Most owners only start calculating once a buyer is sitting at the table. That's when it becomes clear which part of the work actually depends on people, which part has already been placed elsewhere, and which part is in fact being done by software without anyone ever having looked at it. That moment comes too late to still change anything about what a buyer sees. The question of what AI is already taking over today does not belong to the sale. It belongs to today, regardless of whether you sell in two years or in ten.
AI is taking over work. Not as an announcement, not as something that will happen in a few years, but as something that is happening today, in parts, further along at one company than at another. Three categories run through every role and every process: work that AI can fully take over, work where AI makes a proposal and a human approves or rejects it with reason, and work that remains human work. Which tasks fall into which category differs per company. It depends on how standardized the work is, how well the underlying data has been recorded, and whether systems already exist that can feed that work. A company with documented processes and clean administration is further along here than a company where a lot of knowledge sits in people's heads and little is on paper.
That difference is not coincidental. It is the result of choices made years ago about record-keeping, systems, and the transfer of knowledge. Companies where that is in order see the effect of AI reflected fastest in freed-up hours. Companies where that is not in order still see mostly the promise, for now, not the execution.
A buyer does not buy the number of people who work there. He buys what keeps running without those people. Work that AI takes over therefore changes the weight of every value driver a buyer factors in, even if you have no buyer in mind today. Owner dependency weighs heaviest: if knowledge and decisions rest with AI-supported processes rather than with one person, that changes the company's dependency structure, regardless of who takes it over later. But profit quality also shifts. A saving that arises because one supplier cheaply takes over a task counts differently than a saving that is built into the company's own process and continues to exist if that supplier falls away. That distinction is already visible in the books now, even though no sale is planned.
If you are not selling now, there is no need to change anything about staffing today. What an employer does with its staff is up to the employer, and if that question arises, its own statutory requirements apply. What this is about is capacity: which hours in the process are freed up because a task is done or supported by AI, and what that means for how the business keeps functioning without you, or without specific people.
At this point you can establish which work falls into which of the three categories: fully transferable, partially with oversight, or human work. You can establish where the gain from AI use lies within the process itself and where it depends on an external party. You can establish whether knowledge about critical tasks is documented or rests only with you or a key person.
What you cannot establish now is exactly what a buyer will pay for that in two years. Market conditions, the type of buyer, and the sector all help determine how heavily each driver counts at the moment of sale. That uncertainty does not disappear by measuring now. What does disappear is the uncertainty about where you stand today, and that is the part you can already start adjusting, well before a sale moment presents itself.
This question touches on a number of situations that seem different at first glance but share the same core: what remains of the business as the work changes. In a partial sale, the question is how AI takeover of work affects the valuation of the remaining stake, in a transfer within the family, it concerns what the successor takes over if part of the work is already being done by AI, and in case of doubt about the underlying figures, it helps to know what is needed to professionalize management information to the point where a buyer accepts it without a discount. The underlying question, which work in this specific business can be taken over by AI, is answered per task by FTE TO AI's work scan.
You don't have to wait for a sale moment to know where you stand. 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 a maturity score per driver, evidence, the owner dependency index, and a two-year calendar toward a possible exit moment, is under construction. Anyone who wants to know now what is weighing on the books, even without a plan to sell, can start with the question what concretely drags down the value of a business and where that can be seen.