The Signal · Column 097

The Dutch State Is About to Relearn Ford's Lesson

Andrei van Noordt · 28 October 2026 · 2 min read

The Dutch government has announced plans to cut 25,000 civil service jobs as part of an AI push, framing it as modernisation rather than austerity. The announcement reads like efficiency. The execution will read like something else entirely.

Most coverage of this kind of cut treats it as a headcount story: fewer desks, smaller budgets, a leaner state. The more interesting story is what has to grow underneath it for the system to actually work. You do not replace tens of thousands of people who process permits, benefits, and regulatory filings without building something that understands the logic they held in their heads. Someone has to translate decades of bureaucratic judgment into rules a model can follow, and someone has to notice when the model gets it wrong in ways that matter to a citizen's life.

You do not automate a function away. You automate the routine parts and concentrate the judgment in fewer hands.

That translation work is scarce and it is not interchangeable with the roles being cut. A government does not need fewer people who understand how the system works. It needs a much smaller number of people who understand it exceptionally well, paired with the technical staff who can encode that understanding and keep auditing it. The jobs disappearing and the jobs that matter next are not the same jobs, and treating the cut as pure subtraction is how public-sector AI programmes quietly fail two years in.

For talent markets, this repricing shows up first in government contracting and public-sector tech teams, but it will not stay there. Any large organisation doing the same calculus, trimming routine roles while leaning on AI, will discover the same gap: a shortage of people who can sit between domain expertise and model behaviour, catching the failures before citizens or customers do. Those people become harder to find and more expensive precisely because the headcount around them is shrinking.

For hiring leaders, the practical read is simple. If you are planning AI-driven headcount reductions, identify and retain your translation layer, the people who understand both the old process and the new system, before you finalise the cuts, not after the complaints start arriving.