Genie Ontology
#5 in Dokumenten- & WissensmanagementDatabricks · siet 16. Juni 2026 (Ankündigung, Public Preview) – Databricks Data + AI Summit 2026 · 4× · tolest 05. Okt. 2026
Genie Ontology is a unified context layer developed by Databricks (a knowledge graph) that provides Genie One, Genie Agents, and Genie Code with business-aware context. It automatically and continuously extracts knowledge from Databricks-internal assets (tables, dashboards, queries, pipelines) as well as from 50+ connected external apps (e.g., Slack, Jira, Google Drive, Confluence, SharePoint), ranking this knowledge via a PageRank-style algorithm called "OntoRank" based on authority and freshness. The product was announced in June 2026 at the Databricks Data + AI Summit and is currently in public preview, integrated into the Databricks Data Intelligence Platform. Its goal is to enable AI agents to deliver more accurate, faster, and cheaper answers by grounding them in verified business de
Features
| Output Formats | Context snippets/knowledge graph entries consumed by Genie One and Genie Code for answers and actions |
| Base Model | No proprietary LLM; combines Unity Catalog semantics (metric views, domains, pages) with automatically extracted context (OntoRank ranking algorithm) |
| Integrations | 50+ connected apps/systems, incl. Google Drive, Jira, Slack, Confluence, SharePoint |
| Collaboration | Shared ontology across Genie One and Genie Code; context curated once applies to both surfaces; accessible via MCP for external tools |
| License | Proprietary Databricks feature, currently in Public Preview (not open source) |
| Platform | Databricks Data Intelligence Platform, integrated with Unity Catalog |
| Price | Currently no additional costs for curating ontology snippets; part of Genie usage (from $0.07/DBU beyond free tier); pricing subject to change |
| Release Date | June 16, 2026 (announced at Data + AI Summit 2026, Public Preview) |