How Swedish companies use AI – what the official numbers actually show

Statistics Sweden (SCB) has published the first official measurement of how many Swedish companies actually use AI – not individual employees, but companies as organisations. The numbers give a firmer footing than the employee-survey data on AI habits that has circulated recently, and they paint a more nuanced picture than the "AI is everywhere" headline suggests.

One in three companies, a ten-point jump

According to SCB, 35 percent of companies in Sweden used AI during 2025, an increase of ten percentage points from 2024. That is a fast move for a single year, and it places Sweden among the highest-ranking countries in Europe – well above the EU average of 20 percent.

Within the Nordics, usage ranged from just under 29 percent in Norway to 42 percent in Denmark, with Sweden at 35 percent in between. The gap between neighbouring countries is smaller than you might expect – it's within Sweden, between different types of companies, that the real differences show up.

The size gap is the real story

The figure that says the most isn't Sweden's overall number, but which companies sit behind the average. Among large companies, 250 employees or more, 72 percent used AI during 2025. Among medium-sized companies it was 50 percent. Among small companies, only 31 percent.

That means "one in three companies uses AI" is, to a large extent, a figure carried by a minority of large companies. Most Swedish companies are small, and among them, still fewer than one in three use AI at all. The sector gap points the same way: 89.6 percent of companies in IT and communications use AI, against 12.3 percent in transport and storage. Adoption is concentrated – in large companies, and in sectors where software is already part of daily work.

Why the gap exists – and why it doesn't have to persist

The explanation is rarely reluctance. Large companies have their own IT departments, data already sitting in established systems, and budget to experiment without any single effort having to carry all the risk. A smaller company usually has none of that – AI gets attention when there's time for it, if it gets any at all.

But size is not the same as readiness. A well-run workflow doesn't require a data department or a company-wide initiative. What usually decides the outcome is how well-defined the task is and how clean the underlying data is, not how many employees stand behind it. A company with ten people can build a narrow, reliable workflow around a single recurring task – invoice handling, customer support, quote preparation – just as well as a company with a thousand, if it's done with the same discipline: a clear scope, an owner, and a way to measure whether it actually saved time.

Where Nodal comes in

We work mostly with the kind of company SCB's numbers show is lagging – smaller, often family- or owner-run businesses without a dedicated IT or data team. The size gap isn't a sign that AI fits less well there. It's a sign that fewer of them have had the right help to get started at a small, manageable scale.

If you recognise your company in that part of the statistics, we're glad to talk about where a first, well-scoped AI workflow would do the most good for you. Get in touch with us at Nodal.