Most AI debates focus on training. For banks and insurers, the bigger risk sits in inference: every prompt, document and customer record a model sees at runtime. This session explains what inference is, why it breaks down for regulated enterprises, and what it takes to run any model, on any hardware, without data leaving the country. We close with how this works in practice in Singapore.

Permissioned identity is emerging as critical infrastructure for trusted client relationships in an AI-powered financial ecosystem. This fireside chat explores how building this foundation can enable AI adoption at scale, deliver smarter client experiences, and strengthen risk controls. We’ll also discuss how established institutions and emerging ecosystem players can work together to build sustainable, trusted foundations for long-term success.

As financial institutions move agentic AI into core business, gains in growth, productivity and efficiency may bring consequences that are harder to see—and harder to contain. Within firms, greater autonomy could reshape jobs, decision-making, accountability and operational risk. Across markets and governments, it raises questions about oversight, stability and who is responsible when agents act across institutional boundaries. Closing this series, the session looks beyond the promised gains to ask who benefits, who bears the costs, and how leaders can recognise and govern outcomes they did not intend.