A buyer has never looked at owner dependency in isolation from the rest of the business. The question was always: does revenue keep flowing if the owner is out for three months. What changes is not the question, but the answer a buyer expects to go with it today. In the past, the answer was often an organizational chart and a list of key figures. Now a buyer wants to know which part of that work no longer needs to sit with one person at all, because it is handled by a system with oversight.
That is the shift. Work that used to be doable only by the owner because he had the overview can now be taken over in parts by AI: compiling reports, monitoring margins per customer, flagging deviations before they become a problem. Where this happens with human oversight that approves or rejects with reason, the dependency shifts from a person to a process. Where this does not happen, the dependency stays exactly where it always sat.
A buyer does not weigh the number of hours an owner puts in, but what still keeps running without that owner. Three things come together here. First: is the knowledge locked in a head or in a system that can be consulted without that person. Second: are decisions that matter — pricing, customer acceptance, supplier selection — laid down in a working method that someone else can follow, or do they depend on experience that has not been written down. Third: who has the final say on exceptions, and can that say be placed somewhere other than with the owner himself.
AI changes the weight of each of these three. Work that is fully taken over by a system counts heavily if it has been structurally set up and not as a temporary fix. Work where AI proposes and an employee decides counts to the extent that decision-making authority genuinely lies with an employee and is not silently handed back to the owner. And work that remains human work changes nothing about the existing dependency — it only makes more visible how much of it still leans on one person.
The first place this becomes visible is the owner's calendar. Not how many hours are in it, but what type of decision keeps coming back. Second place: what happens if the owner is unreachable for two weeks. Does invoicing continue, are quotes sent out, is a complaint handled without anyone waiting for approval. Third place: how much of the company knowledge exists outside the owner's head — in a system, a process description, a dashboard someone else can read.
This depends heavily on the sector and on how the company has grown. In a company built around the founder's expertise, that knowledge often still sits in conversations and habits. In a company that has worked with systems from the start, that same knowledge is more often already laid down, and it is easier to see which part of it AI with oversight can take over and which part remains human work. Neither is inherently better for the price — it only determines where the work lies to reduce that dependency.
The owner dependency index maps this by placing three questions side by side: which tasks are today irreplaceably tied to the owner, which of those can technically be transferred to a system with oversight, and which structurally belong to the role of owner and do not change in that regard. This does not produce any percentage without explanation, because the outcome depends on the sector, the customer relationships and the way decisions are made within the company. It does provide an order of priority: where the dependency is greatest and where improvement has the fastest effect on how a buyer looks at the company.
This index does not stand apart from the other drivers. A saving that hangs on the owner counts differently than a saving that is embedded in the process, and that relates directly to profit quality — just as it relates to what succession means for the price a buyer is willing to pay. Anyone wondering how team structure and succession are weighed separately will find that worked out in how a buyer assesses team and succession now AI takes over work. And because systems and processes determine whether knowledge is transferable, that also includes how processes and systems are weighed differently now AI takes over work.
Improvement does not start with setting up AI, but with recording what knowledge and decisions currently sit with the owner alone. Only after that can it be determined which part of that AI with oversight can take over, which part can lie with an employee, and which part should structurally remain with the owner. Where this topic touches on personnel decisions — who takes over a task, whose role changes — separate legal requirements apply to that, apart from what is measured here.
The underlying question of which work in this company can genuinely be taken over by AI is answered per task in the work scan of FTE TO AI.
Anyone who wants to know whether owner dependency is currently the heaviest weight on the price, or whether another driver — such as why recurring revenue is weighed differently now AI takes over work — tips the scale more heavily, starts with 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.