The Genie One MCP is now generally available for Databricks users, giving AI agents a governed way to access Genie Oneโs trusted insights across structured and unstructured data. Teams can bring consistent business context into Claude, ChatGPT, Cursor, custom internal interfaces, and other MCP-compatible workflows.
Key highlights
- One context layer for every agent: Genie One grounds questions in Genie Ontology, using approved business definitions, relationships, source authority, permissions, and context from governed enterprise assets.
- Structured and unstructured insights through one interface: Agents can ask Genie One natural-language questions, retrieve query results, continue conversations, and use insights from across the data estate without building separate connectors for every format.
- Interactive answers, not just text: Clients that support MCP Apps can render Genieโs progress, visualizations, results, and Genie Ontology citations inline. Other clients continue to receive text responses with links to Databricks sources.
- Centralized governance through Unity Gateway: The managed MCP Service, is governed through Unity Gateway. Unity Catalog grants control who can invoke it, service policies govern tool calls, and invocations are recorded for usage and audit.
- Permission-aware access: Unity Catalog permissions are enforced on every request. On-behalf-of user authentication lets Genie evaluate access in the end userโs context, so results and source links reflect what that user is authorized to see.
- Use it across practical workflows: Bring trusted business data into presentation-building agents, customer outreach workflows, coding agents, and other automations that need governed context before taking action.
๐ Read the full story here
