A buyer does not count how many customers are in the file, but what happens if the largest two or three fall away. Customer concentration is one of the eight drivers that determine the price, and one of the few that cannot be disguised by a nice revenue line. A company whose revenue lies for the largest part with one customer is valued differently than a company with a hundred comparable customers, even if the revenue is identical. The buyer is not buying today's revenue, but the risk that this revenue will still exist next year.
This calculation is changing now that AI takes over parts of the work. Concentration was always about revenue: what percentage comes from the largest customer. But a buyer also looks at the cost side, and there a new kind of concentration arises that has nothing to do with customers. A saving that rests entirely on one AI supplier or one automated integration counts differently than a saving that is spread throughout the process and can be repeated with multiple suppliers or methods. Profit quality and customer concentration touch each other here: a margin that depends on one external party is a margin with a risk comparable to a revenue that depends on one customer.
A second layer comes on top of that. For part of the customer relationships, AI can take over the task: reporting, standard communication, first-line service. For another part, AI works alongside a human under supervision, where someone approves or rejects with reason — think of custom quotes or contract negotiations. And for a third part, it remains human work, especially where the relationship itself is the value: the account manager who has been visiting the largest customer for twelve years. A buyer wants to know into which of these three categories the most important customer relationships fall. Does the knowledge of the largest customer sit in a system that anyone can operate, or does it sit in the head of one person whom AI does not replace and whom the company also does not replace if he leaves?
A few signals can be uncovered with existing figures, without new research.
That last point is immediately the recurring question underlying this entire driver: which work in this company can genuinely be taken over by AI, and that is mapped per task with the work scan from FTE TO AI.
Customer concentration itself does not change overnight; winning new customers takes time. But the way a buyer assesses the risk can shift, and that happens in two places.
The first place is the process behind the largest customers. If the knowledge about that customer is fixed in a system that multiple people can operate — with AI keeping track of the overview and an employee assessing the exceptions — then the customer is still large, but the dependence on one person is smaller. This connects directly to how owner dependency is measured in a business where AI works alongside you, because often it is not the customer that forms the risk, but the one employee who manages the customer.
The second place is the contract. A large customer with an annual contract and a fixed payment structure weighs differently than a large customer who buys per assignment. Here customer concentration connects to how recurring revenue is measured in a business where AI works alongside you: a concentrated customer base with contractual repetition is a different risk than a concentrated customer base that has to be won anew every quarter.
Whether a buyer sees this depends on what is recorded. A figure on customer concentration without substantiation does not convince; a dashboard that shows per customer which work is automated, which work requires oversight, and which work hangs on one person, does. This connects to how management information is measured in a business where AI works alongside you — without that information, customer concentration remains a number without proof.
This is not advice on which customer should receive more or less attention, and no substantiation for personnel decisions around customer management. Anyone considering adjusting roles around customer management as a result of automation does so under their own applicable legal requirements.
The question is not only how concentrated the revenue is, but how that concentration is built up and how visible that is to someone who does not know the company. A first picture of that, and of the other seven drivers, is available in the free value check: eight short questions, one per driver, with a picture of which driver weighs most heavily on the price today. The full value scan, with evidence per driver and a two-year calendar toward the exit moment, is under construction. Anyone who wants to know first why a buyer looks at the management team and the structure around it in the first place can find that at why a buyer looks at the management team, and anyone who wants to see how all these drivers together form the price can read how the value of a company is determined.