[Databricks Data + AI Summit 2026] The Dawn of the Agentic Data Era, as Seen Through the Keynote

For AI Agents to successfully take root in enterprise environments, Context and Control are essential beyond mere model performance. Agents must understand the relationships between data, business terminology, user permissions, and work processes, and enterprises must have a governance system in place to control this.

To this end, Databricks presented a new direction called Agentic Data, unveiling an architecture that goes beyond analytics to serve as the real-time execution foundation for AI Agents.

• Lakehouse//RT: A real-time data warehouse built on the Reyden engine that delivers millisecond-level responsiveness, supporting real-time workloads without data movement.
• Lakebase: A Serverless Postgres for Agentic applications that functions as an operational database and provides high resilience.
• LTAP (Lake Transactional + Analytical Processing): A direction that connects OLTP and OLAP to simultaneously secure real-time performance and analytical performance.

Databricks also introduced Genie, an AI Assistant that lets users gain data insights and automate tasks through natural language. Genie learns an enterprise's unique context through Genie Ontology, and aims for operational automation—autonomously monitoring and optimizing data pipelines and models—through Genie ZeroOps.