Wisdom AI
Connect WisdomAI to the full upstream lineage so its Context Layer stays in sync with the platform, agent answers are traceable, and new data sources are modeled faster.
Agentic AnalyticsWhy Wisdom AI matters
WisdomAI lets business users ask questions in natural language and get answers backed by its Context Layer, an enhanced semantic layer that captures business definitions, table relationships, and query rules. Proactive agents monitor metrics, detect anomalies, and deliver insights automatically. But the Context Layer describes the data in the warehouse. It does not track what happens upstream: which source systems feed each table, what transformations built it, or whether a pipeline change has silently altered the data underneath.
Typedef connects WisdomAI to the full upstream lineage, so when source schemas change or transformations are refactored, teams can see which Context Layer definitions need updating before agents start returning stale answers. And when an answer looks wrong, the trail goes all the way back to the source system, not just to the warehouse table.
What Typedef unlocks
End-to-end answer traceability
Trace any WisdomAI answer from the Context Layer definition through the warehouse, transformations, and ingestion back to the source system that originated the data.
Context Layer accuracy
Surface when upstream changes have altered the data that Context Layer definitions describe, so teams can update definitions before agents and Proactive Agents return stale results.
Faster data source onboarding
When modeling new tables into the Context Layer, Typedef provides the upstream lineage and transformation logic so data teams spend less time investigating what feeds each table.
Cross-tool answer reconciliation
Show where WisdomAI answers diverge from traditional dashboards or other analytics tools and trace both paths to the point of divergence.
Related integrations
Trace any Cortex answer back through the semantic layer, warehouse, and source systems so users can verify why the agent produced a specific result.
Trace any Databricks AI/BI answer back through definitions, the lakehouse, and source systems so users can verify the reasoning.
Verify that every metric means what you think it means, from definition to source.
Surface the business logic hidden inside LookML so impact analysis doesn't stop at the BI layer.