Cyrine BenFarha

Agentic AI differs from traditional predictive or generative AI in autonomy and goal-driven execution. As such, the data requirements for agentic AI are critical and the various data types must be considered—structured, unstructured, real-time streaming—to ensure data quality and completeness to facilitate autonomous decision-making. The challenge for bank executives is how to handle multi-source data integration (core banking, CRM, third-party feeds, open banking APIs, etc.) and create an infrastructure for scalability. The roundtable will discuss issues related to data governance and regulatory compliance, security and risk (for agent-to-agent communication. among other things), and interoperability and integration with legacy core banking systems to enable cross-domain decision-making.