What Okta learned governing context for AI agents at scale

Okta moved its AI agent context out of the Snowflake UI and into dbt, using a three-layer framework of agent metadata in model YAML, markdown context docs, and semantic views in dbt's semantic layer spec, all shipped through normal pull requests. The context syncs into Agents Schema, so Claude (over MCP) and Snowflake Cortex read the same governed rules, and an analytics engineer fixes a wrong answer by merging one change that every agent inherits.