The installation sector consists of a combination that doesn't weigh the same everywhere: calculation and quotation processes, planning and scheduling of technicians, remote fault diagnosis, administrative processing of work orders and inspections, and the actual work on site -- installing, connecting, testing, delivering. That last part remains human work as long as it requires physical actions in a physical space. But the work around it -- the hours that go into planning, quoting, documenting and diagnosing -- is precisely the part where the nature of the work changes without the job title changing.
What drives the outcome is not the sector as a whole but the circumstances per company: how much standardized work there is compared to custom work, how well the historical data on faults and delivery times has been recorded, and whether oversight has been set up that can approve or reject an AI proposal with reason. Two installation companies with comparable revenue can therefore have a very different answer to the question of what is already shifting now.
AI work in the installation sector does not move in a single direction. There are tasks a system can handle independently, tasks where a system makes a proposal and a human assesses it, and tasks that remain unchanged human work.
This division shifts as well. What requires oversight today may lie closer to takeover a year from now as data and systems mature. That is not a prediction but an observation: the boundary moves, and a strategy that assumes a fixed boundary therefore quietly becomes outdated.
The difference between companies that already benefit from this and companies that do not rarely lies in ambition. It lies in three things: the quality of the underlying data, the setup of oversight, and the question of whether someone has recently tested the assumption underlying the strategy.
A company that has recorded fault data in a structured way for years can have a diagnosis proposal assessed alongside it faster than a company where that data is scattered across work orders and phone notes. A planning department accustomed to an inspection step can simply incorporate an AI proposal for the schedule into that step; a department without that step must first set it up before takeover or oversight gains meaning. That is not a matter of willingness, but of what is already in place.
The strategic pressure does not lie in the question of whether AI is taking over work -- that is already happening, differently per task and per company. The pressure lies in what a management team assumed a year or more ago about its own capacity: how many calculation hours are needed, how many planners a given region can handle, what the turnaround time of a quotation needs to be to remain competitive. Those assumptions were valid at the moment they were set. They do not expire with a decision; they expire because the world around them changes while no one revisits the assumption.
That expiration is exactly why a strategy that still holds up on paper can start to shift from the inside. The figures in the plan have not become wrong, but the assumption on which they rest has. What chafes there is not always visible in the reporting, but it is visible in the behavior of teams that take different routes than the plan prescribes.
This pattern is not unique to the installation sector. The same shift plays out in how takeover and oversight of AI work are putting pressure on strategy in healthcare, in the shift in planning and inventory work in wholesale due to AI and in how production and quality work in manufacturing is shifting due to AI takeover. Each time, it is not the sector that changes, but the assumption underlying the strategy that quietly expires.
This shift says nothing about what an employer should do regarding its personnel. If a decision on positions or staffing levels comes up, its own legal requirements and procedures apply, separate from what is described here. What is at issue here is: which assumption underlying the strategy rests on a situation that no longer exists, and who in the organization should actually be able to see that. That is also why recording why a strategic choice was once made yields so much: see how to record why a strategic choice was made, and why retesting it deserves a recurring rhythm rather than a one-time exercise, as described in how often a strategy needs realignment.
Whether a task in this company can truly be taken over by AI, supported with oversight, or remains human work, is precisely the question the work scan of FTE TO AI answers per task.
The strategy itself does not need to be revised immediately to know where the pressure lies. A 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 pressure test -- the outside view alongside the management team's own self-plot, with an assumption list that stays current -- 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.