XNews

Databricks Co-founder Explains Genie Ontology for Trusted AI Analytics

Databricks Co-founder Explains Genie Ontology for Trusted AI Analytics

Databricks Co-founder and VP of Engineering Patrick Wendell explained that Genie Ontology learns how teams define and calculate revenue from existing analytical work, with official definitions in Unity Catalog outranking agent inferences. Wendell stated that when Genie runs a job, it works from numbers the business already trusts rather than guessing from documents and emails.

Event version 1.

Key insights

Revenue definitions precede agent output

Before an agent hands over a revenue number, it needs to know how the team defines and calculates revenue. Genie Ontology learns that from the analytical work teams already do.

Unity Catalog definitions take precedence

Where an official definition exists in Unity Catalog, it outranks anything the agent infers.

Genie works from trusted numbers

Patrick Wendell stated that when Genie runs a job, it works from numbers the business already trusts, not guessing numbers from documents and emails.

Source posts