Beyond AI pilots: New European study explores what it takes for public administrations to make Generative AI work
How can public administrations move beyond experimenting with Generative AI to creating lasting public value? A new European Commission report, co-authored by CBS Professor Rony Medaglia, explores how organisational capabilities, governance and leadership shape the responsible adoption of AI across the public sector.
Generative AI is rapidly finding its way into public organisations across Europe. Civil servants are already using AI to draft documents, summarise reports, search for information and support everyday administrative work. Yet the biggest challenge is no longer whether public administrations will adopt Generative AI. It is how they can integrate it responsibly, strategically and at scale.
These questions are at the centre of a new report from the European Commission's Joint Research Centre (JRC), The adoption of Generative AI in EU public administrations: Exploring individual behaviours and organisational approaches. The report is co-authored by Professor Rony Medaglia from the Department of Digitalization at Copenhagen Business School together with Patrick Mikalef, Luca Tangi and Paula Rodriguez Müller.
Based on interviews with public servants across eight European public administrations, the study offers one of the first empirical examinations of how Generative AI is becoming embedded in the daily work of government organisations. Rather than focusing solely on the technology itself, the report investigates the organisational and managerial conditions that determine whether AI creates lasting public value.
AI adoption is as much an organisational challenge as a technological one
The study finds that many public employees are already experimenting with Generative AI to improve productivity and support knowledge-intensive tasks. In many organisations, however, this development is taking place faster than governance structures can keep pace.
Employees frequently rely on commercial AI tools outside official organisational systems, creating what the report describes as "shadow AI". While these practices often emerge from genuine attempts to work more efficiently, they also introduce new questions concerning data protection, oversight, accountability and appropriate use.
“One of the key insights from this research is that successful adoption of Generative AI depends less on the technology itself and more on how organisations learn to integrate it into everyday work. This requires experimentation, but also clear governance and continuous organisational learning.” Rony Medaglia
Professor
The research highlights a series of tensions that public organisations increasingly need to manage. Generative AI can empower employees while simultaneously creating new forms of dependency. It can stimulate innovation while requiring stronger organisational control. It encourages experimentation, yet public administrations must also ensure compliance with regulation and public sector values.
These tensions suggest that successful AI adoption depends on much more than selecting the right technology. It requires organisational capabilities that enable experimentation without compromising trust, transparency or public accountability.
From experimentation to strategic assimilation
One of the report's central contributions is a framework describing different stages through which organisations assimilate Generative AI.
Rather than viewing AI implementation as a simple technology rollout, the framework describes an organisational journey. Early stages are characterised by exploration and isolated experimentation. More advanced stages involve operational integration, strategic transformation and ultimately a reflective approach where governance, continuous learning and human oversight become integral parts of organisational practice.
The framework shifts attention away from measuring adoption alone. Instead, it asks a more important question: what kind of public value should Generative AI create, and what organisational capabilities are needed to realise that value responsibly?
Connecting research with Europe's AI agenda
The publication arrives at a time when Generative AI has become a central priority for European digital policy. European institutions are increasingly focusing on how public administrations can move beyond fragmented pilot projects towards coordinated, trustworthy and sovereign AI adoption.
The report contributes practical recommendations covering governance, organisational readiness, workforce capabilities, secure technical infrastructure and cross-agency collaboration. Together, these recommendations support a transition from isolated experimentation towards long-term organisational strategies for AI in the public sector.
For researchers at Copenhagen Business School, the publication also illustrates the growing importance of interdisciplinary research on digital transformation. Questions surrounding AI are no longer purely technical. They involve management, organisational design, public governance and the relationship between technology and society.
Research that informs practice
Rony Medaglia's contribution to the report reflects his longstanding research on digital government, public sector innovation and the organisational implications of emerging technologies. By combining empirical evidence from European public administrations with conceptual frameworks for AI assimilation, the study contributes both to academic research and to ongoing policy discussions across Europe.
As governments continue to explore how Generative AI can support public service delivery, the research suggests that success will depend less on the capabilities of the technology itself and more on the organisational capacity to govern, integrate and continuously adapt its use.
The full report, The adoption of Generative AI in EU public administrations: Exploring individual behaviours and organisational approaches, is available through the European Commission's Joint Research Centre.