Oracle Corp. is extending its partnership with Google LLC’s Cloud to simplify how enterprise users interact with data, introducing a new natural-language interface for queries directly against Oracle databases.

The company today announced the Oracle AI Database Agent for Gemini Enterprise, a tool that allows users to ask questions of enterprise data without writing SQL or understanding underlying schemas. The agent will be available through Google Cloud Marketplace for customers running the Oracle Autonomous AI Database on Google Cloud.

The offering builds on Oracle’s AI Database@Google Cloud service, which deliver Oracle database services, including Exadata and the Oracle Autonomous AI Database, from Google Cloud data centers. The service is currently available in 15 regions, with expansion planned.

The new agent is intended to address a bottleneck in enterprise analytics caused by the need for technical teams to translate business questions into database queries.

“When most business users want a question answered or a report, they contact IT,” said Arpan Shah, senior vice president of database product marketing at Oracle. “With the announcement, business users will be able to ask questions right through Gemini Enterprise in natural language.”

Unlike tools that extract or replicate data into separate AI pipelines, Oracle keeps query processing entirely inside the database for performance and security reasons. “There’s no data moved; it’s all done securely,” Shah said.

Oracle executives said the architecture is intended to preserve data governance and reduce latency. Security controls are enforced at the database layer using existing access privileges. User identity is propagated from the front-end interface to the database, ensuring that query results reflect only the data each user is authorized to see. “A user can only ask questions on data he or she has access to,” Shah said.

This model builds on Oracle’s recently introduced “deep data security” capabilities, which enable fine-grained controls at the row and column level. “We’re relying on the data tier to make the right decisions about who’s got access to what,” said Nathan Thomas, senior vice president of product management.

The companies are positioning integration as a way to streamline both business intelligence and AI development workflows. In addition to end-user querying through Gemini Enterprise, developers can incorporate the agent into applications built on Google Cloud’s Vertex AI platform. Oracle said the system supports agent-to-agent interactions for more complex workflows.

Performance considerations vary depending on deployment architecture. Oracle co-locates database services with application and AI endpoints within the same cloud region to minimize latency. For cross-region or multicloud scenarios, the company relies on dedicated interconnects, though executives acknowledged that latency tradeoffs remain in distributed environments.

Oracle also highlighted potential cost benefits. By shifting semantic interpretation and query generation into the database, customers reduce reliance on token-based processing in external AI services.

The agent will be available at no additional cost for existing Oracle Autonomous AI Database customers on Google Cloud. Initial access will be limited during a rollout period, with broader availability expected later this summer.

Image: SiliconANGLE/Ideogram