S&P Global Energy opens data to AI agents via Databricks Genie Agents and MCP

S&P Global Energy's move shows how the division of labor around enterprise data is shifting. Previously, opening a new data domain to conversational use meant a cycle of requirements gathering, API design, engineering and testing. Now, subject-matter experts add table descriptions, example queries and business definitions instead of writing code, and those agents are exposed as MCP servers. A shared proxy layer groups the domain-specific agents, so a calling agent can send a query to one or several data domains at once.

The engineering role has not vanished; it has moved up a layer. Instead of building a separate conversational interface for every domain, engineers maintain the common infrastructure that connects and combines those agents. That same governed access layer can be reused across customer AI environments, rather than building a separate connection for each application, which also changes how additional conversational data products get created.

What the outcome hinges on is less clear. Databricks lists its managed MCP servers, including Genie Agent servers, as Public Preview, so wider availability may depend on that status advancing. The joint announcement did not disclose benchmark scores, measured deployment-time data, or figures on external-customer adoption, leaving the real-world accuracy and uptake of the setup unquantified for now.