A risk is on a list. Someone has named it, a probability and an impact have been attached to it, and it is reviewed periodically. An assumption is on no list at all. It is woven into the plan itself: "the lead time of this process will remain roughly the same," "this department will keep doing this work," "our competitor cannot do this faster than we can." No one has ever written that sentence down separately, because it seemed self-evident at the moment the strategy was set.
The difference is not academic. A risk is monitored because it has been recognized as a risk. An assumption is not monitored, precisely because no one has recognized it as an assumption. It disappears into the text of the plan and stays there, even as the world around it changes.
AI taking over work is not a risk in the classic sense. It is not an event that occurs and then is over. It is a slow shift that has a different pace per task, per department and per company. For one company it means that part of drawing up reports is now done by a system within a few weeks, with an employee approving or rejecting with reason. For another, nothing changes this quarter, because the work requires too much context, negotiation or responsibility to hand over to a system.
That difference in pace is not a matter of chance. It lies in what kind of work it is: routine and well-defined, or dependent on judgment and exceptions. That same structure also determines which assumption in a strategy comes under pressure first. A plan that relies on the lead time of a department that mostly does work that can be taken over ages faster than a plan that relies on work that, for now, remains human work.
The shift itself is therefore not the risk. The risk is that no one recognizes the assumption affected by it as something that needs to be checked.
The strategic pressure test places two pictures side by side: what sector data, regulatory clocks and public competitive signals show from the outside, and what the company's own executive team sees and expects internally. Where those two pictures diverge, an assumption emerges that deserves attention. Alongside that is a list: per assumption, when it was established and whether it still holds.
This is an estimate, not a measurement. The pressure test does not state with certainty that a task will be taken over by AI, nor does it guarantee that. It indicates where the likelihood of change is greatest, given what is currently known about the nature of the work and the movements in the sector. Where the underlying data is thin, or where it concerns work that is highly dependent on individual judgment, the outcome is more uncertain, and that is not disguised with a precise percentage.
An outcome also says nothing on its own, apart from the moment. An assumption that still holds today may no longer hold in a quarter. That is why the question of how often an assumption about AI should be reconfirmed is just as important as the first time it is established.
Some executive teams have already built this in: a fixed moment at which assumptions are reviewed, a way to record why a strategic choice was made at the time, and a rhythm for how often the strategy as a whole is recalibrated. For them, a shifted assumption is an agenda item, not a surprise.
Other companies lack that rhythm. The strategy was set, the assumptions were never named separately, and the only time anyone notices is when a result falls short and no one can explain why anymore. Often it is also a matter of who shouts loudest about AI within the executive team: how do you prevent the loudest voice from determining what AI takes over first is then more relevant than any technical question. And sometimes the picture within the executive team itself is not even shared: what do you do with an executive who does not share the AI picture is then the question that must be answered first, before an assumption list yields anything. A pilot with AI in a single team does not resolve that difference; see why an AI pilot is not yet a strategy.
The underlying question — which work in this specific company can genuinely be taken over by AI, which part remains under human oversight and which part remains human work — is answered per task by the FTE TO AI work scan.
Which assumption in your strategy has gone unchecked the longest is usually not hard to guess once you ask the question. Anyone who wants to organize this without immediately building a judgment on top of it can take the free assumption check: a short conversation in which you name the key assumptions underlying your strategy and see, per assumption, when it was last confirmed. The full pressure test, with the external and internal picture side by side, is under development.
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.