Recurring revenue is one of the eight drivers a buyer uses to determine the price of a business. That a portion of revenue returns every year is not, in itself, proof of value. The question a buyer asks is why that revenue comes back, and whether that why keeps working when something changes in how the work gets done. Now that AI is taking over work in parts of every business, that exact why is changing, and with it the weight a buyer assigns to this driver.
A buyer does not look at the percentage of recurring revenue as an isolated number. He looks at the source of that repetition. Revenue that returns because a customer has to choose between you and an alternative every year, and happens to keep choosing you, is different from revenue that returns because a contract, an integration, or a process makes that choice unnecessary. The first is habit. The second is structure. A buyer pays for structure, not for habit, even though both look identical in the books.
The shift now underway changes where that structure sits. Work that used to be doable only by a fixed point of contact, and that thereby tied the customer relationship to that person, is being taken over in parts by AI, with human oversight approving or rejecting with reason. That changes the nature of the recurring revenue itself: it shifts from "the customer comes back for this person" to "the customer comes back for this process." For a buyer, the second is considerably less risky.
The first place to look is the contract itself. Recurring revenue on automatic renewal weighs differently than revenue that is renegotiated annually. But the second place, which is more often overlooked, is the execution behind the contract. Two businesses with identical contract terms can have very different risk profiles, depending on whether the service delivery behind that contract is captured in a system or sits in the heads of a few people.
A signal a buyer takes seriously here is what happens when a regular employee is on vacation or leaves. Does the service continue without the customer noticing anything, or does delay, noise, or a need for the owner to step in personally arise. The latter connects directly to another driver: how much of the customer contact runs through the owner at all, something that can be measured separately and concretely via how you measure how much revenue, decisions, and customer contact run through the owner.
A useful way to start is to record, per major recurring customer or contract, which part of the work behind it follows fixed steps, which part depends on judgment by a specific person, and which part is fully automated already or technically ready to be automated. The latter is exactly the distinction between the three categories that run through everything: tasks AI can take over, tasks where AI does most of the work under oversight, and tasks that remain human work.
Where that split lies differs strongly per business, and that difference is usually not accidental. It relates to how processes are documented, which systems are in place, and whether the work is broken down in such a way that part of it can be handed over to a machine without changing the quality of the service. Businesses where this has already happened find that their recurring revenue becomes sturdier: less dependent on individual availability, more dependent on a process that demonstrably keeps running. Businesses where this has not yet happened often have the same revenue, but not the same certainty about it, and a buyer prices in that difference.
This measurement does not stand apart from the other drivers. A saving or efficiency that only arises thanks to one specific supplier or tool counts differently for a buyer than a saving embedded in the process itself, and that affects earnings quality just as much as recurring revenue. Whether management information makes this kind of dependency visible is also relevant, as described in how management information changes when reports and analyses are largely generated automatically. And because AI taking over work also affects how a team is structured and who actually does which part of the work, there is a connection with what happens to team and succession when tasks shift to AI with oversight. Where any of these insights touch on personnel decisions, separate legal requirements apply; that is for the employer to assess, not a scan.
To make recurring revenue weigh more heavily, it is not necessary to sign more contracts. It is necessary to make visible that the service delivery behind existing contracts does not depend on one person or one informal arrangement. That means recording which steps have a fixed, repeatable form, which steps require judgment that is given or withheld with reason, and which steps still rest entirely on one person's experience. Where the latter is the case, room opens up to determine whether AI can take over part of that work there, and which part remains human work, regardless of what happens with that time afterward.
This question, which work in this specific business can genuinely be taken over by AI, is answered task by task in the FTE TO AI work scan. For those who first want to know how recurring revenue relates to the other seven drivers, such as market position or the moment at which preparing for a sale becomes worthwhile, an overview of the moment at which preparing for a sale starts to pay off and how AI changes the negotiating position toward customers and competitors offer an additional perspective.
Those who want to know where to start can take the free value check: eight short questions, one per driver, resulting in a picture of which driver is weighing most heavily on the business's price today. The full value scan, with evidence per driver and a two-year calendar toward the exit moment, is under construction.