The thesis
Augmenting human judgement through the system.
What I argue, in a single page. It is the thread running through my ERP work, my acquisitions and my writing — and the reason I believe reindustrialisation will be decided on the shop floor rather than in grand plans.
The starting point
Industry 4.0 is not a story about software.
The “4.0” label was sold as a catalogue of tools: sensors, ERP, MES, digital twins, and now AI. That framing is a mistake. None of these tools has any value in itself. Their only function is to improve one thing: the quality and the speed of decisions. Industry 4.0 is, at bottom, a discipline of decision-making.
A company is not a collection of machines, it is a loop: information → decision → action → measurement → learning. When that loop runs badly — data that gets ignored, intuition where measurement already exists, know-how that retires without being passed on — no technology saves the organisation. When it runs true, ordinary people obtain extraordinary results.
What AI reveals
AI is not a threat. It is a revealer.
I refuse both easy postures: catastrophism and solutionism. AI is a new layer of abstraction, in the lineage of the previous ones — machines, electricity, computing, the internet. After the automation of the gesture comes the automation of reasoning. The real question is not “will it replace us?”, but: where will the value that only a human can create sit tomorrow?
When an AI reaches 90% accuracy on a decision, that is less a feat of the machine than a confession about our organisations: they had already standardised that judgement to the point where a procedure could imitate it. AI does not defeat human competence — it reveals where we had already hollowed it out. That is excellent news, on one condition: that we decide what we want it to amplify.
The goal is not to automate in order to remove the human. It is to remove the tasks without value, so as to give the human back their judgement on the ones that count.
What it means for industry
The scarce resource is no longer execution. It is systems thinking.
If execution becomes a commodity, then value moves up towards those who know how to design and govern systems: connecting the shop floor to the data, the data to the decision, the decision to the result. It is the people formed by industrial management — the field, lean, supply chain — who gain in value, not the reverse. Technology is a means; the end is to make companies more solid, and technicians freer in their craft.
Why this matters for the country
Reindustrialisation will come from the shop floor.
French industry has been impoverished; for a long time its survivors could not invest. It will not be reborn by decree, nor through a handful of subsidised gigafactories. It will be reborn through the successful handover of thousands of small and mid-sized firms — those solid workshops whose owners retire this decade — and through sober technology, in the service of real performance: a system that standardises without dispossessing, that measures without surveilling, that trains a junior in weeks rather than years.
That is exactly what I am building. One proof at a time.
This text is deliberately compressed. I unfold it, subject by subject, in my writings — always starting from the field and working back up to the idea.