where dutch smbs really stand

In 2025, 33 percent of Dutch companies with ten or more people used at least one AI technology. In 2024 it was 23 percent. In 2023, 14 percent. That’s Statistics Netherlands doing the counting (CBS, 2026). Sounds like a boom. And then you ask how deep that use actually goes, and every measurement comes back way lower. Software company Exact went and asked: only 6 percent of small business owners say AI is a real, built-in part of how they run the place (Exact, 2025). So “using AI” keeps becoming an easier and easier bar to clear. Actually embedding it? Still rare. And if you want to know where Dutch small business really stands, you’ve got to keep those two curves apart. Don’t let anyone glue them together.

how many dutch companies use ai?

A third. That’s the short answer. In its report Digitalisering en kenniseconomie 2025, the yearly one on digitalization and the knowledge economy, CBS found that 33 percent of companies with ten or more people used at least one of eight AI technologies it asked about (CBS, 2026). A year before, 22.7 percent. The year before that, 14.0 percent (CBS, 2025). So it tripled in two years. In official statistics, mind you. You don’t see that pace often.

Now, behind that one percentage sit three completely different worlds. In early 2026 CBS put out a separate report on micro companies, the two-to-nine-people outfits, for the first time. There, AI use in 2025 came in at 13.8 percent. Ten to 249 people? 29.8 percent. From 250 people up? 66.2 percent (CBS, 2026). And even inside that ten-plus group the gap is huge: in 2024, 17.8 percent of the ten-to-nineteen-people companies used AI, against 59.2 percent of the ones with five hundred or more (CBS, 2025). So every time you see an adoption figure, check the size class first. Otherwise you’re comparing a five-man installation firm with a bank.

Zoom out to Europe and the Netherlands is doing fine. Eurostat clocked an EU average of 20.0 percent for 2025 among companies with ten or more people, against 13.5 percent in 2024 and 7.7 percent back in 2021. Among the big European companies it was 55 percent (Eurostat, 2025). Over 2024 the Netherlands came sixth in the EU, behind Denmark among others; the 2025 ranking isn’t confirmed yet. And here’s a figure that quietly flips how the whole thing looks: of all the Dutch companies using AI in 2024, 65 percent were small ones, ten to fifty people (CBS, 2025). So in raw numbers, most AI users are small companies. Even though the percentage inside the big-company group runs a lot higher. That same CBS piece, to its credit, hangs a warning on the link between AI use and company size: you can’t tell from these figures whether AI users are bigger and stronger on revenue because of AI. Correlation. Not proof.

the bar to count as an ai user drops every year

Underneath the CBS number sits one survey question: does your company use at least one of eight AI technologies, things like text generation, image recognition or speech processing? One is enough. So the shop where somebody knocks out quotes with a chatbot lands in the exact same bucket as the one running its whole planning on AI. The statistic tells you AI walked in the door. What it’s actually doing in there? That it barely tells you.

Ask about depth and the number changes completely. Exact runs a kind of ladder in its MKB Barometer, its recurring survey of Dutch small business: 43 percent of small businesses are actively poking at the possibilities, 23 percent are working on implementation, and 6 percent have made AI a real part of the business (Exact, 2025). Now, that’s commercial market research from a software vendor, with its own definition and its own sample, so take it as a hint and never as official statistics. But the direction lines up with what the government finds in the qualitative work: research agency Dialogic, commissioned by the Dutch Ministry of Economic Affairs, found that AI in small business is rarely embedded in any structural way and usually stalls at loose bits like text and image generation (Dialogic, 2025).

And this is where that tired line, “small business is lagging,” needs a correction. On the experimenting bit, Dutch small business hardly lags at all. A third of the ten-plus companies now does something with AI, and the growth curve is steeper than the EU average. The trouble sits one floor down: the step from loose use to actual process. And that step is hard everywhere, even at outfits with their own innovation department. Same echo abroad, honestly: MIT researchers reported in 2025 that 95 percent of generative AI pilots deliver no measurable business result, though that one covers mostly large, international organizations and says nothing specific about Dutch small business (Fortune, 2025). Why that number needs reading with care, I get into over in the 95 percent myth. For here, the lesson’s simpler: the small-versus-large gap lives in the adoption figures, far less in the embedding.

why do the ai numbers contradict each other?

