Databricks brings governed business context to ChatGPT, Claude and more

Genie gives agents a business map

Databricks says its Genie plugins are designed to reduce the uncertainty agents face when searching across an organization's data. Instead of moving between documents, schemas and tables without knowing where the right answer lives, agents can use Genie Ontology as an enterprise context layer.

Genie uses the ontology to understand business terms, identify relationships and find relevant data across tables and documents. It then writes and runs SQL over data the user is allowed to see, returning answers with citations to their sources. For specialized questions, Genie can route work to domain-specific Genie Agents, such as a marketing agent for pipeline attribution or a finance agent for operational costs.

The system also supports insights from unstructured data. For clients that support the MCP Apps extension, Genie can return interactive visualizations and Genie Ontology citations directly in the conversation.

How the plugins fit into agent workflows

A complex business question can be handed from an agent to Genie for investigation. Genie interprets the request, retrieves permitted data, runs the required SQL, routes the question when specialized analysis is needed and returns a cited answer. Agents can also request missing business context while continuing a larger task, such as editing files, drafting a plan or posting to Slack.

Databricks says the plugins work natively in ChatGPT, Claude, Microsoft Copilot Cowork and Grok Bot. Agents without a native plugin can reach Genie through the Genie MCP server. Organizations can also use the MCP server directly in tools, including AI assistants, coding agents and client interfaces.

Examples span product, finance, accounts and marketing

In ChatGPT, a product manager's always-on dot could ask Genie about feature adoption using the certified definition of an active user, compare results with launch targets, investigate changes in enterprise usage and post a summary to Slack.

Claude could combine certified revenue, order volume and average deal size metrics with pricing memos, sales call notes and competitor reports to help explain a regional revenue miss. Copilot Cowork could use Azure Databricks Genie to prepare customer-meeting materials, including a Word briefing, an Excel overview with charts and a presentation, while waiting for approval before sending communications.

With Genie connected, a Grok Bot growth marketing teammate could review campaign performance, investigate campaigns missing targets, draft budget changes and prepare re-engagement audiences using governed campaign and pipeline definitions.

What this means for enterprise AI users

The announcement gives organizations a way to define business meaning in Databricks and make that context available across several agent workflows. According to Databricks, this can help teams keep metric definitions and source context consistent while agents answer questions or carry out longer tasks.

Users still work through their preferred agents, while Genie handles data retrieval and governance. Databricks says teams can enable the plugin in an agent workspace and begin asking business questions or assigning tasks, with the MCP server available for agents that do not have native plugin support.