Implementing Generative AI in Swedish Companies : A Qualitative Study on Key Challenges and Success Factors

Sammanfattning

The use of generative artificial intelligence (GenAI) has increased rapidly globally. This study aims to identify key challenges and success factors related to the implementation of GenAI in Swedish companies. The study is based on a qualitative research design with semi-structured interviews with individuals involved in GenAI projects within Swedish organizations. The People-Process-Data-Technology (2PDT) theoretical framework has been used to analyse the empirical material. The results show that organizational and data-related challenges, such as poor data quality, insufficient data management, and hype-driven initiatives without clear requirements, pose greater obstacles than technical factors. End user skepticism also emerged as a barrier to adoption and value realization. In contrast, agile methodologies, cross-functional teams, established system implementation processes, and internal training were identified as success factors in implementing GenAI. Although technology, such as Retrieval-augmented generation (RAG) architecture and cloud-based solutions, helps with the implementation, the study highlights the importance of internal training efforts and understanding of user needs. The conclusion shows that Swedish companies are primarily hindered by organizational immaturity and data limitations rather than by technology itself. A holistic approach, where people, processes, data, technology, and culture are aligned, is essential for successfully implementing GenAI, which corresponds well with the principles of the 2PDT framework. The thesis provides insights that can be used in further research.