<p>The advent of generative Artificial Intelligence (AI), popularized through ChatGPT and other similar platforms, has in recent years become a focal point for governments’ ambitions to enhance the quality of public services and policy. Whilst governments have been using technological systems and platform for decades, such as big data analytics and automated decision-making, recent developments in generative AI, and its perceived capability of simulating human thought, is accompanied by both opportunity and risk. AI is not only significant for the future delivery public services, it also constitutes a challenge to how we define and prioritize public value. This article uses a ‘value-chain approach’ to explore and summarize the immediate experiences of the use of AI in public service contexts, and highlights some of the hurdles and challenges associated with the adoption of this technology. While value-chain approaches are traditionally associated with identifying sequences in a (commercial) production (manufacturing) process - as an analytical tool to realize desired outcomes - recent literature has successfully applied this approach to public service contexts, including in relation to digital service delivery and public policymaking. This article provides an opportunity to reflect on some of the promises and pitfalls associated with this technology, as well as presenting some elements for better diagnostic tools used for forecasting and evaluating digital platforms and systems utilizing generative AI in a more systematic manner. This includes value issues associated with data quality, intellectual property, surveillance, privacy, and transparency. The article also highlights issues around trade-offs between different values, and in doing so, argues that assessing the interlinked relationship between values and digital services are key to understanding the future nature of public service delivery.</p>