Manufacturing, wholesale and transport all raised their AI use between 2023 and 2025, and Belgian and Dutch companies sit near the top of all three. Our analysis of Eurostat data shows what separates the leaders: they use AI inside the core process, on the production line, in the order flow and in transport planning, at three to five times the EU rate.
This piece is for the IT manager, CIO or informatiemanager at an organisation of 250 to 5,000 people who has to tell the board where AI goes next. Crux Digits builds AI systems alongside your existing ICT partner or IT team; we are not an ICT partner ourselves.
Which industries are pulling ahead with AI?
Of the three sectors in this piece, wholesale moved fastest, nearly tripling from 7.9% to 23.4% of EU companies between 2023 and 2025.
Manufacturing went from 6.8% to 17.3%. Transport and storage grew from 5.3% to 11.2%, the smallest rise of the twelve sectors in Eurostat table isoc_eb_ain2, and still a doubling. Information and communication remains the most AI-intensive sector at 62.5%, construction the least at 10.8%. No sector went backwards.
That is good news for the Low Countries. Belgium leads all of EU manufacturing at 39.8% of firms and ranks third in both wholesale and transport. The Netherlands is in the EU top six in nine of the eleven sectors where it has a ranked value. In all ten sectors where both countries report, Belgian companies use AI more often than German ones.
The headline rate hides the more useful number. Eurostat also asks what the AI is used for. On those process lines the leading country runs at three to five times the EU rate, and that is where the next twelve months are decided.
What do manufacturing leaders do that others can copy?
They put AI on the production line: Belgium leads AI in production processes at 10.2% of manufacturers, against 3.5% across the EU.
Only about one in five EU manufacturers that use AI at all use it in production. The rest use it in the office: a drafting assistant, a translation tool, a summary of the 8D report. Those are useful, and they are the easy part.
The leaders go a step further. In machine learning for data analysis, Belgian manufacturers are at 12.2% and Dutch ones at 7.4%, against 3.7% in the EU. Dutch machine builders rank second of 21 at 13.9%. In practice that is a model reading historian data and SCADA tags to flag a bearing before it fails, a vision check at the end of the line that books rejects back into the MES, or a scrap forecast per batch.
What to copy: pick one line where the data already sits in the historian or the MES, and build one model your process engineers trust. The Netherlands is sixth on production AI at 6.3%; Belgium shows what the next step looks like, next door. Getting the data out is a project of its own, covered in exporting OT data from the historian.
What do wholesale leaders do that others can copy?
They automate the order flow: Danish wholesalers use AI for workflow automation at 18.8%, against 5.6% across the EU.
Denmark leads wholesale overall at 54.9% of firms. Belgium is third at 41.1% and the Netherlands sixth at 34.5%. That makes wholesale the widest Dutch gap to the leader among the sectors where 25 or more countries report: 20.4 percentage points. On workflow automation, Belgian wholesalers are at 10.4% and Dutch ones at 8.7%.
The work behind those numbers is concrete. Orders arrive by e-mail, PDF and EDI and get retyped into the ERP. Supplier price lists are updated line by line. A pakbon has to be matched to the purchase invoice. Each is a narrow task with a clear right answer, the kind AI handles well while a person approves the exceptions.
As plain arithmetic, closing half the Dutch wholesale gap would put Dutch wholesalers at about 44.7%, above Belgium today. For a single company the step is smaller: one order flow running in production.
What do transport and logistics leaders do that others can copy?
They put AI into planning: Lithuania leads AI for logistics at 10.2% of transport firms, against 2.1% in the EU.
Transport is where the Low Countries pulled ahead fastest. Dutch transport and storage firms went from 11.0% in 2024 to 19.7% in 2025, now 1.8 times the EU rate. Belgium is third at 22.4%. On AI for logistics specifically, Belgium is fourth at 5.3% and the Netherlands fifth at 4.5%. For workflow automation, Danish firms are at 14.0%, Belgian 9.8% and Dutch 5.3%, against 2.8% in the EU.
In a carrier or a warehouse the leaders' list looks like this: route and load planning in the TMS, slot booking at the dock, ETA prediction for the customer desk, and reading CMRs and delivery notes into the WMS without retyping. Each sits next to a system the organisation already runs, so the integration with the TMS or WMS vendor is most of the project.
What should an IT manager do in the next twelve months?
Make five moves: benchmark the process line, use your size class, start at the widest gap, ship one system, and set the 2027 bar. Each ties to a finding in the study.
1. Benchmark on the process line, not the headline. Compare yourself with the leader on AI in production processes, AI for logistics or workflow automation. The headline rate counts a single pilot the same as a system in daily use; the process lines do not hide that difference as easily.
2. Use your size class as the peer group. In 2025, 55.0% of EU companies with 250 or more staff used AI, 67.6% in the Netherlands and 76.4% in Belgium, according to Eurostat table isoc_eb_ai. An organisation of 800 people at its sector average is behind its own size class.
3. Start where your sector's gap is widest. For a Dutch wholesaler that is the order flow into the ERP. For a manufacturer it is the production line, where Belgium is at 10.2% and the Netherlands at 6.3%. For a carrier it is planning and the dock.
4. Put one system into production through your normal change process. On the process lines, where the EU rate sits between about 2% and 6%, one system in production puts an organisation in the leading group. Run it through your ICT partner's change and release procedure, with documentation an auditor can follow, and own the code.
5. Set the board's bar with the two-year rule. In eight of twelve sectors the 2025 EU average had already passed the level the leading country reached in 2023. Plan as if today's leader level is roughly the average your competitors will reach by 2027.
What does 2027 look like if the trend holds?
If the 2023 to 2025 pattern repeats, today's leader levels become the European average around 2027: roughly 40% in manufacturing and 55% in wholesale.
That is the study's rule of thumb applied forward, not a forecast. Transport is one of the four sectors where the rule did not hold, because its 2023 leaders were further ahead, so read it there as a direction only.
Two cautions keep this honest. In 2025 Eurostat added generating pictures, video or sound to its list of AI technologies, used by 9.6% of EU companies, so part of the latest rise is a wider definition. And use is self-reported and binary. Neither changes the direction: all twelve sectors grew.
The positive reading for a CIO is that the methods putting Belgian manufacturers and Danish wholesalers ahead are not exotic. They are models on data the organisation already holds, inside systems it already runs. By 2027 the question will be less whether you use AI and more which process it runs in. The full country ranks per sector are in the sector study.
Where does Crux Digits fit?
Crux Digits builds one AI system for one core process, next to your ICT partner or IT team, and hands over code you own.
That can be a model on historian and MES data, an order intake flow into the ERP, or document reading into the TMS or WMS. We do not run infrastructure, hosting or support desks. Your ICT partner keeps that role, and we work through its change process.
How we approach this for organisations of your size: applied AI and machine learning for mid-sized businesses.
Last updated: 28 October 2026.
A consultant tells you where AI pays off; Crux Digits also builds it. A fixed price per step, one named expert, from Utrecht.
AI consultancy in the Netherlands →
