In an organization where AI is starting to take over tasks, this does not happen simultaneously for everyone. The employee who performs the task daily usually notices it first: part of the work goes faster, or a tool does something for which a colleague was previously needed. The manager sees it later, when the pattern repeats and someone chooses to report it — or not. The executive board sees it only when it translates into a figure: less hiring, a vacancy that is not filled, a team that handles more work without expansion.
That order is the problem. Not because anyone is negligent, but because the place where a shift becomes visible first — the work floor — is usually not the place where strategic assumptions get adjusted. By the time the shift reaches the boardroom table, the assumption has sometimes already been outdated for months.
The three categories that keep recurring here: tasks AI can take over completely, tasks where AI does the work under human oversight that approves or rejects with reason, and tasks that remain human work. None of the three is fixed. A task that fell into the third category last year may sit in the second this year. This is not something established once and then valid forever — it is something that shifts, per task, per team, and not everywhere at the same pace.
This is why two comparable companies in the same sector see a different reality. One company has processes that are easily split into repeatable steps; there the shift is visible and fast. The other company has work that is heavily intertwined with exceptions, context, and negotiation; there less changes, or only later. The difference does not lie in ambition or budget. It lies in the nature of the work itself, and in how well someone in the organization knows that work.
The executive board steers on aggregates: revenue per team, lead time per process, cost per FTE. A shift at the task level usually only shows through there once it is large enough to affect the aggregate. Until then, the assumption underlying the strategy — "this process requires this staffing level," "this competitor cannot do this faster" — simply continues to run, even when it no longer holds. No one withdraws it, because no one has the task of doing so.
This is what makes the topic difficult to place correctly. What an individual manager does with their team when a task disappears or changes is up to that manager and falls under their own legal requirements; we offer no advice on that. What matters here is something else: whether the assumption on which the strategy rests still matches what is actually happening.
It is not necessary to wait until the shift appears in an aggregate. There are three places where the signal can be picked up earlier than at the executive board itself.
The first place is the task itself: who performs it, and is that person still needed for the full set of tasks or for part of it. The second place is the assumption attached to the task — what a strategic assumption exactly is and how you check whether it still holds determines whether anyone ever deliberately recorded that assumption, and therefore whether it can ever be deliberately revised. The third place is the picture within the executive board itself: why the pictures within an executive team about the same topic diverge shows that three board members can see the same market and hold three different timelines for when something becomes relevant.
A self-plot of the executive team alongside external signals does not produce a forecast. It produces a snapshot of this moment: this is what the sector shows, this is what the team thinks, and here is where the difference lies. Whether that difference still exists in six months cannot be guaranteed — sector data changes, regulations shift, a competitor may itself move faster or slower than its public signals suggest.
The outcome also means nothing if the input is not accurate. An assumption list is only as good as the honesty with which someone admits that an assumption has actually not been checked for a year. And the method says nothing about what an executive board should do with that information — that remains a choice that lies with the executive board itself, not with an estimate.
The underlying question — which work in this specific company can genuinely be taken over by AI, in which category and with which oversight — is answered per task with the work scan from FTE TO AI.
Do not start with an estimate of how much work is changing, but with the question of which assumptions underlie the current strategy and when they were last tested. This connects to how you keep a strategy current without doing a full review every quarter, and to the question which decisions are better postponed until it is clear how AI is developing in this specific case.
Anyone who wants to know where that stands for their own organization can start with the free assumption check: a short round in which you identify the key assumptions underlying your strategy and see, per assumption, when it was last confirmed. The full proof print, with sector data, self-plot, 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.