Building Governed Agents with Databricks
Databricks proposes Unity Catalog as the governance layer to solve enterprise agent sprawl
“Everything in UC gets logged as well. So you can actually see the request response, the token latency, all the errors that's happening in each individual call.”
Databricks is positioning Unity Catalog as the centralized governance layer for AI agents, addressing the growing 'agent sprawl' problem where any engineer can spin up an agent in 20 lines of code but no organization-wide visibility, audit logging, or policy controls exist. The session frames this as a tension between developer ease-of-use and IT governance needs. This matters because it signals Databricks' strategy to own the enterprise agent governance layer, competing directly with other platforms trying to solve the same productionization gap.