Educational institutions run on a mix of work that doesn't summarize well in a single term. There is preparatory work: compiling teaching materials, drafting tests, updating schedules and planning. There is assessment work: grading, giving feedback, monitoring progress. There is administrative work around enrollments, absence registration, subsidy accountability and accountability to the inspectorate or the board. And there is the work that no one puts on paper but that still takes up most of the hours: the face-to-face contact with pupils, students and parents, in which explanation, correction and guidance come together.
These four types of work are unevenly sensitive to change. Preparatory work and administrative work largely consist of repeatable steps with a fixed structure: drafting a test according to a framework, optimizing a schedule within rules, filling in a subsidy application according to a format. Assessment work is in between: recognizing the pattern in an answer can be done mechanically, but the weight of an assessment for an individual pupil requires someone who knows the context. Direct contact remains largely human work, not because it cannot be technically supported, but because its value lies largely in the relationship.
In some institutions, teaching materials and test materials are now largely drawn up by AI, with a teacher checking and adjusting the result before it reaches the classroom. Grading structured assignments — multiple choice, short answers, standardized rubrics — is in some schools already largely automated, with a teacher assessing deviating cases. Timetabling and capacity planning are shifting where the underlying data is in order; where that data is scattered or outdated, manual work remains.
What does not shift, or shifts much more slowly, is the work in which the outcome depends on who stands in front of the class: mentoring conversations, guidance of a pupil with a specific support need, the judgment call in a borderline case in an assessment that determines a follow-up track. There, oversight is not limited to approving or rejecting with reason — there, the initiative itself is human work.
The difference between institutions that are already benefiting from this shift and institutions where everything remains as before rarely lies in ambition. It lies in three conditions: whether the underlying work process has already been broken down into recognizable, repeatable steps; whether there is someone who actually checks the output instead of adopting it blindly; and whether the board and the participation council have already had this conversation or are still avoiding it. Institutions where these three are in order see the shift in practice. Institutions where that is not the case are still talking about it as something that is yet to come.
A multi-year strategy for an educational institution often rests on assumptions about staffing needs, about where investments of time and money yield the most, and about what capacity a teacher, dean or staff member will need in the coming year. Those assumptions are usually correct at the moment they are established. The problem is not that they were wrong, but that no one marks the moment at which they stop being right.
If grading tests becomes largely automated with teacher oversight, the question of how much staff capacity is needed for testing periods changes — an assumption that was still stated differently in last year's staffing plan. If teaching materials are largely compiled and checked by AI, what constitutes good preparation time changes — an assumption that was still adopted unchanged in the collective labor agreement negotiation or the workload assessment. The assumption is still on paper; the reality behind it has already shifted. What an institution does with that — whether and how that affects staffing — is up to the board and falls under its own legal and collective labor agreement requirements; that is not a choice that is prescribed or substantiated here.
The underlying mechanisms are not unique to education. Retail and the IT sector also see work shifting into the same three categories: fully transferable, partly under supervision, and largely human work. What differs is which work falls into which category and how quickly that changes. For a management team that wants to know whether its own assumptions have already absorbed that difference, it is useful to first get a clear picture of what a strategic assumption actually is and how you test whether it still holds, and to examine why the perceptions within a single management team about the same development often diverge. These two questions form the basis of the pressure test: an outside view placed alongside the board's self-image, with an answer for each assumption as to whether it still holds.
The question of which work at this institution can genuinely be taken over by AI cannot be answered with an impression or a single example from the staff room conversation. The work scan from FTE TO AI answers that question per task, with a rationale indicating whether the task is transferable, falls under supervision, or remains human work.
Anyone who wants to know now which assumptions in their own strategy are due for replacement can start with the free assumption check: a short round in which you name your key assumptions and see for each one when it was last confirmed. The full pressure test, with the external view placed alongside the board's self-image, 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.