Accessibility long counted as a result of workforce planning. Whoever put more people on the phone or behind the mailbox responded faster. Customers felt that difference directly: the supplier who responded within an hour won the deal that sat unanswered for three days at a competitor. The comparison between suppliers was in effect a comparison between staffing levels, opening hours, and the discipline with which people kept up with their inbox.
That assumption no longer holds everywhere. In places where AI takes over the first response, the triage, and part of the handling, accessibility is no longer a function of FTEs but of which tasks in that process have been handed over to a system that does not wait for the next working day.
Accessibility consists of a series of sub-tasks, and they do not all shift equally far.
Receiving and categorizing a query -- recognizing what someone needs, which department it belongs to -- is work that AI can take over in parts today. A first answer to a frequently asked question, a confirmation, a status update: that shifts too, provided the data the system relies on is in order.
In between sits a layer that runs partly automatically, with an employee who approves or rejects with reason: a proposed solution that a human reviews before it reaches the customer, or an escalation that a system flags but an employee handles.
And a third category remains human work: the customer who is angry and wants to be heard, the situation that does not fit a script, the decision that warrants an exception. Accessibility that consists only of speed misses this layer; accessibility that is competitive knows exactly where the boundary between the three categories runs within the organization itself.
The difference between companies lies not in ambition but in what is already in place. An organization with structured customer data, clear categories of queries, and a process that was already standardized for people can transfer part of the response without the customer losing quality. An organization where knowledge sits in people's heads and every contact is handled in its own way has little to hand over to a system; there, accessibility remains a staffing question, and the comparison with competitors changes more slowly.
This inevitably touches on the question of what happens to people when part of their tasks is taken over. Separate legal requirements apply to that; it is a decision for the employer and not a conclusion that follows from a task analysis.
If accessibility partly becomes a systems question, what the customer notices as the difference shifts too. Not just response time anymore, but the extent to which the first contact already resolves something without the customer waiting for a human. Two suppliers who used to score comparably on accessibility because both responded within a day can now be far apart: one resolves three-quarters of queries in the first contact, the other forwards everything to a queue that looks no different than it did five years ago.
The size of that difference depends on the sector, the type of query, and how much of the process has already been standardized; a specific percentage cannot be given without knowing the specific company and the specific tasks.
What can be said: accessibility as a claim -- "we are always accessible" -- means something different today than it did ten years ago, and a claim that does not indicate which part of the contact is handled by a system and which part by a human is harder to verify than it sounds.
Accessibility is often mentioned in the same breath as other factors on which customers compare suppliers, and the same shift plays out there just as much. How far an organization has come in this respect is linked to how far a company's digital processes have already developed, and anyone operating across multiple locations or countries will notice that accessibility is also a question about how consistent coverage is across different locations and time zones. Anyone who wants to know whether a claimed advantage on accessibility also holds up against what competitors actually demonstrate can test that via the method for checking whether distinctiveness is based on evidence.
The underlying question is always the same: which part of the work behind accessibility can, for this specific company, with these specific customer queries, actually be taken over by AI, which part requires oversight for a reason, and which part remains human work. That question is answered per task with the work scan from FTE TO AI, not with a general estimate about the sector.
What you can do now is hold your own claim up against your own body of evidence. The free dimension check lets you state where you believe you win on accessibility, and shows which of those claims can be defended with evidence and which rest on assumption. The full benchmark, with comparison against your peer group, is under construction.