Databricks Opens Genie One’s Business Context to External AI Agents

Databricks has made the Genie One Model Context Protocol server generally available to all Databricks users. The service gives external AI agents a managed route to retrieve Genie One data, answers and query results while using a shared layer of business context.

One business context across several AI tools

Genie One MCP lets another AI agent ask Genie One questions and retrieve structured and unstructured data, insights, answers, query results and visualizations. Databricks says those responses are grounded in Genie Ontology, its layer of business context spanning data, documents and tools.

The company’s pitch is that employees can keep using the AI tools already in their workflow. Databricks says the connection works with agent clients including Claude, ChatGPT, Cursor and custom internal interfaces, while drawing each request from the same governed business context.

The gateway is the control layer

The connection runs inside Unity Gateway as a managed MCP Service. Databricks says that puts centralized governance, fine-grained policies and audit logging around MCP invocations, rather than leaving each connected agent with a separate path to the underlying business information.

Where Databricks sees the connection fitting

• Presentation agents can retrieve business results and context while preparing slides.
• Customer-outreach agents can investigate product-usage signals before preparing CRM emails.
• Coding agents can retrieve current product definitions and queries before a developer changes logging.

A shared route does not mean an identical interface

Databricks’ companion MCP App can show interactive visualizations and Genie Ontology citations, but only in supported agent clients. The release can standardize where an agent gets its business context and how that access is governed; what users can inspect inside a chosen assistant may still vary by client.