Oracle introduced new agentic Artificial Intelligence capabilities for Oracle Artificial Intelligence Database to help enterprises build, deploy, and scale secure agentic applications for production workloads. The platform is designed to bring Artificial Intelligence and data together across operational databases and analytic lakehouses, allowing agents to securely access real-time enterprise data wherever it resides and combine business data with large language models trained on public data. Oracle said the new capabilities are available across platforms from multicloud environments to on-premises deployments, with additional benefits on Oracle Exadata through Exadata Powered Artificial Intelligence Search for high-volume, multi-step agentic workloads.
The new product additions focus on simplifying development and accelerating deployment. Oracle Autonomous Artificial Intelligence Vector Database is intended to provide the simplicity of a vector database with the broader capabilities of Oracle Artificial Intelligence Database. Currently in limited availability, Autonomous Artificial Intelligence Vector Database is accessible through either the Oracle Cloud free tier, or a developer tier with low-cost pricing. Oracle said customers can upgrade with one click to Oracle Autonomous Artificial Intelligence Database as requirements grow, while retaining support for graph, spatial, JSON, relational, text, and parallel SQL. Oracle Artificial Intelligence Database Private Agent Factory adds a no-code builder for data-driven agents and workflows that runs as a container in public clouds or on-premises, and includes pre-built agents such as a Database Knowledge Agent, a Structured Data Analysis Agent, and a Deep Data Research Agent. Oracle Unified Memory Core is designed to let users store context for agents in one system across vector, JSON, graph, relational, text, spatial, and columnar data.
Oracle also emphasized security controls intended to reduce risk from external attacks, insider misuse, accidental disclosure, and unintended exposure to large language models. Oracle Deep Data Security applies end-user-specific access rules in the database so each end-user or agent acting on that user’s behalf can only see permitted data. Oracle Private Artificial Intelligence Services Container is aimed at organizations with strict security requirements that want to run private instances of models without sharing data with third-party providers or sending data outside the firewall. Oracle Trusted Answer Search is designed to provide a deterministic way to answer user questions by matching them to previously created reports instead of relying directly on a large language model response.
The company also highlighted support for open standards and flexible deployment choices. Oracle Vectors on Ice adds native support for vector data stored in Apache Iceberg tables, enabling Artificial Intelligence Vector Search to read vectors directly from Iceberg, build indexes, and update them as source data changes. Oracle Autonomous Artificial Intelligence Database MCP Server is intended to let external agents and MCP clients securely access Autonomous Artificial Intelligence Database capabilities without custom integration code or manual security administration. Oracle said customers and developers can use the new agentic Artificial Intelligence capabilities now to build applications without moving data, learning new skills, or working around database scalability and security limitations.
