For a long time, headcount was a reasonably useful way of describing the size of a company. A 1,000-person business was, broadly speaking, a 10x bigger organisation than a 100-person business because most of the capability required to run it came from the people it employed. It wasn’t particularly sophisticated, but neither is counting people, which is probably why we’ve become so attached to it.
That relationship is starting to break. Two companies can employ exactly the same number of people and have radically different capacity to get things done. One might have 100 people operating largely within the limits of their own time, knowledge and expertise. Another can have those same 100 people surrounded by specialist partners, automated workflows, AI agents and an intelligence layer containing much of what the organisation already knows. Same headcount, very different company.
This matters because we’ve spent decades using headcount as shorthand for things it doesn’t actually measure particularly well: scale, capability, capacity and sometimes even progress. We approve headcount, freeze headcount, benchmark it and occasionally announce with considerable enthusiasm that we’ve reduced it. Then we wonder why the remaining people are spending half their week in meetings coordinating all the work that apparently became more productive.
Same headcount, very different company.
From the org chart to the work chart
There’s a broader rethink underway. Microsoft and LinkedIn have started talking about moving from the traditional “org chart” towards a “work chart”, mapping how work actually gets done rather than simply where people sit. Deloitte is making a similar argument around designing organisations from activities, decisions and outcomes rather than roles and reporting lines. Both are useful, but I think there’s another layer to it. The org chart tells you who works for the company. The work chart tells you how work gets done. Increasingly, neither tells you the full capability the company can bring to bear on it.
Designing a business that never had an org chart
We’re seeing this first-hand in AI-native businesses we’re helping build from the ground up, from personalised healthcare to pet services. These conversations start somewhere quite different from conventional workforce planning. Rather than asking how many people we need, we’re asking what the business needs to be capable of, what work actually creates value, where exceptional humans materially improve the outcome and what can be handled by technology or specialist capability around them.
In personalised healthcare, for example, we’re helping design a business where clinicians, diagnostics, health data, AI and continuous digital support are part of the operating model from day one. That changes the workforce question considerably. It isn’t simply how many clinicians or operators you need. It’s where clinical expertise and human judgement create disproportionate value, what intelligence can sit around those people and how the two combine to deliver something that would be difficult, and probably prohibitively expensive, through a conventional appointment-led model.
That’s one of the more interesting things about designing an AI-native business rather than retrofitting AI into an existing one. Established businesses understandably start with what they already have and ask how AI can make it faster. When you’re building from scratch, you can ask the slightly more uncomfortable question: knowing what we know now, would we actually design the company this way at all?
Quite often, you wouldn’t. Roles don’t have to arrive pre-packaged from an industry template. Capability doesn’t necessarily need to sit permanently on the payroll. A specialist you desperately need for six weeks doesn’t automatically need a desk for three years. And if technology can perform work perfectly well, we don’t need to turn it into somebody’s job simply because that’s what the org chart at our last company looked like.
Headcount as the consequence, not the starting point
Headcount starts becoming the consequence of organisational design rather than the starting point for it.
Historically, capability and headcount have been joined at the hip. Need more marketing capability, hire another marketer. Need more analysis, hire an analyst. Need more coordination because you’ve hired all those people, hire a manager. Before long the organisation has solved every capability problem with another salary and created several entirely new capability problems in the process.
There’s now a broader set of choices. You can build capability in the people you already have, buy it by hiring exceptional permanent talent, borrow specialist expertise when you need it, or bot work that doesn’t require a human doing it anymore. The point isn’t to replace “hire someone” with “use AI”, which would be an equally lazy answer. It’s to start with the work and make a deliberate decision about where the capability should come from.
The economics of exceptional talent
It also makes the economics of exceptional talent more interesting. If AI significantly increases what an outstanding engineer, marketer, product leader or clinician can do, it doesn’t necessarily make those people less valuable. It can make the difference between exceptional and average talent more economically significant because you’re multiplying judgement as well as output. Giving everyone a chainsaw doesn’t suddenly make everyone a good carpenter.
Giving everyone a chainsaw doesn’t suddenly make everyone a good carpenter.
Which is why I think the endless debate about how many jobs AI will eliminate is focused on the least interesting number. The better question for a business is how much capability each dollar of workforce investment can create. Revenue per employee becomes useful here, although even that starts behaving strangely when more of the productive capacity of the company sits outside the employee base.
What we should be measuring instead
Perhaps what we actually need to get better at measuring is organisational capacity itself. Who you employ remains enormously important, but so does what those people can access: the expertise around them, the workflows underneath them, the agents working alongside them and the accumulated intelligence of the company available when they make a decision. A brilliant person inside a terrible system and a brilliant person inside an exceptional one might look identical on the headcount report. They won’t look remotely identical in the P&L.
That’s why two 100-person companies can increasingly be completely different-sized businesses in everything except payroll. One might genuinely have the productive capacity of 100 people working within a conventional operating model. The other can have 100 exceptional people sitting at the centre of an ecosystem of intelligence and capability extending far beyond its payroll.
A brilliant person inside a terrible system and a brilliant person inside an exceptional one might look identical on the headcount report. They won’t look remotely identical in the P&L.
Calling them both 100-person companies is technically correct. It’s just not telling us very much anymore.
The better question
Headcount isn’t going away. People cost money, boards quite reasonably care about workforce investment and Finance is unlikely to surrender the spreadsheet without a fight. But as a shorthand for how much capability a company has, how much it can produce or ultimately how large it can become, headcount is losing some of its meaning.
For years, the workforce question was how many people do we need to build this company? Increasingly, the better question is what capability do we need to build it, and where should that capability come from?
The companies being designed around that question are going to look very different from the ones we’re used to counting.
Sources
- The 2025 Annual Work Trend Index: The Frontier Firm is born — Microsoft and LinkedIn. https://blogs.microsoft.com/blog/2025/04/23/the-2025-annual-work-trend-index-the-frontier-firm-is-born/
- From jobs to skills to outcomes: Rethinking how work gets done — Deloitte Insights. https://www.deloitte.com/us/en/insights/topics/talent/future-of-workforce-planning/planning-work-outcomes.html