In many companies, the price is not calculated, but determined. There is a spreadsheet, a rule of thumb, a feel for what a customer will still accept and what a competitor would offer. That feel usually sits with one person, often the owner or a calculator who has been there for years. New employees get the spreadsheet, not the reasoning behind it. This creates a calculation that exists on paper, but in practice only works as long as that one person is watching.
This did not arise from an unwillingness to transfer it. It arose because calculating relied largely on experience that is difficult to write down: which line item you stretch with a difficult customer, which discount can still be given without losing margin, when a quote is too sharp to be healthy. That kind of judgment was never fixed anywhere, because there was never a reason to fix it. It worked, so it stayed as it was.
The work of a calculator consists of three kinds of actions, and these three do not change at the same pace. Adding up cost prices, applying margins based on fixed rules, comparing with earlier quotes: that is work a system can perform consistently and verifiably. The judgment of whether a margin should be lowered in this specific case because of the relationship with the customer, the competitive pressure at that moment, or a strategic reason to want an order: that remains a human decision, supported by an overview of what was decided before and why. And a third category, such as finding patterns in hundreds of old quotes to see which exceptions are structural, is work where oversight remains necessary: the system proposes a pattern, a human approves or rejects it, with reason.
In companies where the calculation has already been partly documented -- in software, in an explicit set of rules, in recorded exceptions -- that third category can shift quickly to the first. In companies where the calculation still lives in the head of one person, that documentation must happen first before AI can do anything with it. So the difference between companies does not lie in how suitable the work is, but in how much of the reasoning already exists outside the calculator's head.
A buyer reviewing the books sees margins and quote history, but not why the margin is what it is. If the answer to that question lies only with the current calculator, the buyer is buying a pricing structure that cannot be checked and cannot be continued without that same person. This carries through into profit quality: a margin that demonstrably follows from documented rules counts more heavily than a margin that simply "feels" correct. It also carries through into the owner-dependency index, because calculation is typically one of the tasks that stalls first upon transfer, since no one else knows the exceptions.
This is not a personnel issue and not a reason to reconsider roles; what an employer does with its personnel falls under its own legal frameworks. The question here is whether the knowledge behind the price has been documented, independent of who applies it today.
The first step is making visible the rules that already exist, even if they were never written down. This is done by comparing old quotes and their outcomes: which discounts recurred more often, with which type of customer, under which circumstances. That overview is not a replacement for the calculator, it is a representation of what that calculator was already doing, but made readable for others.
After that, part of that overview can be recorded as a fixed rule, and another part as a guideline that a supervisor -- not necessarily the original calculator -- can apply and justify. The relationship with the calculator is not damaged in this process, as long as the documentation is done to support that person's knowledge, not to replace it for the moment that person is still present. What remains is a calculation that is demonstrable, even when the person who once built it is no longer watching.
Calculation is rarely the only place where this plays out. The same question comes back in how the planning depends less on one scheduler, in how customer relationships become transferable without losing the connection and in how you measure how much of the company runs through the owner. Which part of the calculation work in this specific company can genuinely be supported by AI is precisely the question that the FTE TO AI work scan answers per task.
A good starting point is to determine how much of today's pricing exists only in the head of one person, and how much of it can already be found in quotes, rules, or documented exceptions. The free value check offers a first indication: eight short questions, one per value driver, giving a picture of which driver is weighing most heavily on your company's price today. The full value scan, with maturity scores per driver and a two-year calendar toward the exit moment, is under construction.