Greg Rokita

As enterprises move from AI prototypes to production-grade agents, one challenge dominates: deploying systems that reason deeply over proprietary data, securely and at scale. This session shows how Anthropic Claude's frontier reasoning, natively integrated into the Databricks Platform with Unity AI Gateway for governance, and the Databricks Lakehouse and Vector Search powering their functional skills, provides the full-stack foundation for domain-specific agents enterprises can trust. We will hear from Edmunds on how they built an agentic digital workforce powered by Claude on Databricks. We'll deep-dive into how a "hunting party" of specialized agents autonomously writes code, opens merge requests, and audits 101 legacy schemas in one afternoon (vs. two weeks manually). Attendees will learn how 40+ composable skills yield exponential capability, how a self-improving flywheel slashed prototype-to-production lead times by 50% and how tasks that once cost $640 in human effort now run for $25 in agentic compute.

The data warehouse era is ending. The data lake house has emerged as the new data platform for the cloud by providing one unified platform for both traditional analytics and machine learning.