Two years before a sale, you are at a different point than a year before, or at the moment a buyer is already at the table. There is still time to change something about how the company works, but not so much time that the consequences of that change only become visible five years later. What you establish now about the work in the company can still become demonstrable in the remaining two years — not as a promise to a buyer, but as something reflected in the figures and processes by the time they look in.
A buyer does not buy the number of people on the payroll. They buy what keeps running when part of those people are no longer there — the process, the customer relationship, the knowledge that isn't held in one head. AI is changing what keeps running without people, and that is already happening today, in parts. At one company, software already handles part of the administration independently; at another, the same work remains entirely manual, not because it couldn't be done otherwise, but because no one has separated the process from the person who has been doing it for years.
Three categories run through all that work. Some tasks AI can take over entirely, without a look from an employee still needed. Other tasks run partly automatically, with someone approving or rejecting and giving a reason for that. And a third part remains human work, because it requires judgment, relationships, or context that a system does not have. That division differs per company, and even per department within a single company. Two years before a sale is the moment to know which category the work in your company falls into, not to guess at it the moment a buyer asks about it themselves.
One year before a sale is the point past which there is time to still change something fundamental about a process — what happens then is described at what you do a year before the sale with this work. At two years, that is different. A process that currently leans partly on one person can, in that time, still be documented, redistributed, or supported by a system — and that can become visible in two consecutive years of financial statements instead of in an explanation a buyer has to take on your word.
That is also why owner dependency weighs most heavily here. Work transferred to AI changes the weight of every value driver, but most directly for owner dependency: if a task that currently sits with you can be transferred to a system with human oversight, part of the risk a buyer assigns to your departure disappears. A saving that arises because you spend less time on a task also counts differently than a saving that today depends on a single supplier or contract — earnings quality is not just how much is saved, but also how firmly that saving is anchored in the process. How a buyer weighs that exactly also depends partly on who will be sitting around the table; why a buyer looks at the management team plays a role in that which goes beyond the work itself.
What can be established at this point: which tasks in the company are already demonstrably being taken over by AI today, which run partly with oversight, and which remain entirely human work — and with whom that knowledge and responsibility currently lie. That is a factual inventory, not an estimate of the future.
What cannot be established: exactly how much a buyer will pay for that in two years, or what part of the work will then actually have been taken over. That depends on how the technology develops, on the sector, and on choices still to be made. Nothing is guaranteed here, then, about which task AI will actually take over or on what timeline. If the subject touches on personnel decisions — who continues to do which work and who does not — separate legal requirements apply to that; that assessment is up to the employer, here only facts about the work itself are mapped out.
The underlying question — which work in this specific company can genuinely be taken over by AI — is answered per task by FTE TO AI's work scan, regardless of whether a sale is already on the agenda. Even if a sale is not yet a concrete plan, that changes nothing about the relevance of the question; what that means is described at what you do with this work if you don't want to sell yet. And not every sale is a full transfer — at what you do with this work in a partial sale the question plays out just as much, but with a different outcome for what you still keep in your own hands.
Two years before the sale, there is room to establish what is weighing on the price today, and to see whether that can still be changed in the remaining time. A first step in that is the free value check: eight short questions, one per value driver, resulting in a picture of which driver is weighing most heavily on your price today. The full value scan — with a maturity score per driver, the owner dependency index, and a calendar toward the exit moment — is under construction. How exactly those eight drivers are weighed is explained at how the value of your company is determined.