LAS VEGAS – October 27, 2026 -- Oracle will roll out AI-generated business ontologies, a semantic reasoning engine and an AI agent registry within the next 12 months as part of its AI Data Platform, and has launched a services offering to move enterprise AI projects into production. The company announced the moves at Oracle AI World.
AI Discovery agent drafts ontologies from schemas and documents
Oracle's AI Discovery agent generates business ontologies by analyzing database schemas, semantic models, application and API documentation, and business knowledge held in unstructured documents. Domain experts review and refine the output before approving it as the shared foundation for all AI solutions. The platform also monitors how agents use the ontologies and recommends updates for expert approval.
Oracle contrasts the approach with traditional ontology projects, which it says often depend on extensive manual modeling and specialist maintenance.
Semantic Reasoning Engine assembles context before it reaches the model
The Oracle Semantic Reasoning Engine uses the ontologies to build business-ready context, orchestrate retrieval across federated systems, resolve business relationships and apply governance controls. The model is then left to handle language understanding, reasoning, explanation and user interaction. By reusing context and prior analytical findings across queries, the engine is designed to cut repetitive reasoning and token consumption.
Deep Insight Agents build on this layer. They identify drivers, run forecasts and scenarios, and recommend next actions, with human review and approval. Their output can be delivered through Oracle Analytics Cloud, Oracle Fusion Data Intelligence and custom applications.
AI Registry inventories agents, tools, skills and MCP servers
Oracle AI Registry maintains a centralized inventory of agents, tools, skills and MCP servers, including ownership, versions, dependencies, permissions and lifecycle status. Oracle AI Governance and Control is planned to add testing, policy assessment and ongoing monitoring of agent behavior, quality, performance, usage and cost. Oracle aims to help customers avoid agent sprawl, duplicated effort and unclear ownership.
Business users can reach registered agents through Oracle Agent Hub or supported third-party assistants such as ChatGPT and Claude. Access and execution remain subject to enterprise security policies.
Innovation Service pairs Oracle engineers with customer teams
The new Oracle AI Data Platform Innovation Service has Oracle experts work with customer teams to define business outcomes, then design, build, validate, deploy and operate solutions on the AI Data Platform and Oracle Cloud Infrastructure. It combines forward deployed engineering with ongoing operational support.
IDC research vice president Megha Kumar said the platform helps organizations "move beyond experimentation and toward intelligent, governed business outcomes." BARC vice president of research Kevin Petrie said ontologies let models and agents grasp relationships between key concepts without consuming unnecessary tokens. Oracle executive vice president T.K. Anand said agents are moving from information retrieval into business decisions and execution.
Oracle said timing and features may change at its discretion.