Point of view
In five years, your org chart will have two kinds of workers on it.
One kind takes holidays. Work is already being reorganised around outcomes rather than functions — and a growing share of the resources delivering those outcomes are agents. The leaders who get ahead will be the ones who know, at any moment, how many of each are at work in their business.
A new KPI
Revenue per employee. Span of control. Time to hire. And now: the human-agent ratio.
For most of a career, the leadership KPIs that mattered described the shape of the organisation. They assumed the organisation was made of people.
In a frontier firm, work is no longer organised by function. It is organised by project and outcome. Resources come together for a defined purpose, deliver, and move on. Some of those resources are people. A growing number of them are AI assistants. Each one shifts the ratio a little.
We are already seeing the early signals. Sales teams deploying agents to retrieve pricing and prepare call notes. Operations teams monitoring supplier performance. Finance drafting variance commentary. None of it announced itself as a transformation. All of it changed the ratio.
What is your human-agent ratio today? And where do you want it to be in twenty-four months?
Where companies actually are
Three postures. Most companies are on the first rung.
-
Rung 01
AI Curious
Experimenting with tools. No executive sponsorship, no governance, no plan. Real energy, no direction. Most of the market is here.
-
Rung 02
AI Adopters
Tools deployed and licences bought. Copilot is live. Usage exists. But nobody can connect any of it to a business outcome, and the CFO is starting to ask.
-
Rung 03
Frontier Firms
AI treated as a strategic business capability. Workflows redesigned around it, agents governed and owned, adoption measured, and outcomes reported to the board every quarter.
The gap between the second rung and the third is not technology. It is an operating model.
What leaders will be asked
Five questions your board has not asked yet, and will.
- How much of our work is done by agents?
- Not how many licences — how much work.
- Who owns them?
- Every agent in a frontier firm has a named business owner, not an IT ticket.
- How do we know they are still right?
- Agents drift. Data changes. Somebody has to review them.
- What did the last quarter's AI investment return?
- In the same units as every other investment.
- What happens to the people whose work changed?
- The answer to this determines whether any of the rest of it sticks.
AI success is less about technology and more about structured human change. Companies that treat it the other way round spend twice and land nothing.
The frontier firm is not a technology outcome. It is a management outcome.
The companies that pull ahead in the next three years will not be the ones running the most experiments. They will be the ones that treated AI as an operating model from the start.