Databricks previews Genie Ontology to provide business context for AI agents

Databricks is advancing a new context layer for enterprise AI called Genie Ontology, currently in preview. [1] It automatically extracts business context from sources like documents, pipelines, and dashboards to create a living graph for AI agents. [1] Announced at the Data + AI Summit by CEO Ali Ghodsi, the system uses a ranking method inspired by Google’s PageRank to identify authoritative business definitions. [1] Organizations can also upload their own definitions through Databricks’ Unity Catalog Semantics platform. [1]

The shift toward ontology aims to solve consistency issues common in standard retrieval-augmented generation (RAG) deployments. Analysts note that older RAG approaches simply retrieve similar-looking information without understanding specific business meaning, often yielding conflicting answers to the same question. By providing a single unified definition through a context layer, Genie Ontology promises to improve trust and governance for enterprise AI. What the sources don't address: when Genie Ontology will move from preview to general availability.