For more than two decades, Denmark has been at the forefront of digitalization of its public sector through innovative new technologies. Recently, Artificial Intelligence (AI) has caught on as a technology to bring about further improvement on such diverse topics as welfare, climate, and government processes. Within this context the Danish Agency of Digital Government launched the initiative Signature Projects—AI pilot projects whose purpose is to gather practical knowledge of the use of AI in a public-sector context. It is thus the goal of this exploratory case study to look into these Signature Projects as devices of public-sector innovation to gain knowledge of the manifold experiences gathered through these Signature Projects. Based on a grounded analysis of six embedded cases chosen among a range of municipal and regional AI projects, this study finds that the knowledge-sharing process employed between these is ineffectual. Based on this finding this report makes four direct recommendations to improve in this area. This study contributes to the theoretical field of public pilot
projects by defining types of experiences acquired.