An earn-out splits the purchase price into two parts. One part is paid at transfer, the other follows later and depends on results that still have to prove themselves after the transfer: revenue that holds up, customers who stay, a margin that remains intact without the departed owner. The buyer does not propose this because he doubts your past figures. He proposes it because he doubts their repeatability once the situation changes — and that transfer is exactly such a change.
The earn-out is therefore not a negotiating trick. It is a pricing mechanism for uncertainty that the buyer cannot remove himself simply by calculating better. The greater that uncertainty, the greater the part of the price that is deferred.
The biggest source of that uncertainty has always been the owner: whether the customer relationship, the pricing agreement, the knowledge of exceptions runs through one person who leaves after the deal. That risk still exists, but a second layer has been added. In an increasing number of businesses, part of the work is now done or supported by AI — in some fully, in others with an employee who assesses and corrects the outcome, and in yet others not at all, because the nature of the work leaves no room for it or because the organization has not yet set it up. That division differs in every business and is measurable.
For a buyer, that is relevant because it changes the question an earn-out tries to answer. Not only: does the customer and the revenue stay. Also: is the saving or the capacity that AI has freed up something embedded in the process, or something tied to one person, one subscription, or one incidental setup. A saving that disappears as soon as the owner leaves weighs just as heavily during the earn-out period as a customer who leaves.
Businesses where this is already visible generally have one thing in common: the work AI takes over is described at the level of the task, not the role. It has been recorded which part of a process runs without supervision, which part with an employee's check, and which part remains human work and why. That record exists independently of the person currently doing it. A buyer who sees that has less reason to put the uncertainty into an earn-out — the evidence for repeatability is already there.
Businesses where the earn-out instead grows are often businesses where nobody has asked that question. The saving is mentioned in a pitch, but not substantiated at the level a buyer can verify. Those who want to know what that substantiation looks like will find it worked out in how you substantiate a saving generated by AI to a buyer.
An earn-out reduces the buyer's risk, but changes nothing about the structure of the business. If a large part of the profit runs through the owner — decisions, customer contact, knowledge of exceptions — that remains the case until the earn-out period is over, and then it becomes visible again. How much of that actually runs through the owner, and not through the process, is a question that can be measured; how you measure how much runs through the owner describes what that measurement is based on.
In addition, an earn-out says something about the profit that serves as the basis for the price, not only about continuity. A profit that has been cleaned up with one-off savings or with costs that return after the transfer stands on loose ground. What normalization of profit means when AI plays a role in it is worked out on what normalization of profit means for a business that uses AI.
The underlying question in every earn-out is the same: which part of the work in this business continues without the people currently doing it, and which part does not. The work scan from FTE TO AI answers that question per task, with the same breakdown a buyer makes himself: automatic, under supervision, or human work. This is not personnel advice and not substantiation for a dismissal decision — separate legal requirements apply for that — it is a description of what the work itself allows.
The earn-out question usually only comes up once negotiations are already underway, but the substantiation for it needs to be ready earlier. When that preparation is best started is described on when you should start preparing for a sale. And because market expectations around AI are shifting in the meantime as well, it is good to know how you keep your business value steady while that shift continues — that is on how you keep your value steady while the market shifts.
This explanation is not an estimate of the earn-out that would be proposed in your case, and no guarantee that a buyer uses or forgoes this instrument. The amount and conditions of an earn-out follow from the negotiation itself, not from a generic description.
The free value check consists of eight short questions, one per value driver, and gives you a picture of which driver is weighing most heavily on your price today. The full value scan — with evidence per driver, the owner-dependency index, and a two-year calendar toward the exit moment — is under construction.