After four years, the Collaboration for AI in Municipalities and Civil Society (Kraftsamlingen) is now concluding. Since 2022, the program has reached a large proportion of Sweden's municipalities and engaged thousands of individuals.
![]()
Kraftsamlingen has served as a national arena for the target groups to cooperate on artificial intelligence. This has included funding 33 development projects and establishing an AI Council for the country's municipalities. We can also see important ripple effects.
![]()
Rebecka Lönnroth
Head of AI Adoption Public Sector at AI Sweden
When Collaboration for AI in Municipalities and Civil Society launched in 2022, ChatGPT had not even been released. However, OpenAI had already shown GPT-3 in 2020, and it was possible to see that something was about to happen with large language models. Furthermore, there were other AI technologies that could create value for municipalities and civil society.
And that is why this investment was made.
– Both municipalities and civil society lacked an arena for cooperation on artificial intelligence. Kraftsamlingen has shown that a national effort focusing on organizational maturity, learning, and cooperation can deepen understanding and capability in the municipal sector and civil society, says Rebecka Lönnroth.
Rebecka Lönnroth.
Kraftsamlingen was funded by Vinnova and run in collaboration between AI Sweden, Vinnova, the Swedish Association of Local Authorities and Regions (SKR), the Agency for Digital Government (DIGG), and the Swedish Agency for Youth and Civil Society Affairs (MUCF). By participating in activities such as training and webinars, thousands of individuals have strengthened their basic understanding of AI thanks to the program.
![]()
We have seen several examples of concrete progression where organizations have moved from interest to action as they started their AI journey and carried out development projects within the framework of the programme.
![]()
Andreas Skog
AI Change Agent at AI Sweden
Such development projects have helped to advance the municipalities that received funding. They also served as important reference cases for others to learn from.
– When Kraftsamlingen started, there were few relevant examples for municipalities to look at. Being able to show concrete value creation early on has been important, says Andreas Skog.
Another central result is the AI Council that AI Sweden started together with SKR. With representatives from municipalities of various sizes and maturity levels, the AI Council gave the sector a voice at a stage when such structures were not yet established.
Representatives in the AI Council: Katarina Lagerqvist (Chief Digital Officer, Kristinehamn), Anna Bengtsson (digitaliseringsledare, Herrljunga/Vårgårda), Stefan Eriksson (utvecklingsledare digitalisering, Gagnef), Iwona Carlsson (Chief Digital Officer, Kungsbacka), Maria Svensson (utvecklingschef, Värmdö), Frédéric Rambaud (digitaliseringschef, Eskilstuna), Fredrik Edholm (digitaliseringsstrateg, Skövde), Mathias Andersson (data scientist/AI lead, Helsingborg), Mårten Lindskog (AI co-ordinator, Stockholm), Fredrik Hallgren (utvecklingsledare AI, Göteborg), Jari Koponenpr (programansvarig Handlingsplan för digitalisering, Sundsvall), Torbjörn Svedung (CDO/CISO, Mellerud).
Rebecka Lönnroth also notes that the work has generated more ripple effects. The most prominent example is the project A Shared Digital Assistant for the Public Sector.
– It is gratifying to see concrete system effects of the work. The fact that so many municipalities are now joining forces in an innovation project of this caliber shows that the sector has taken the step from being passive recipients to actively driving the national development and use of the technology.
Team members from the project 'A shared digital assistant for the public sector'.
In the autumn of 2025, an interview study with participants was published. It clearly shows that the path towards increased AI maturity is not about individual technical solutions, but about building a cohesive organizational capability that includes continuous learning, collaboration, and a strategy that integrates AI into the core of the operation.
– The overall picture from Kraftsamlingen is that the biggest obstacles for municipalities and civil society organizations are not primarily technical. They are rather organizational and structural and concern leadership, legal frameworks, data access, financing, and forms of cooperation, says Rebecka Lönnroth.
Download and read the report (in Swedish).
AI development in Swedish municipalities is carried out with relatively little coherence or shared direction. Within individual municipalities, there are islands of high AI engagement, and there are also major differences between municipalities. This means that potential synergy effects are often missed and that there is also a risk that some municipalities fall behind in the use of AI.
– Kraftsamlingen shows that the capacity for initiative is high, but municipal managers and decision-makers need to capture the curiosity and lead the work within an already complex governance and management structure. We see the same tendencies in civil society, says Rebecka Lönnroth and continues:
– National coordination is needed that not only stimulates individual projects but also strengthens collaborations for sustainable and more equitable AI development and thereby helps to create the conditions for scaling, sharing resources, and realizing the effects of AI.
Katarina Lagerqvist
Kristinehamns municipality
Networks and a deeper knowledge of AI solutions—what is required and what benefit they can provide—are the most important things we took away from Kraftsamlingen.
Our participation has, among other things, resulted in all municipalities in Värmland now being involved in A Shared Digital Assistant for the Public Sector, thanks to the knowledge I was able to pass on to them. I have also been able to contribute to work with Compare and county-wide issues concerning the AI Catalyst in Värmland with experiences and lessons learned. The work in the project clearly showed how important it is to learn from what others are doing.
