A customer who places three quotes side by side is not comparing your process but its outcome: response time, the precision of a proposal, how quickly a question gets a concrete answer. If a competitor has already handed part of that work over to AI, with or without human oversight of it, that outcome changes before any report appears internally about it. The customer sees the difference at the third quote. You only see it at the quarterly figures, and only then if you manage to link the loss to a cause.
This is not a matter of attentiveness. It is a matter of where the evidence lands first. Internal signals travel through reporting cycles. Customer signals travel through the offer itself, so in real time.
The shift does not lie in "AI does more work". It lies in what a dimension in the comparison means once the work behind it changes character. Delivery time, based on planning capacity, used to be a matter of staffing and workload. Once planning itself largely runs automatically, with a human approving or rejecting outcomes, delivery time becomes a matter of how well that system is set up, not of how many people are sitting at the desk. Whoever implements that earlier wins on a dimension where staffing used to set the ceiling.
That does not happen everywhere at the same pace. Some tasks lend themselves to being taken over entirely by AI, others only partly with oversight that approves or rejects for a reason, and part remains human work because the nature of the task requires it. Which category applies to which dimension differs per company and per process. That is exactly why the question of whether your strongest point is now being matched by AI is not rhetorical: a distinguishing capability that rested on human work can become a commonplace once competitors have automated it and you have not.
A benchmark that wants to make this visible works with estimates, and those estimates are not equally certain everywhere. Job postings, turnaround times in public communication, complaint patterns on review platforms: these are indirect signals. They tell you something about probable setup, not about the exact state of affairs behind a competitor's wall. When a competitor publishes no figures, it is up to the researcher to work with that kind of indirect source, and the approach to obtaining usable information about a competitor with no published figures is explicitly an approach of probabilities, not certainties.
There are also outcomes that say nothing at all. A score on a dimension on which you will never compete is noise in the report, not insight. Before a score carries weight, it must be established that the dimension in question actually decides the deal with the customers you want to win. How you make that distinction is described at the question of how you weigh a dimension on which you will never win anyway. Without that step, every benchmark produces a precise score on an unimportant question.
A third limitation: all parties in a market often phrase their value proposition in similar language. "Personal approach", "tailor-made", "short lines of communication" recur with almost every provider, regardless of what actually lies behind those words. That makes it difficult to steer by the claim itself; the reason competitors sound almost identical in their promises lies not in a lack of originality but in the nature of marketing language, and it is exactly why a claim test is needed before a score is based on a claim.
Not every signal is equally soft. Staff turnover, for instance, is a hard indicator that says something about internal stability, and with a competitor that changes frequently in key positions that is an indication that processes are less firmly anchored than the outside suggests; what a competitor's staff turnover says about its position is therefore one of the few dimensions where public data and actual internal state run reasonably parallel. Something similar applies to AI maturity: that can be measured from the outside based on public traces, and the way to measure a competitor's AI maturity from the outside shows which traces those are and what they do and do not prove.
The underlying question, which work in your own company can genuinely be taken over by AI, is answered by the FTE TO AI work scan per task, with the distinction between full takeover, oversight with review, and work that remains human work. For decisions about staff deployment, separate statutory requirements apply, incidentally; this page and the accompanying method provide no substantiation for that.
The free dimension check lets you name where you think you win, and tests per claim whether there is evidence to support it or not. The full benchmark, with the complete peer group and evidence matrix, is under construction.
Vraag maar waarop er in uw markt gewonnen wordt. Ik vergelijk liever dan dat ik uitleg.
Answers come from this site’s knowledge base. Not tailored advice, and not a scan of your company.