As AI systems move from answering questions to taking actions, the data layer behind them is being asked to do a great deal more. An agent carries conversation state, long term memory, operational records, vector data, tool definitions and permissions, and it queries continuously rather than in human sized bursts. This session looks at how Gen AI and the data layer are advancing together, and what it takes to make the two work as one system. We will explore themes such as vectorization and retrieval, how agents reach data through open protocols like MCP, where agent memory lives, and what changes once an agent can act on data rather than only read it. The aim is to leave you with a clearer view of how data architecture is evolving for agentic AI, and what to weigh when designing for it.