A buyer looking at customer concentration is actually asking a different question: what happens to profit if the largest customer leaves. Until recently, the answer was mostly a matter of dividing revenue by number of customers. That calculation is now changing as part of the work behind that revenue is done by AI, with varying degrees of oversight, and not at the same pace in every company.
Customer concentration says something about dependency, but the question that follows is always: dependency on what. On the customer itself, or on the people who work for that customer. A company where one customer represents a large share of revenue, but where the underlying work largely runs through fixed processes and systems, is different from a company where that same revenue depends on a small team that happens to get along well with that customer.
AI changes that second scenario first. Work that used to exist only in the heads of specific employees — reporting, coordination, first drafts, checking for deviations — can partly be taken over by systems that work with oversight: an employee approves or rejects, with reason, but no longer has to do the work itself. Where that already works this way, a large customer becomes less a risk tied to people and more a risk tied to process. A buyer prices in that difference.
In practice, a buyer does not look at the revenue percentage of the largest customer, but at what happens if that customer leaves: which work stops, which people become redundant for other customers, which knowledge disappears. Concretely, a buyer pays attention to:
This also directly touches on how owner dependency is weighed, because a large customer that runs only through the owner counts more heavily than a large customer served through a team and a system.
Measuring starts with the work itself, not with the revenue figure. For the largest customers, whether there are two or ten, it is possible to review which part of the work falls into three categories: work a system can already take over, work that can be done with oversight, and work that remains human work. The larger the share of the first two categories, the less the relationship with that customer rests on individual employees.
In addition, it is worth looking at the profit quality of that customer revenue. A saving or margin that hinges on one supplier or one contact person weighs differently than a saving that is anchored in the process itself. This connects to the question of how a buyer looks at recurring revenue now that AI is taking over part of the work behind it, because there too, it is not the revenue itself that counts, but what that revenue still costs in human work.
Anyone wanting to know how much of the contact with the largest customers actually runs through one person will find a concrete approach in how you measure how much runs through the owner or key people.
Reducing customer concentration itself — more customers, smaller shares — takes time and is not always desirable or achievable. What can be influenced in the shorter term is the dependency on people within those customer relationships. That means documenting work in processes, standardizing reporting, and where possible, placing parts of the work with systems under human oversight instead of with one specialist.
Good management information plays a role here: a buyer who can see which part of the work at the largest customer already functions independently of specific people needs evidence for that, not just an assurance. See also why management information weighs differently now that AI is taking over reporting work. A team trained more broadly than a single account holder also makes a difference, as described in why team and succession weigh differently now that AI is taking over work.
Where this touches on personnel decisions, separate legal requirements apply; that is not part of this consideration. This is about facts regarding work: what a system can take over, what requires oversight, and what remains human work — not about what an employer does with that.
The underlying question, which work in your company can genuinely be taken over by AI, is answered per task with the work scan from FTE TO AI. When that process can be started and how much preparation time it requires depends on the situation; an indication is given at when you can start preparing a sale.
To see whether customer concentration is currently the driver putting the most pressure on your price, or whether another driver weighs more heavily, the free value check is a first step: eight short questions, one per value driver, giving a picture of which driver has the most effect on your company's value today. The full value scan, with maturity scores per driver and a two-year calendar toward the exit moment, is under development.