Every new release of business AI statistics tells roughly the same story. Adoption is climbing quickly, and it is climbing fastest among the largest companies. For the owner of a 20-person firm in Switzerland, the headline average is less useful than the gap underneath it, and what causes that gap.

What the numbers say

In 2025, one in five EU enterprises with ten or more employees used at least one AI technology, up from 13.5% a year earlier, according to Eurostat. That is a sharp rise, but the average hides the shape of it. Eurostat's detailed breakdown shows 17% of small firms (10–49 employees) using AI, against roughly 30% of medium-sized firms and 55% of large ones.

The OECD sees the same pattern across its members. Writing for Euronews in September 2026, OECD Secretary-General Mathias Cormann put it as one in two large firms using AI last year, against one in six small ones.

The gap is not only about whether a firm uses AI, but where. Among AI-using enterprises, Eurostat found large firms far more likely than small ones to apply it to ICT security (47.5% against 14.5%), business administration (43.4% against 28.5%) and production processes (33.5% against 19.0%). Small firms tend to use AI at the edges of the business. Large firms are pushing it into the core.

Where Switzerland sits

The Eurostat enterprise figures above cover EU member states, so they do not give a clean comparison for Swiss companies. The Swiss signals that do exist are mixed in an instructive way.

Swiss workers are among Europe's heaviest individual users. Eurostat data reported by Euronews put workplace use of generative AI in Switzerland at 34.4% of people aged 16 to 74 in 2025, just behind Norway and well above the EU figure of 15%.

At company level the picture is less settled. A Sotomo survey of 300 SMEs for the insurer AXA, reported by Swissinfo, found that 34% had deliberately built AI into their work processes in 2025, up from 22% the year before, while a further 37% were testing it. A larger UBS survey of around 2,500 companies, also covered by Swissinfo, found six in ten using AI. UBS economists added that few were doing so systematically, and almost a third of companies had no plans to start. In the same survey, smaller companies mostly used AI for analysis, while larger ones most often pointed to automating business processes.

Read together, these figures suggest that Swiss staff are using the tools, but many Swiss SMEs have not yet turned that individual habit into designed, repeatable processes.

Why the gap exists

The causes are not mysterious. Among EU enterprises that had considered AI but not adopted it, Eurostat found the most common obstacle was a lack of relevant skills, cited by about 71%. Unclear legal consequences (53%) and data protection concerns (49%) followed. Only about a fifth said AI was simply not useful to them.

That last figure matters. The gap is not mainly about belief in the technology. It is about capacity. Someone has to understand the tools, map the work, connect the systems and take responsibility for the data. A large company can assign a team to that. A small one usually relies on the owner's evenings.

The OECD's discussion paper for the G7 on SME adoption lists connectivity, data, skills and finance as the prerequisites. Most Swiss SMEs already have the connectivity. The shortfall usually sits in data that is scattered across systems, in skills, and in the time needed to apply both.

What a small firm can do about it

Closing the gap does not mean copying a large company's programme. It means using the advantages small firms do have: short decision lines and processes that one person can see from end to end.

  • Find out what staff already do. If employees use AI tools informally, learn which tasks and which tools. That is your real starting point, and a data protection question you need to answer in any case.
  • Choose one core process, not a peripheral one. The Eurostat data suggest small firms stay at the edges. Pick one workflow that touches customers or revenue, such as enquiry handling, quoting or order follow-up.
  • Name an owner. One person, with a few protected hours a week, accountable for the result and for the data flows.
  • Configure before you build. Most small firms will get further with the tools they already pay for than with custom development.
  • Write down the rules. Which tools are approved, which data may go into them, and who reviews the output. One page is enough to begin.
  • Measure one number. Hours spent, response time or error rate, recorded before and after.

The bottom line

The adoption gap between small and large firms is real, but it is mostly a gap in time, skills and structure rather than in usefulness. Swiss SMEs start from a reasonable position, because their staff are already comfortable with the tools. What remains is turning scattered individual use into one or two well-designed processes that the business owns and can measure.

Want to know which of your processes is the right first one? The free AI Readiness Audit maps your recurring work and ranks what is worth automating.

Sources

  • Eurostat, "20% of EU enterprises use AI technologies", 11 December 2025. link
  • Eurostat Statistics Explained, "Use of artificial intelligence in enterprises", data extracted December 2025. link
  • Euronews, "Europe cannot afford to miss AI transformation, with fiscal dividend within reach, OECD chief says", 21 September 2026. link
  • Euronews, "AI use at work in Europe: Which countries use generative AI tools most, and why?", 19 March 2026. link
  • Swissinfo, "Swiss SMEs ramp up adoption of AI in workplaces", 8 October 2025. link
  • Swissinfo, "Six out of ten Swiss companies use AI", 19 May 2026. link
  • OECD, "AI adoption by small and medium-sized enterprises", 9 December 2025. link