Short answer. In 2025, 33 percent of Dutch companies with ten or more employees used AI technology. Among companies with 100 to 250 employees it was 52 percent, among companies with 500 or more 73 percent. Half of the companies with ten or more employees use ERP software, and 8 percent use robotic process automation. Source: Statistics Netherlands (CBS), ICT use by enterprises 2025.
On this page I collect the numbers I use myself when I talk to companies about automation. Primary sources only, with the source next to each figure, and broken down by company size. That last part matters. The national average hides how big the difference is between a company of fifteen people and one of two hundred.
Most figures come from the CBS StatLine table ICT use by enterprises 2025 (CBS StatLine, 2025). Percentages are always the share of all companies in that size class. I update the page as soon as new figures are published.
AI use by company size
| Company size (employees) | Uses AI (2025) |
|---|---|
| 10 or more (total) | 33% |
| 10 to 20 | 25% |
| 20 to 50 | 33% |
| 50 to 100 | 43% |
| 100 to 250 | 52% |
| 250 to 500 | 61% |
| 500 or more | 73% |
Source: CBS StatLine, ICT use by enterprises; company size, 2025.
The difference between small and large is almost a factor of three, and AI use rises by eight to ten percentage points with each size class. There is no sudden threshold. There is one in the business software underneath, further down this page.
Growth is fast. According to CBS, AI use among companies with 50 to 250 employees rose from 20 percent in 2023 to 45 percent in 2025. Across all companies with two or more employees, including the smallest, it was 17 percent (CBS, 2025). That lower number explains why you see different percentages in the media: it depends on which companies are counted.
What companies use AI for
| AI used for | 10 or more | 50 to 100 | 100 to 250 |
|---|---|---|---|
| Administrative processes or management tasks | 14% | 18% | 25% |
| Marketing or sales | 12% | 15% | 18% |
| R&D or innovation | 9% | 11% | 15% |
| Accounting, controlling or financial management | 8% | 11% | 13% |
| Production or service processes | 7% | 9% | 13% |
| ICT security | 7% | 10% | 17% |
| Logistics | 2% | 3% | 4% |
Source: CBS StatLine 2025. Share of all companies in the size class.
Administration and marketing lead. In logistics and production, where processes are most fixed and costs highest, AI is still rare. That is where I think the most room is, and also the most risk if you start without a mapped process. Where AI does and doesn't work is in automating business processes with AI.
How companies get their AI software
| AI software | 10 or more | 100 to 250 |
|---|---|---|
| Commercial software used as is | 19% | 34% |
| Commercial software modified in-house | 9% | 16% |
| Open-source software modified in-house | 8% | 14% |
| Developed by external suppliers | 9% | 17% |
| Developed by own employees | 6% | 12% |
Source: CBS StatLine 2025. Companies can give multiple answers.
Most companies use AI as it comes out of the box. Modifying it or having it built happens much less often, but grows with company size. Among companies with 100 to 250 employees, 17 percent have AI software developed by an external party.
Why companies don't use AI (yet)
Ten percent of companies with ten or more employees considered AI without starting to use it. The reasons they give:
| Reason for not using AI | Share of all companies (10 or more) |
|---|---|
| Lack of relevant expertise in the company | 8% |
| Concerns about data protection | 5% |
| Existing ICT systems are not compatible | 4% |
| Difficulties with data availability | 4% |
| Unclear legal consequences | 4% |
| Costs are too high | 2% |
| Ethical considerations | 2% |
Source: CBS StatLine 2025. Multiple reasons possible.
Cost is at the bottom. What holds companies back is expertise, followed directly by things to do with the process and the systems: ICT that doesn't connect and data that isn't readily available. I see the same in practice. AI rarely gets stuck on the model. It gets stuck on the data and the systems around it.