Because definitions, populations and years fall out along the way. Here’s a real one: Exact’s 6 percent “integrally embedded” got read by ictmagazine.nl as general AI use in 2024, and from there you could claim a leap “from 6 percent to 70 percent now”. Two measurements that don’t belong next to each other, glued into one spectacular growth story, which then gets quoted by the next person, and the next. That’s how a number starts to echo. Claims like “67 percent SMB adoption” and “95 percent of Dutch organizations run AI programs” do the rounds too. Vendor or content-marketing surveys, every single one, each with its own definition, and you can never lay them beside the CBS figures as if they counted the same thing.

And even inside CBS you have to watch which table you grab. In the autumn of 2025, two publications on AI use per sector came out that look, at first glance, like they contradict each other. The September news release, on companies with ten or more people, puts construction at 8.9 percent in 2024 and hospitality at 9.0 percent (CBS, 2025). A December one, with a broader set of companies and a different question, lands at roughly 4 and 3 percent for that same year (CBS, 2025). Both right. Each within its own definition. The ranking holds up fine, though: information and communication way out front, construction, hospitality and transport bringing up the rear. And for the niche trades a lot of my readers are in, installation, cleaning, events, there just isn’t an official adoption figure. The SBI-sector numbers, that’s the Dutch standard industry classification, are too coarse for it, so I’m not going to sit here and extrapolate them either.

Three questions save you from most of the accidents. Which companies does it actually cover, and from which size class up? What exactly counts as AI use here? And who paid for the research? Run those three over anything you read, and most of the mess sorts itself out.

what holds smbs back on ai?

Knowledge and experience. Way more often than money or distrust. Of the Dutch companies that looked at AI and then said no, 74.6 percent gave lack of experience as the reason, across every size class and sector (CBS, 2025). Among micro companies it’s 71.6 percent (CBS, 2026). And across the EU, Eurostat lands in nearly the same spot: 71 percent point to a lack of relevant expertise as the main barrier (Eurostat, 2025). Three separate measurements this close together? That makes it one of the best-backed facts in the whole file.

The OECD sees the exact same thing outside Europe: across six G7 countries, more than half of small and medium-sized companies name lack of knowledge about generative AI as the biggest barrier, and of the SMBs already using generative AI, fewer than 30 percent offer any training for it (OECD, 2025). For the Netherlands, Dialogic throws in costs, how digitalized the organization already is, and whether the data is even available (Dialogic, 2025). Same picture every time: the barrier is learnable. And honestly, that’s good news. A learnable barrier is something you can actually go and work on. “Too expensive” or “our customers don’t want it”, now those are a lot harder to shift.

what should a small business do with these numbers?

One thing. Pick a single process and embed AI in it properly, with everything that comes with that. The adoption statistic isn’t a race you’re losing. As a micro company, any AI use at all already drops you into a minority of 13.8 percent, and structural embedding is rare everywhere you look, on every measurement going.

What you really don’t need to do: write a thirty-page AI strategy, try ten tools at once, or panic at headlines about adoption exploding. Those headlines are measuring the low bar. And two thirds of the companies in your size class are still sitting at zero.

What does work is a lot more boring, and it’s the order I actually run myself.

First, pick one process that comes back often and that you can measure: drafting quotes, answering customer questions, planning, chasing invoices. Then, before you touch a thing, measure the baseline for two weeks. How long does it take, where does it go wrong, what quietly never gets done. Only then do you bring in one tool, and you describe the process around it: who checks the output, what happens when it’s wrong, where the customer data lives and whether it’s even allowed to live there.

Now the step I care about most, and it’s the one everybody skips: train the people who actually work with it. Lack of experience is the biggest barrier in every single measurement, and training is exactly what happens at fewer than 30 percent of users (OECD, 2025). That gap? That’s your head start. Sitting right there.

After six to eight weeks, hold it against your baseline, in hours and euros. And only when the thing runs without you standing next to it do you go and pick up the next process. Not before.

Get through that and you’re already further along than almost every “AI user” in the statistics. That’s not me overpromising. It’s just what the numbers on the gap between use and embedding literally say.

Where I land: the adoption percentage is the least interesting number in this whole file. All it tells you is whether you cleared the lowest bar, and that bar keeps dropping every year. The question that actually matters is whether there’s one process in your company that demonstrably runs better since AI became part of it, with someone who genuinely owns it. In my own work for small service companies I see exactly what Dialogic describes, over and over: the chatbot is already there. The process around it? Missing.