A strategic document rarely disappears because it is bad. It disappears because no one knows anymore which part of it still holds true. At the moment of writing, every assumption was fixed somewhere: the market moved this way, the competitor did not do that, this work cost that much capacity. A few quarters later, part of those assumptions has quietly expired, without anyone withdrawing them. The document itself does not change. The reality underneath it does. That gap is what a board feels as "the plan no longer lives", even before anyone can point to where it went wrong.
The most underestimated expiry date currently lies with AI. Not because AI was in the plan as a separate track, but because the plan silently assumed who does which work and how much time it takes. That assumption is precisely where AI is already breaking in today, in parts and not everywhere at the same pace. For part of the work, a task can now largely be taken over. For another part, that happens with human oversight that approves or rejects with reason. For a third part, it remains human work, for reasons that differ per task. A plan that last year calculated with a fixed number of office hours for a task that has now partly shifted, is quietly miscalculating, even if no one has ever adjusted that figure.
The difference between companies that track this and companies that do not rarely lies in the sector or size. It lies in whether there is a place where an assumption is periodically tested, and whether there is someone who is allowed to say that an assumption no longer holds without this being read as criticism of the original choice. Where that place is missing, an outdated assumption simply stays in the document, because no one has the task of revising it. Where that place exists, an expired assumption becomes an agenda item instead of a surprise afterwards. That difference is organizational, not technical: it is about who has which voice and when, something that how you prevent the loudest voice from determining what AI takes over first addresses.
A stress test places two images side by side. One comes from outside: sector data, regulatory clocks running out, public signals about what competitors are already doing. The other comes from inside: how the board itself assesses the situation, independently of each other, before anyone has heard the rest. Where these two images diverge, a list of assumptions emerges in which it is established, per assumption, when it was last confirmed. That is an estimate, not a statement with certainty. The estimate is less reliable the older the underlying data is, the less transparently a sector reports, and the more an assumption revolves around something the board itself knows better than any external source. A stress test that nonetheless suggests a hard figure on such points says nothing. The value lies in making visible where the image from outside and the image from inside diverge, not in an outcome that erases all uncertainty.
Boards often confuse these two, and that is precisely where plans start to shift. A risk is something you know can happen and for which you have a response ready. An assumption is something you no longer know you are assuming, until someone asks the question. AI taking over work usually falls into that second category: no one decided that the assumption about required capacity would expire, it just became so. What that distinction means in practice is worked out in what the difference is between an AI risk and an AI assumption. Whoever treats these two interchangeably ends up managing an assumption as a risk, with a response plan for something no one knows still holds true.
Some markets are currently moving faster than the cycle in which a board revises the plan. Then the question is not whether the strategy still held true on the day it was adopted, but whether it still holds true on the day you read it. That situation, and what can be done about it without rewriting the entire plan from scratch, is addressed in what you do when the market changes faster than your plan. Even a good board can differ on this: not everyone sees the same urgency, and what to do with a board member who does not share the view on AI is a separate question that is not solved with more persuasive power, as can be read at what you do with a board member who does not share the AI view.
This page is about whether an assumption in a strategic plan is still valid. It is not about whether, and in what way, an organization deploys personnel as a result of that outcome. That decision rests with the employer and has its own legal requirements; we deliver the factual picture with which an assumption can be tested, not the justification for a personnel decision.
The question of which work in your company can actually be taken over by AI cannot be answered at the strategic level alone; that question is answered per task with the work scan from FTE TO AI. If you would first like to know how you can test whether your strategy still holds before proceeding, that is a good starting point. A direct first step is 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 the image from outside alongside your board's self-plot, 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.