Many SMBs try AI. Few actually anchor it.

How many Dutch companies "use AI" depends entirely on what you measure, and that one fact explains why published numbers range from under 20 percent to well above 70 percent. They are not contradictory. They sit on a ladder, where each rung is a stricter definition of use. Read that way, the data tells a clear story: one-off experiments with tools like ChatGPT are now widespread, but AI anchored in core processes remains rare. The gap between the two is where the real opportunity sits.

A funnel, not a contradiction

At the wide end, Exact's MKB Barometer 2026, a self-reported survey of 1,719 SMB owners and managers (companies of 2 to 250 FTE, fieldwork Q1 2026), found that 70 percent say they use AI in some form, up from 33 percent in 2025 and just 6 percent in 2024. A step further, the Dutch Chamber of Commerce (KVK) reported in September 2025 that roughly three in ten entrepreneurs use ChatGPT or Microsoft Copilot, rising to 46 percent in professional services.

Sharpen the definition once more and the number collapses. Statistics Netherlands (CBS), which uses a strict definition of AI technology, found that 17 percent of companies with two or more employees used AI in 2025. That is still a doubling versus 8 percent in 2023. At the narrow end, ABN AMRO estimates that only 20 to 25 percent of SMBs with 10+ employees are actively experimenting with AI, and fewer than 10 percent apply it structurally in their core processes.

The lesson for anyone presenting these numbers: pick one definition and stick to it, or show the funnel explicitly. The CBS numbers form the most authoritative baseline (national statistics office, official methodology); the Exact numbers are the most recent and most SMB-specific, but rest on self-reporting and a broader definition. Both are correct, but about different things.

The bigger the company, the more AI

The clearest and most consistent pattern in the data is that adoption rises sharply with company size. Based on the 2025 CBS numbers, AI usage climbs from 13.8 percent among micro-businesses (2 to 9 employees) to 27 percent (10 to 49), 45 percent (50 to 249), and 66 percent at large companies with 250+ employees. A large company is roughly five times more likely to be an AI user than a micro-business. ING found the same gap within a single sector: in manufacturing, 64 percent of large companies in 2025 used at least one AI application, versus only 12 percent of small companies (10 to 50 employees).

For international context: Eurostat put AI usage in the EU in 2025 at 20 percent (companies with 10+ employees), with the Netherlands among the higher-adopting member states, above the EU average. The Eurostat threshold of 10+ employees is not directly comparable to the CBS headline of 2+ employees, so the two percentages should not be placed side by side without that caveat.

ICT leads, hospitality trails

Adoption varies enormously by sector. The 2025 CBS data puts information and communication (ICT) far in front at 54 percent, followed by specialised business services (31 percent) and financial services (28 percent, the sharpest jump of the year). Real estate (21 percent), energy and water (18 percent), manufacturing, healthcare, and other services (around 16 percent), and trade (12 percent) sit in the middle. At the bottom: transport and storage and construction (each 7 percent) and hospitality (6 percent, the lowest of all sectors). The ICT lead is heavily driven by generative AI specifically, such as text and speech generation.

What holds the rest back

The dominant barrier, by a wide margin, is a lack of knowledge and experience. Among companies that considered AI but chose not to, CBS found 73 percent cited lack of experience, 49 percent privacy concerns, and 42 percent legal or liability uncertainty. Only 15 percent said AI simply wasn't "useful." For micro-businesses the pattern is nearly identical (lack of experience 71.6 percent, privacy 48.8 percent), and notably, cost is a relatively small and declining barrier (around 20 percent).

Government-commissioned research by Dialogic for the Ministry of Economic Affairs (November 2025) adds the qualitative core: the real block is a lack of vision and strategy around AI, compounded by limited skills, weak data foundations, low existing digitalisation, and for smaller companies, simply the absence of the time and people that larger companies can free up. Regulatory uncertainty is also unresolved: KVK found that half of entrepreneurs don't know the European AI Act exists, and only 7 percent feel well informed. The primary driver, by contrast, is efficiency: automating repetitive work, cost savings, and absorbing persistent labour shortages.

What SMBs do use AI for

Among companies that do use AI, the most common applications in 2025 were marketing and sales (35 percent) and administrative and management tasks like planning, translation, and candidate screening (32 percent), followed by R&D and innovation (25 percent), production and service processes (18 percent), finance and accounting (17 percent), IT security (11 percent), and logistics (4 percent). The most-used underlying technologies are text-focused: text mining for summarising and analysing text (12 percent of all companies) and natural language generation (8 percent).

Exact's 2026 survey points the same way, with writing text and generating images as the most reported use (56 percent), followed by meeting notes and translations (37 percent) and insight into business performance (36 percent). On specific tools, the clearest signals come from KVK and online bank Knab: KVK's 30 percent for ChatGPT and Copilot is the best institutional reading of usage per named tool. Knab's 2025 study of more than 2,000 freelancers found that 51 percent use AI tools for work (up from 38 percent a year earlier), 94 percent of whom use ChatGPT.

The bottom line

AI adoption in the Dutch SMB sector is real and accelerating, but it remains shallow. A large and growing share have tried generative AI, a smaller share use it regularly, and only a small minority have woven it into how the business actually runs. The dividing line runs along company size and sector, and the binding constraint is not cost but know-how, vision, and trust. Closing that gap, rather than simply running more first trials, is the prize.

A note on the numbers

All percentages are dated and attributed to their original source. CBS figures for 2025 are provisional and cover companies with 2+ employees unless otherwise noted. Exact figures are self-reported and use a broader definition of AI use than CBS, which is why they are substantially higher. Eurostat uses a 10+ employee threshold. Various widely shared "2026" statistics from commercial and marketing blogs could not be verified against primary sources and have been deliberately excluded.

1. 01Exact, MKB Barometer 2026 (research by Markteffect, n=1,719, companies 2-250 FTE, Q1 2026).
2. 02KVK, Research: entrepreneurs taking few steps around European AI legislation, 11 September 2025 (KVK Omnibus, n=651).
3. 03CBS, Companies use AI most often for marketing or sales, 12 December 2025.
4. 04CBS, Use of AI technology by Dutch micro-businesses, 17 March 2026.
5. 05ABN AMRO, sector research and AI insights, 2025-2026.
6. 06ING Research, thematic study on AI in manufacturing, February 2026.
7. 07Eurostat, 20% of EU enterprises used AI technologies in 2025, 11 December 2025.
8. 08Dialogic / Ministry of Economic Affairs, Research on AI use in SMBs: ambition or hesitation?, November 2025.
9. 09Knab, Research: AI use among freelancers is rising rapidly, 2025 (n>2,000 self-employed).