In the media

AI is used in four ways in Nordic companies

By Sverre Stenseng

Computerworld

24 June 2025

Artificial intelligence is used by most companies in the Nordics, but according to PA Consulting, most have not yet reached the point where AI is used as a strategic resource.

PA Consulting recently carried out an AI survey across the Nordics. In it, 1,500 people from Norway, Sweden, and Denmark responded to questions about how AI is used in their organisations.

Based on the answers, PA Consulting believes it has identified four different ways AI is being used what they term four ‘AI archetypes’, which can also be seen as stages.

The first stage is familiar to everyone. This is the use of general task assistants, such as ChatGPT and Microsoft Copilot.

Some also use what they call knowledge agents these are AI agents trained on the company’s own data.

More advanced

Among the more advanced users, AI is also used in decision support systems. Here, larger analyses are carried out ones that cannot be done manually in order to generate the best possible proposals for solutions.

At the fourth stage, some use autonomous agents in other words, AI agents. These are virtual or physical robots that automatically carry out tasks without human intervention, based on given instructions.

“28 percent of companies report that they have so far only reached the first stage, where AI options are used as task assistants and knowledge agents,” says Henrik Bibow, Associate Partner at PA Consulting.

“30 percent of respondents have reached the second stage, but that still means that 58 percent of companies have not advanced beyond implementing general knowledge agents,” he adds.

More productive

He also highlights that 27 percent state they have implemented decision support systems, while only 15 percent have made it to the final stage, where machines automatically carry out tasks without human intervention.

“This final group reports a significant improvement in productivity compared to the others. A full 50 percent of respondents in this category say they have seen an increase in productivity of more than 15 percent compared with 25–30 percent of respondents in the other groups,” says Bibow.

As for overall usage, the survey confirms what several other studies this spring have shown: a very high proportion of Nordic companies some 98 percentare using AI, but most are still beginners.

“Full potential is only realised when AI is used to change how companies work and how decisions are made,” says Henrik Bibow, Associate Partner at PA Consulting.

Limited impact

“Most of us, over the last few months, have experienced how generative AI can help automate and streamline manual tasks. But even though the AI revolution has made many daily tasks more efficient, so far it has had little impact on companies’ overarching work processes and decision-making,” says Bibow.

“This means that only a fraction of AI’s potential has been exploited. In other words, AI is primarily a tool that employees use to carry out various tasks not a strategic resource that companies have integrated into the organisation and the way they work,” Bibow explains.

Thinking holistically

The survey also shows that companies which have integrated AI across the organisation use it not only to make employees more efficient, but also to get more out of their investments.

“The most forward-thinking companies take a holistic approach. They have moved from improving individual functions to optimising entire value chains, and they invest in AI skills across the organisation. In addition, they invest significantly in data management to ensure the quality of their data and trust in that data,” says Bibow.

He has some clear views on how companies can get even more out of AI going forward.

“Full potential is only realised when AI is used to change how companies work and how decisions are made. To unlock this potential, companies must start working with AI as an integrated part of their business strategy, and rethink how work is carried out and how they make decisions.”

This article was first published in Computerworld (in Norwegian).

Explore more

Contact the team

We look forward to hearing from you.

Get actionable insight straight to your inbox via our monthly newsletter.