growth strategy.

Customers arrived because somebody told somebody else they were good.

The founder’s mobile number became the sales department. Their judgement became the quality-control system. Their memory became the CRM.

Quoting might involve instinct, a spreadsheet somebody built six years ago and someone called Dave to sign it off. A member of staff may be keeping a crucial customer happy almost entirely through a relationship nobody else in the company quite understands.

3 From the outside this can look unsophisticated.

Sometimes it is.

But quite often underneath all of that is a fundamentally good business. Customers trust it. People have careers inside it. It makes money. It has spent decades learning how to do something useful.

Then the owner turns sixty-two.

Suddenly an apparently ordinary business has a very large problem.

The businesses hiding in plain sight

We devote enormous amounts of attention to new companies.

I understand why. I have been one of those founders.

Startups contain possibilities in concentrated form. They have compelling stories, recognisable heroes, financing rounds and the seductive idea that something entirely new is being created.

But they are not most of the economy.

Most of the economy is made up of businesses that already exist.

The engineering company on an industrial estate. The commercial laundry servicing hotels. The HVAC contractor. The testing company. The electrical contractor. The specialist maintenance company with twelve vans that everybody in the area seems to know.

Individually they can look unremarkable. Collectively they make everyday life possible.

Many were built over twenty, thirty or forty years, and many of the people who built them are now getting older.

There may be no son or daughter who wants to take over. The management team may be perfectly capable but have nowhere near enough money to buy the company. The business may be too small to attract the sort of institutional buyer that appears in the financial pages and too dependent on local knowledge to make sense to a passive investor.

The founder may also have perfectly understandable feelings about what happens next.

If you have spent thirty years building a name in a town, employing people whose families you know and serving customers who have your mobile number, “maximise the exit multiple” may not be the only thing you care about.

You may want your money, of course.

But you might also want the business to continue.

You might care about the people.

You might prefer not to watch the name disappear six months after the deal.

So sometimes the founder keeps going for longer than they really want to. Investment slows. Decisions get postponed. The company becomes increasingly dependent on somebody who is increasingly trying to leave it.

Eventually, something that took a lifetime to build can simply disappear.

A lifetime of relationships, competence and local trust gets dissolved at Companies House for a fee.

That seems to me an extraordinary waste. Not simply financially, but socially too.

For a while I thought of this mainly as a succession problem.

Then I started looking at what was happening at the other end of working life.

The other end of the ladder

There is another argument happening around AI.

Whilst we cannot yet announce that artificial intelligence has destroyed graduate employment, the signs are there. Entry-level hiring is down. Senior managers are increasingly able to automate work they might previously have given to somebody junior.

So what happens if AI becomes exceptionally good at the work we traditionally gave to inexperienced people?

The junior analyst built the spreadsheet. The graduate did the first piece of research. The assistant produced the first draft that somebody more senior then rewrote. A new manager struggled with a rota and learned why what looked sensible on paper didn’t work when six actual human beings arrived on Monday morning.

A great deal of this work was economically inefficient. But it was also developmentally essential.

Nobody arrives in a business aged twenty-three with the judgement they will have at forty-three. They acquire it. They watch other people. They make fairly harmless mistakes before being trusted with expensive ones. They perform tasks that an experienced person could do faster because doing the task is part of how they become experienced.

There is tacit knowledge inside all of this. The ten things nobody writes down because everybody who has been around for long enough already knows them.

If AI automates enough of the bottom of the ladder, we cannot simply celebrate the productivity gain and remain silent about how anybody reaches the higher rungs.

And this is where two things that had been sitting separately in my head began to look different.

On one side are useful businesses whose owners need somebody to carry them forward.

On the other is a generation of capable people who need routes into responsibility, judgement and eventually some economic agency of their own.

Perhaps those are not entirely separate problems.

Perhaps, in some cases, one can become part of the answer to the other.

What if the next generation of entrepreneurs does not always have to start a company? What if some of them can inherit responsibility for one?

There is something important in that distinction for me, because it changes the idea from preserving jobs into something much more ambitious.

Human agency isn’t simply the right to keep doing whatever task a machine has not yet learned to do. It is the ability to act, make consequential decisions, shape an outcome and, perhaps, have some genuine stake in what you are shaping.

Automate the company, not the human

Before that idea really works, though, the business itself has to be worth inheriting.

A tired company that is entirely dependent on the founder, burdened by administration and unable to improve is not much of a platform for the next generation. Succession becomes more interesting if the person taking over can preserve what is valuable whilst building something more capable around it.

This is also where AI starts to look much more interesting.

Go back to the electrical contractor.

The technicians are good. The customers are happy. The founder knows the trade backwards. But the company itself is commercially fairly ordinary.

Leads arrive through email, WhatsApp, the website or somebody’s personal mobile. A quotation waits because Dave is on site and apparently Dave is the only person who understands how the current template works. Invoices go out later than they should. Important knowledge lives in people’s heads.

None of this means the company is bad. It means it was built around doing the work rather than designing a modern operating system.

That is exactly the sort of place where AI could be transformative.

Not because it can replace the electrician, but because it can remove a great deal of what makes being an electrician inside a small business unnecessarily frustrating.

An enquiry could be captured properly the first time. A quotation could be largely prepared from a site visit rather than reconstructed three days later from somebody’s notes. Technical knowledge that currently requires asking the founder could become available to the whole team. Scheduling and customer communication could improve. The owner might find out which work actually makes money before the accountant tells them six months later.

A ten-person company can start to possess capabilities that once belonged to a company with fifty. That leaves us with two very different questions.

How many people can we remove?

Or:

How much more capable can these people become?

I find the second question much more interesting.

