A firm in financial services consists of layers that don't all do the same work. There is casework: intake, checking, processing changes, reporting. There is advisory work: the conversation in which a client makes a decision about a mortgage, an insurance policy, a pension question. And there is the layer around it — compliance, regulatory reporting, record-keeping that must be able to be checked by a regulator or an accountant.
These three layers don't weigh equally in a purchase price. Casework is measurable and repeatable, so relatively easy to transfer. Advisory work often hangs on a person and a relationship this person has built. Compliance is part of the licence itself, and a buyer isn't just buying revenue but also the right to be allowed to generate that revenue.
Within these hours, three categories run through each other. Part of the work can be taken over by a system: summarising policy terms, filling in an intake form based on documents supplied earlier, flagging a deviation in a file. Another part goes with oversight: the system draws up an advisory proposal, an adviser approves it or adjusts it, with a reason that is recorded. And part remains human work: the conversation in which a client hesitates, or a situation that doesn't fit the model.
Which category a task falls into depends on how standardised the question is and how heavy the consequences of an error weigh. An insurance change without customisation lends itself differently to being taken over than advice on the dissolution of a marriage with pension division. That difference doesn't sit in the sector as a whole, but in the firm: one firm has already largely standardised its casework, another still does everything the same way it did ten years ago. Both exist side by side, and the purchase price feels that difference sooner than the owner thinks.
If part of the casework is freed up, it's not just the cost structure that changes. It changes what a buyer sees when he looks at the eight drivers that determine the price.
Owner dependency weighs the heaviest. If the advisory work mainly sits with the owner personally — the client calls the owner, not the firm — then AI takeover of casework changes little about that. The question that counts is how do you measure how much runs through the owner, and that question sharpens as more execution work falls away and the remaining work becomes more visibly concentrated with one person.
Profit quality is the second driver that shifts. A saving on file processing that makes the process itself cheaper counts differently than a saving that depends on one software supplier or one subscription. A buyer wants to know whether the margin continues to exist without that one party, or whether the profit evaporates as soon as the contract changes.
In addition, transferability changes: work that runs through a system with recorded rules is easier to transfer to a new owner than work that sits in the head of one adviser. And it touches risk and compliance, because a regulator wants to know who is overseeing what, even when part of the process is automated.
This shift is not a reason to say anything about who a firm keeps employed. It is about which work, in hours and FTE capacity, can or cannot be transferred to a system, and what that means for the price a buyer is willing to pay. Decisions about personnel fall under their own statutory requirements, and that question lies outside what a value scan answers.
The way AI takeover shifts the value drivers looks different in every sector, because the ratio between standardised work and person-bound work differs. In construction, owner dependency often sits in estimating and client relationships, in the installation sector in planning and craftsmanship on site, and in the recreation sector precisely in the guest relationship that is hard to have taken over. Financial services stands out because part of the work falls under regulatory rules that require record-keeping, which makes automation both easier — rules can be coded — and riskier, because an error in an automated process repeats itself.
The question of which work in this particular firm can genuinely be taken over by AI cannot be answered at sector level. That is precisely what FTE TO AI's work scan answers per task. And for those already thinking about a sale, timing also plays a role: when should you start preparing a sale determines whether there is still room to adjust a driver that is currently weighing on the price before a buyer assesses it.
Anyone who wants to know where to start can take the free value check: eight short questions, one per driver, with a picture of which driver is weighing most heavily on the price today. The full value scan, with a maturity score per driver, an owner-dependency index and a two-year calendar towards the exit moment, is under construction.