Through the AI Council, we have had at least some political impact by being able to provide input with a united voice from the municipalities on what we saw as important moving forward, for example, the need for an AI strategy and the importance of learning and acting together.
Therese Perner
Lidingö Stad
The most important thing we gained is the importance of a clear vision and having a concept for what a project should deliver and how the result should be managed. I also think the importance of cooperation was a clear common thread in all projects.
Following the project, we have continued to build on our masking service. It has been scaled up within the organization and we continue the development. We have also started a new project called Document Search, which is based on the prototype developed within the project.
Overall, we have learned to work iteratively and cooperate across municipal boundaries and in close collaboration with the business sector. This has led to our supplier developing something they call the "Lidingö Model," which involves a close and transparent partnership between the municipality and the supplier where both work together towards a common goal. However, this requires the organization to allocate resources for it.
Åse Andersson
Linköpings municipality
We set learning as our most important goal in the project, and we achieved it in four areas: What AI can be used for in social services, what prerequisites are needed within the organization, knowledge about information management, and lessons learned about our own data.
The work also clarified the need for coordination in the municipality as a whole and for having a common structure. However, the great need for national coordination also became clear, particularly regarding data sharing and legal guidance.
Furthermore, the project contributed insights into the need for operational management in our own organization, where we must be prepared to incur management costs in both money and personnel resources to achieve sustainable digital solutions. And that we need to work on our data structure. We have continued working and see positive development in these areas today. Even if Kraftsamlingen itself did not bring this about, the venture and our project contributed additional arguments. I see that as a strength.
Sara Hugosson
ChildX
For us, the most important thing was to get started, to dare to test new technology and learn along the way! Simply put, "start our AI journey." We were six organizations (1000 Möjligheter, ChildX, MÄN, Novahuset, Storasyster, and Tjejzonen) that together developed an AI tool (FRONTA) to streamline the analysis of chats that support organizations have with children and young people.
After the project concluded, we have continued to develop the solution and received more funds from, among others, the World Childhood Foundation to implement it in a real-world environment. FRONTA is used by the organizations today, but given how fast everything is moving in AI development, we will soon need to update the tool again.
The work has also helped us understand what opportunities exist—how we can more effectively analyze data and information. It has sparked a new curiosity and desire to improve even more. I believe Kraftsamlingen was crucial for many organizations to start thinking about how AI can be used in their operations. There will be a big void now that it's over!
Ylva Christiansson
Friends
For us, Kraftsamlingen was a way to start small and thereby build an understanding of the necessary collaboration for it to work, and to start thinking about the possibilities that exist with AI in other parts of the operation.
Following the conclusion of our project, we have experimented further with AI categorization of open-ended text responses. Kraftsamlingen was significant for the individual civil society organizations that participated, but they were quite few in relation to the number of municipalities.
It is difficult to see an impact on Sweden's civil society organizations as a whole because it is a very large and diversified group of actors, most of whom did not come into contact with the project. However, all extra support to civil society organizations is valuable as we have such small internal resources to invest in experimental activities!
Build upon the structures, networks, and forms of cooperation established within the framework of the Kraftsamlingen Initiative to ensure continuity, shared understanding, and a common direction for Sweden's municipalities. National forums for sharing experiences, learning, and prioritization need to remain and be developed to support municipalities and civil society in the next phase, where the focus is on scaling and long-term integration of AI.
Develop models for how AI solutions, working methods, and expertise can be shared and reused between organizations and sectors. This includes common approaches to procurement and financing that reduce duplicate work. This can also enable smaller organizations that lack their own capacity to take part in the opportunities presented by AI.
Direct future investments toward the structural prerequisites identified as crucial obstacles to scaling, namely data access and data sharing, legal frameworks and guidance, and technical infrastructure that can be used broadly by municipalities and civil society organizations.
The Collaboration for AI in Municipalities and Civil Society program served as the engine for a national initiative from 2022–2025 to strengthen the capability of Sweden's municipalities and civil society organizations to use artificial intelligence. Through a strategic partnership between AI Sweden and Vinnova, in collaboration with the Swedish Association of Local Authorities and Regions (SKR), the Agency for Digital Government (Digg), and the Swedish Agency for Youth and Civil Society Affairs (MUCF), the venture created a national arena for joint work and organizational AI maturity.
A central part has been the financing of around 30 development projects and the creation of extensive structural capital. The initiative has reached over 200 of Sweden's municipalities and 50 civil society organizations.
The experiences show that the biggest obstacles to AI development are not technical in nature but concern leadership, legal frameworks, data preparedness, and the capacity for long-term collaboration. Through concrete development work and experience sharing, the program has led to increased AI maturity and consensus, not least through the establishment of the AI Council for municipalities, which gave the sector a united voice in national policy dialogues. An important system effect is the emergence of cooperation between organizations in the Kraftsamlingen Initiative and new initiatives that have been started based on identified needs.
Related content