Every board hears the same claims: AI changes everything, or AI is overrated. Both statements are too crude to act on. The question that can actually be answered is smaller and more specific: is AI, for a given task, in your sector, already taking over work, partly with oversight, or does it remain human work? That is not an opinion. It is a fact that differs per task and that can be verified.
You recognize hype by the scale of the claim: it applies to "the sector" or "the profession". You recognize a lasting shift by the scale of the evidence: it applies to a task, with a date on which that was last established, and with a reason why human oversight is still needed there, or no longer is. As soon as someone cannot break a claim down to that level, there is a good chance it is noise.
In one company a task is already being done with AI today and signed off by a human. In a comparable company, in the same sector, that is still done entirely by hand. That difference rarely comes from ambition or budget. It comes from who notices it first. An employee who works with a task daily often sees a shift sooner than a board that steers by quarterly reports; who in the organization notices first that AI is taking over work is therefore not always the person with the mandate to act on it. That delay between signal and decision is precisely where the distinction between hype and a lasting shift is made, or missed, in practice.
Every statement about "AI can take over this task" is an estimate, based on what examples, regulation and vendor offerings currently exist. That estimate is stronger the more easily the task can be isolated, and weaker the more the task relies on context that has not been documented. An estimate that is three months old may still be valid, or may already be outdated: that differs per task and per the speed at which the underlying model or regulation changes. An outcome that says "AI can take this over" says nothing about whether that actually happens; that depends on choices outside the estimate. And an outcome that says "this remains human work" is a snapshot, not a guarantee for next year.
Where this touches on who still performs which task, the following applies: an estimate is not grounds for a dismissal decision. That is subject to its own legal requirements, separate from what a sector picture or a self-assessment shows.
A shift is probably lasting, not incidental, when three things come together. First: the change is visible at multiple parties in the sector independently of each other, not just at one early adopter. Second: a reason can be identified why human oversight is or is no longer needed, rather than a general statement about "quality". Third: the shift persists across multiple measurements, not only the first time someone noticed it. If one of the three is missing, caution about withdrawing an assumption is warranted; it may be a temporary outlier rather than a structural change.
A strategy contains assumptions about what work costs, who does it, and how quickly a competitor can follow. As long as those assumptions hold, the strategy holds. The problem is not that assumptions become outdated; that always happens. The problem is that no one tracks it at the level of the assumption itself, so that a plan remains intact on paper while the reality beneath it has already shifted. That is also why how you keep a strategy current without revising it every quarter is not a matter of meeting more often, but of knowing per assumption when it was last tested.
Some decisions can wait until that becomes clearer. Which decisions you are better off postponing until the AI question is answered is itself a question that differs per situation, but the premise is the same: making a decision based on an assumption that no longer holds is more costly than parking a decision until the assumption has been reconfirmed.
A self-assessment by your own board is useful, but incomplete without a picture from outside. Sector data, regulatory clocks and public signals from competitors show whether a shift you already feel internally is also already playing out elsewhere, or whether you are ahead or behind. How you know whether your competitor is shifting faster than you is therefore not a matter of market research, but of the same kind of factual testing you apply to your own assumptions. The same applies to regulation: it often changes outside a board's field of view, until the moment it has to be put on the agenda, and how you get regulation structurally onto the board agenda is a way to bring that moment forward instead of being caught off guard by it.
All of this stands or falls on a distinction that is often overlooked: what the difference is between a strategy and a plan. A plan describes steps; a strategy describes assumptions. Only the latter can be tested at the moment the world beneath it shifts.
The underlying question, which work in this company can genuinely be taken over by AI, is answered per task with the work scan from FTE TO AI, independent of what sector-wide claims assert.
Anyone who wants to know where their own strategy currently stands can start with the free assumption check: a short round in which you name your key assumptions and see, per assumption, when it was last confirmed. The full stress test, with sector data, self-assessment and assumption list side by side, is under construction.
Stel uw vraag. Vaak zit de echte vraag een laag dieper — daar mag ik naar vragen.
Answers come from this site’s knowledge base. Not tailored advice, and not a scan of your company.