Artificial intelligence (AI) is a fast-paced technology which can be found in different organizations, including the public sector of Sweden. This advanced tool implies many new work processes and an executive of sufficient basic information for such implementation. While finding a usage within the public sector, a problem occurs around how to and what is needed regarding the integration and processes for a complete normalization of the AI usage. Based on the following research question “What are the main challenges for public sector organizations when it comes to normalizing the use of AI in daily tasks?”, the aim of this thesis is therefore to investigate the key obstacles hindering the integration of AI tools into routine tasks. Through eight semi-structured interviews grounded in the Normalization Process Theory (NPT), the study explores theoretical frameworks surrounding AI, which includes generative AI, and an examination of AI implementations within the public sector context. The conclusion of this thesis reveals several obstacles preventing a complete normalization of AI within the public sector. These include the absence of clear guidelines regarding AI usage, lack of legitimacy for AI tools in current workflows, insufficient competence and development opportunities, and limited resources for AI advancement and utilization. These insights show the many challenges the public sector encounters in embracing AI, and furthermore a need for comprehensive strategies to address these obstacles to facilitate the seamless integration of AI technologies into daily operations.