ERP, CRM and BI: which business software companies use
| Software | 10 or more | 10 to 20 | 50 to 100 | 100 to 250 | 500 or more |
|---|---|---|---|---|---|
| Business software (total) | 69% | 57% | 85% | 92% | 97% |
| ERP software | 50% | 36% | 69% | 78% | 85% |
| CRM software | 54% | 45% | 66% | 73% | 79% |
| BI software | 36% | 21% | 57% | 74% | 94% |
Source: CBS StatLine 2025.
There is a clear jump around 50 employees. Of companies with 20 to 50 employees, 54 percent use ERP software; at 50 to 100 it is 69 percent. For BI software it goes from 37 to 57 percent. Still, almost a third of companies with 50 to 100 employees have no ERP system. That doesn't mean they have no administration. It usually means orders, stock and finance live in separate packages and spreadsheets held together by hand. That is the situation I describe in a patchwork of tools.
Process automation and data exchange
| 10 or more | 50 to 100 | 100 to 250 | 500 or more | |
|---|---|---|---|---|
| Robotic process automation (RPA) | 8% | 12% | 16% | 35% |
| Sales via EDI (orders received electronically, 2024) | 7% | 13% | 19% | 19% |
Source: CBS StatLine 2025.
Structured, automatic exchange of orders between companies is still the exception in the Netherlands. Among companies with 100 to 250 employees, 19 percent receive orders via EDI. For the rest, orders arrive some other way, in practice often by email, through a portal or by phone. That is where the re-typing starts. Work out what that costs with the calculator: what does manual re-typing cost you per year?. See also re-typing between systems.
Other numbers on processes that don't run smoothly
- An average of 35 percent of all b2b invoices in the Netherlands are paid late; the average payment term is 45 days (Atradius Payment Practices Barometer, 2025). Part of that delay starts with invoices that go out late themselves; see invoices and claims sent too late.
- In the field audits with the best methods, at least 86 percent of the business spreadsheets examined contained errors (Panko, 2008). See dependent on Excel and one person.
- In Deloitte's global survey, process fragmentation and a lack of IT readiness topped the reasons companies could not scale automation (Deloitte, 2020).
What these numbers say about mid-sized companies
The numbers show a tipping point around 50 employees. That is where the use of ERP and BI software rises sharply and AI use keeps climbing, while structured connections between companies and systems lag behind. Companies buy software, but the work between that software remains manual. And when they consider AI but don't adopt it, they cite expertise, systems and data before cost.
That matches my experience. The question is rarely which tool to buy. The question is how the work between your tools runs, and which part of it can be done smarter. That is why I start with the process; see process optimisation: map the process first, then the software that fits.
About these numbers
The CBS figures come from the annual ICT survey among companies with ten or more employees, published on 12 December 2025 (StatLine table 86118NED, reporting year 2025; EDI figures for 2024). All percentages are the share of all companies in the stated size class, rounded to whole percentages. The CBS news release on 2025 uses a broader group (companies with two or more employees) and divides applications over the companies that use AI; that is why those percentages differ from the tables above. I don't use figures whose original source I could not check myself.
Frequently asked questions
How many companies in the Netherlands use AI?
In 2025, 33 percent of Dutch companies with ten or more employees used AI technology, according to CBS. Among companies with 100 to 250 employees it was 52 percent, among companies with 500 or more 73 percent. Including the smallest companies (two or more employees), it was 17 percent.
How do companies use AI?
Most often for administrative processes and management tasks (14 percent of companies with ten or more employees) and marketing or sales (12 percent). In production, finance and logistics use is still small. Most companies use commercial AI software as is; 9 percent modify commercial software themselves.
Why don't companies use AI?
Companies that considered AI but don't use it most often cite a lack of relevant expertise (8 percent of all companies with ten or more employees), followed by data protection concerns (5 percent) and ICT systems that aren't compatible (4 percent). Too high costs is cited by 2 percent. I'm Ricardo Theijs of RNT Projects. I come from process optimisation (UWV, Centric, G4S, MSc Business Process Management) and now build the systems and AI workflows where standard packages fall short myself. I keep this page up to date with the latest figures.
Running into this yourself?
I review your process and build the solution where a standard package falls short. Remote, with visible results in two weeks.
Let's talk