A customer cancels. The reason you get is vague: "we chose a different provider" or "it fit better with what we needed". Later you hear from a colleague or from the customer themselves that the competitor is doing something with AI: faster quotes, a chatbot that answers instantly, a schedule that adapts at lightning speed. You conclude that AI is the reason. That may be true, but it is one of several possible explanations, and the explanations call for different responses.
If part of the work behind a dimension is taken over by AI, it is not the dimension itself that changes but the standard on it. Delivery time was for years a matter of staffing levels and planning discipline. Companies differed on how good their planners were. If planning largely runs itself, with a human approving or rejecting exceptions, the comparison shifts: it is no longer about who has the best planners, but about who has already placed their planning work with a system that reacts faster than a human can. Anyone who suspects this cause recognizes it in a competitor who responds remarkably quickly to changes, while the standard lead time is not necessarily shorter. You recognize it in customers who specifically mention that they got a response "within an hour" or "immediately", not that the response was better.
If AI replaces costly human work at the core of an offering, that can translate into a lower price without the customer noticing anything about AI. The customer then simply compares two quotes and picks the cheapest one. This looks like a loss to an AI-driven competitor but is actually a price shift. The pattern that goes with this is described on what it means when a competitor suddenly operates at lower prices than before. If the reason for leaving is consistently price and not speed or convenience, that points more in this direction.
Not every loss to a competitor is connected to AI. Sometimes there is a new entrant that simply operates a different model, or the market itself shifts in such a way that no one is clear anymore about what the winning factor is. Both patterns can be distinguished from an AI-driven shift: with a new entrant that quickly wins customers, the explanation often lies in a different revenue model or a different target audience, not in automation. With a market in which no one knows anymore what makes the difference, the problem is that the winning dimension itself has become unclear, which can just as easily stem from AI as from something else.
One customer leaving is an incident. A declining win rate across multiple deals is a pattern. The distinction matters because the response differs: you respond to an incident by questioning the customer, and to a pattern by remaking the comparison itself. What a declining win rate specifically does and does not say is worked out on what a structurally declining win rate means versus incidental losses.
The shift does not move at the same pace everywhere. In some sectors, most of the quoting work is already largely automated, with a salesperson approving the outcome. In other sectors that remains human work, either because the complexity of the request cannot be captured in a system, or because the customer explicitly wants personal contact. That difference does not lie in the sector as a whole but in the task: one piece of quoting work lends itself to being taken over, another does not. What is genuinely possible for AI to take over in your own company is not something to estimate but something to find out per task; that is exactly what the FTE TO AI work scan is for.
You cannot infer from a handful of departing customers that you have an AI shortfall across the entire board. Nor can you conclude that staff are the cause of the loss; what an employer does with its staffing levels is up to the employer, and decisions about that carry their own legal requirements which are not addressed here. What can be investigated is: on which dimension you lost the customer, whether that dimension has recently shifted due to automation at the competitor, and whether your own peer group shows the same pattern. How you define that peer group is described on how to compile a peer group that truly reflects your market, and which room in the market is still unused is described on what a white space analysis shows about unused market room.
Name the two or three dimensions on which you believe you win against this competitor and test them with the free dimension check: you will see which of those claims can be defended with evidence and which rest on assumption. The full benchmark, with peer group and evidence matrix per dimension, is under construction.