In a partial sale, you don't sell the entire company and not just the remainder either. You sell a stake alongside a buyer who, after the transaction, watches, steers, or eventually takes over the rest. That changes what counts. A full acquisition is assessed on what remains standing once the previous owner leaves. A partial sale is assessed on what remains standing once the new co-owner starts watching how the work already runs today, with whom, and with what dependencies.
That is why the question about work that AI has taken over comes into sharper focus here than in a full exit. A buyer taking a hundred percent stake can restructure after the deal if needed. A buyer taking a partial stake must already be able to see today what the process itself does and what still hangs on people, because they will share that ratio with you for years to come.
AI is taking over tasks, not as a promise but as something that is already happening in parts. Three categories run through every company: work the software handles independently, work where an employee approves or rejects with reason, and work that remains human work. That division differs by company and by department, and that difference is exactly what a buyer looks at in a partial sale.
In one company, reporting, the initial triage of requests, or the drafting of quotes already largely runs through a system, with an employee assessing the exceptions. In another company, exactly the same work still rests entirely on one person, because no one has taken the time to break the process down in a way that lets a system take over part of it. Both companies can have the same revenue and the same workforce. In a partial sale they weigh differently, because the buyer has to keep working with exactly that difference after the deal.
Owner dependency weighs heaviest here, because a buyer taking a stake wants to know how much of the remaining work will still hang on you personally after the transaction. But the shift affects more than that one driver. Profit quality counts differently when a saving stems from a process that stands on its own, than when that saving depends on a single subscription with a single supplier that could stop or become more expensive tomorrow. Recurring revenue, customer concentration, and operational scalability shift along with it, each in its own way, because the question always comes back to what continues without a specific person or party being present.
This is also where this moment differs from a transfer within the family with the work AI is taking over or a management buy-out with the work AI has taken over: in those routes, someone who already knows the process often stays at the helm. In a partial sale to an external party, that knowledge is not automatically present, and the work itself has to tell the story.
You can establish now which work in your company is currently done by a system, which work runs with human oversight, and which work is still entirely human work. You can establish how much of that work hangs on you personally, and how much on a process that keeps running without you. You can establish whether a saving in operations stems from something repeatable, or from a construction tied to one contract or one person.
You cannot establish now what a buyer will think of that in two years, exactly what amount is attached to it, or which part of the work a buyer will want to organize differently after the deal. That depends on the buyer, on the market at that moment, and on how fast the underlying technology in your sector keeps advancing. It equally depends on questions that fall outside the scope of this page: if the redistribution of work touches on employment contracts or personnel, its own legal requirements apply, separate from what a buyer values.
The underlying question — which work in this company can genuinely be taken over by AI, now and in two years — is answered per task by the work scan from FTE TO AI, with evidence instead of an estimate.
The way you present that saving and that dependency to a buyer determines whether it counts as profit quality or is read as a risk; that is what how you substantiate profit quality is about. Also relevant is the question what recurring revenue is worth in a sale, because a partial buyer often pays as much attention to the stability of revenue as to the cost side. And if the figures have been moving the other way for a while, then what you do when your value declines while your revenue rises is a question that often shares the same cause as the one central here.
You can start by getting a picture of what is weighing most on your price today. The free value check consists of eight short questions, one per value driver, and gives an initial picture of which driver currently weighs heaviest. The full value scan, with a maturity score per driver, evidence, the owner dependency index, and a two-year calendar toward the exit moment, is under construction.