When two competitors deploy the same AI application for quotes, delivery times or customer contact, the difference that dimension once caused disappears. What used to be an advantage because it was scarce — fast planning, an accurate quote, a helpdesk reachable 24/7 — becomes a basic requirement as soon as the underlying task is taken over by AI. Not because the dimension becomes unimportant to the customer, but because everyone scores equally on it.
The question, then, is not whether AI makes dimensions worthless. That is already happening, task by task, in the company you are looking at today. The question is which dimensions this concerns, and which ones actually gain value because they remain the only thing left on which a difference can still be made.
The speed at which a dimension empties out depends on how much of the underlying work AI can take over, and how much of that continues to run with human oversight. A planning dimension that can run entirely on AI hollows out faster than an advisory dimension in which a specialist still assesses and corrects the outcome. And work that remains human work — negotiation, handling exceptions, building trust — remains a distinguishing dimension, precisely because it cannot be copied with the same software.
This also explains why one company already notices that a dimension no longer yields anything, while another still claims an advantage on something that is already commonplace elsewhere. It is not a matter of sector, it is a matter of what has already been taken over per company. Those who see how a new entrant without staff builds a market position often see that this entrant skips exactly the dimensions that existing players still staff, and starts directly with AI on dimensions where established parties still think they are winning.
The benchmark scores your position and that of your peer group on the dimensions that decide deals in your market. Every score can be traced back to an evidence matrix: observations, not assumptions. That is the core of the method, and also its limit.
The method establishes whether a dimension is currently distinguishing within the compared group. It does not establish whether that will still be the case a year from now. If competitors in your market currently still largely work manually on a dimension, that dimension can empty out tomorrow as soon as one of them deploys AI for the underlying work. The score is a snapshot, not a guarantee of that snapshot's shelf life. That is also why repetition is needed: how often a comparison with competitors needs to be redone depends on how fast AI takes over the work behind the measured dimensions in your sector, not on a fixed calendar.
There is another limit. A dimension can score low in the evidence matrix without this saying anything about the quality of the underlying work. It means there is insufficient externally observable evidence to substantiate a claim. That is a different problem from weak execution, and it requires a different approach: how you substantiate a claim about your own quality is about gathering evidence, not about improving the work itself.
Those responsible for commercial results often lose deals not because a dimension is weak, but because a claim on that dimension no longer holds while the sales story has remained unchanged. "We are faster" was true when speed depended on staffing. As soon as a competitor runs planning on AI with an employee who now only assesses exceptions, that same claim is no longer true, even though nothing has changed about what your company does.
The claim test lays every claim to an advantage alongside the evidence available for the peer group. This prevents a comparison from being built on an assumption that still lives internally but has already been overtaken externally. The answer to what there still is to compete on when AI does the executing work therefore also differs per company, and changes as soon as something shifts within the peer group.
If a dimension becomes worthless because the entire market scores equally on it, the first question is not how you are going to win on it again. Often that is no longer possible, and the energy is better spent on a dimension that still distinguishes. What to do with a lag on a dimension depends on whether that dimension is still movable, or whether the market has by now outgrown it.
This inevitably touches on the question of which work in your own company can already be taken over by AI and which work remains human work. That question is answered by the FTE TO AI work scan, task by task. Where an outcome touches on personnel decisions, its own statutory requirements apply; the scan and the benchmark provide facts about work, not substantiation for those decisions.
Name the two or three dimensions on which you believe you are winning. For each one, check whether that winning still rests on something scarce, or on something that by now moves equally fast everywhere. The free dimension check shows which of those claims can currently be defended with evidence within your peer group. The full benchmark, with the complete evidence matrix per dimension, is under construction.