The opportunity may be to automate the company around the human being rather than automate the human being out of the company.

And if the company becomes significantly better as a result, another question begins to appear.

Who gets to participate in that improvement?

The ones that interest me are the useful, durable businesses whose ownership problem is harder than their underlying business problem.

A bridge that doesn’t require everybody to be wealthy already

There is an obvious practical objection to the picture I have just painted.

Suppose the sixty-two-year-old electrician owns a good business worth a meaningful amount of money. Suppose there is also a thirty-year-old person who could, with time, become a very good leader of it.

The younger person probably doesn’t have enough money to buy the company.

Most people don’t.

If the route into ownership requires somebody to have accumulated a few million pounds before they begin, it is not much of a route.

But a fundamentally good company produces cash.

The retiring founder does not necessarily have to receive every pound of the purchase price on the day they hand over the keys. Part of the value can be paid over time from the cash the business continues to generate.

There are various terms for this — seller financing, deferred consideration, a vendor loan — but the concept is more straightforward than the language makes it sound.

In plain terms, a good business can pay for a meaningful part of its own generational handover.

This is not financial alchemy. The founder takes risk by waiting for some of their money, and the business still needs to be good enough to support the arrangement.

But it changes something quite profound.

You do not necessarily have to arrive already wealthy to become the next steward of a valuable business.

There is at least the possibility of earning your way into ownership.

I find that much more interesting than the conventional story we tell about entrepreneurship.

What energy taught me about participation

I came at the ownership part of this from somewhere else entirely.

Energy.

For most of the last century, households were passive consumers of a centrally controlled energy system. Distributed energy changes that relationship.

Put solar panels on the roof, a battery in the house or an EV on the drive and the household starts to possess useful assets of its own. With the right software, those assets can contribute to the wider energy system.

And if my battery or car is helping somebody else create economic value, I think a meaningful part of that value should come back to me.

That was the idea behind what I called the Participation Economy.

What interests me is not really the technology. It is the relationship.

People don’t feel part of something simply because somebody tells them they do.

In my experience if you can contribute something useful, you have some ability to affect the outcome and when value is created, some of it comes back to you in a meaningful way you can recognise. Then you feel part of something.

The more I thought about this, the less convinced I became that the principle belonged only to energy.

Do people feel that this value they create flows back to them in a meaningful day in their work today? Probably not always. To provide contribution and value add they look after customers, solve problems, teach colleagues, protect the reputation of the business and carry bits of knowledge that turn out to matter when something goes wrong on a wet Tuesday afternoon.

They create value.

Of course they are paid for their work, and I am not suggesting that wages somehow cease to matter because somebody has discovered equity.

But if somebody progressively takes more responsibility for making a company better, perhaps there should be a credible path towards ownerships towards participating in the value they create.

Responsibility, capability and economic participation can sit closer together.

That is what I have started to think of as Participation Ownership.

I am still working out exactly what I mean by that and how this model might work, but hopefully that is a healthy starting point.

A different kind of entrepreneur

The more I explore the idea, the more I wonder whether we have made entrepreneurship unnecessarily binary.

Either you work for somebody else or you start something from nothing or maybe get share options on the promise of an established valuable business one day.

Find an idea. Win the customers. Hire the people. Build all the systems. Raise some money if you can. Spend several years wondering whether the whole thing will survive.

For some people, that is exactly the right path.

But there are already millions of companies.

Some of them need the next generation much more urgently than the world needs another startup.

Why shouldn’t an ambitious engineer join one, learn the business, prove themselves, lead it and eventually become an owner?

Why shouldn’t a talented manager in Birmingham or Bristol take responsibility for a local business and consider that every bit as entrepreneurial as launching another software company in London?

Perhaps ownership can be earned through stewardship as well as invention.

I like that idea.

And what appeals to me is not only the ownership. It is what has to happen on the way there.

Somebody has to learn a business properly. They have to make decisions that carry consequences, earn the trust of people who may have worked there for twenty years, understand what the founder knows that nobody has ever written down and slowly acquire judgement of their own.

That is part of how people become capable.

In that sense, the succession problem might also create a new version of the ladder that technology is beginning to remove elsewhere. Not by inventing artificial training exercises, but by giving people real responsibility inside businesses where the outcome genuinely matters.

It also feels strangely suited to this particular moment because the person taking over does not have to reproduce the founder who came before them.

9 The founder may have built the company through thirty years of personal relationships, intuition and effort. The next person can preserve that accumulated trust whilst adding tools and capabilities the founder never had.

Better systems. Better information. AI.

A little more commercial sophistication without turning the place into a branch office of something enormous.

The previous generation’s knowledge becomes the starting point rather than something we accidentally throw away.

This feels less like disruption to me than evolution, and perhaps we have spent too long assuming that disruption is always the more interesting of the two.

Back to Charlotte

Which brings me back to the restaurant.

The point of the story is not really that Charlotte is bad.

For many things, Charlotte is probably extremely useful. If a human had answered the call with a friendly hello and I had simply wanted to make a booking, Charlotte might have been the ideal person to transfer me to. She is remarkably efficient, never gets tired and can deal with routine without interrupting somebody who is trying to serve dinner to a hundred people.

The mistake was not using artificial intelligence.

The mistake was failing to notice when the nature of the problem had changed.

And that is ultimately a leadership decision rather than a technology one. Somebody has to decide where these systems make people more capable and where they begin to get in the way of the judgement, curiosity and presence that the situation actually requires.

As the technology improves, that boundary will keep moving. We will have to keep deciding where to draw it.

At first I wanted information.

Then I needed presence.

I needed somebody to look.

And perhaps to stay curious.