Interest in the integration of artificial intelligence-based systems in the public sector follows in the footsteps of earlier movements toward widespread information, communications and technology's adoption in the public sector given the expected improvement in public service delivery, responsiveness and efficiency gains within public administrations. However, AI adoption in the public sector comes with challenges that cut across technological, management and governance dimensions and thus require enhanced competence both at the individual and organizational level. The chapter builds upon research conducted by the lead author as part of the European Commission's AI Watch for the public sector, and the experience of the EU-funded Master in artificial intelligence in public services (AI4GOV). It further leverages the work conducted by the authors for UNESCO to operationalize a framework for capacity building in the area of digital governance and AI at the individual and organizational level. Based on results from these efforts, the chapter outlines the key competences and training requirements needed to adopt AI in government as well as the overall conditions in which these capacity building efforts should be anchored. After discussing the conceptual framework developed for capacity building the analysis presents the key conditions needed to ensure sustainable learning outcomes for both individuals and organizational units. This discussion is complemented by presentation of the outline of an AI and Digital Governance Compass that aims to guide public sector organizations in identifying a mix of hard skills, soft skills and managerial competences that they need to put in place to achieve the organization's objectives. The chapter further emphasizes the importance of interdisciplinary education and methodologies integrating data science and service design. To nurture such an approach is the unique feature of the AI4GOV Master aimed at training "functional specialists" in AI for public services, it is crucial to experiment with teamwork and project-based learning. This approach goes beyond the operational use of AI by the public sector for carrying out administrative duties, and addresses the governance challenges that are the core of policymaking in the digital age. The chapter concludes with recommendations for policy and future research directions, to identify and possibly anticipate open-ended future issues that are emerging in the realm of AI governance in the public sector.