The Signal · Column 048

A restructuring is really a knowledge decision

Andrei van Noordt · 5 September 2026 · 2 min read

D'Ieteren Automotive is reorganising, and the restructuring puts almost 350 jobs in Belgium at risk. There is no artificial intelligence anywhere in that headline, which is precisely why I am writing about it. This is what most of the next two years in the Benelux labour market will actually look like. Not a machine taking a job. A reorganisation, a consultation process, and several hundred people carrying their working knowledge of an industry out of the building inside the same quarter.

The reading most people miss is that a restructuring sorts by cost and by function. It does not sort by knowledge. An org chart records what someone is called and what they are paid. It does not record who knows why a process was built the way it was, whose supplier estimates are reliably optimistic, or which exception rule exists because of a mistake nobody ever wrote down. That knowledge leaves in the same envelope as the salary, and only one of the two appears in the savings calculation.

A reduction is not a cost decision. It is a decision about which knowledge your company keeps, taken by people who cannot see it on the org chart.

That matters more than it used to. Any company putting AI into its operations needs exactly this kind of undocumented, sector-specific judgment, because it is the raw material the tooling has to be trained against and corrected by. Access to good models is available to everyone now. Knowing what good looks like inside automotive distribution is not. When a few hundred people holding that second thing come loose in a small market at once, the companies hiring get a short window at a price they will not be offered twice.

I would treat that window as genuinely narrow. In Belgium and the Netherlands the pool of people with deep operational experience in one sector plus enough data literacy to work alongside modern tooling is thin, and clusters like this get absorbed fast by firms that were already looking. The organisations doing the reducing tend to come back for the same profiles later, at a rate they no longer set.

For hiring leaders, the practical read is simple. Before you sign off on any reduction, write down the people whose judgment your systems would have to learn from if you automated their work tomorrow, and take that list out of scope. It will be shorter than you expect, and it will not line up neatly with your seniority grades.