Oracle adds agentic AI tools to Oracle AI Database, Fusion Cloud apps

Oracle said it has added a set of agentic AI features to Oracle AI Database in a move that brings autonomous agents closer to transactional data. Separately, Oracle announced new agentic AI capabilities to its Fusion Cloud applications.

The news, announced at Oracle AI World Tour in London, revolves around using Oracle AI Database to bring together AI and data to enable agents to access data to operate. Customers can use multiple AI models, frameworks and open data formats.

Here's what Oracle AI Database is getting in its latest update.

• Oracle Autonomous AI Vector Database in limited availability. Autonomous AI Vector Database is built on Oracle AI Database and gives developers the ability to build vector-powered apps. Customers can upgrade with one click.
• Oracle AI Database Private Agent Factory, which has a no-code AI agent builder that runs in public clouds on on-premises and a set of pre-built AI agents. Pre-built agents include a Database Knowledge Agent, Structured Data Analysis Agent and Deep Data Research Agent.

• Oracle Unified Memory Core, which gives user one store for context used by AI agents.
• Oracle Deep Data Security, which provides end-user specific data access rules in the database. An end-user also includes an AI agent.
• Oracle Private AI Services Container, which covers the security requirements needed to run private instances of AI models.
• Oracle Trusted Answer Search, which adds deterministic responses to AI answers. Trusted Answer Search uses IA Vector Search to minimize LLM hallucinations.
• Native support for Apache Iceberg tables and Model Context Protocol (MCP).

Holger Mueller, an analyst at Constellation Research, said Oracle's agentic AI moves with its database are compelling. "Oracle future proofs the Oracle Database for the agentic era, by bringing all required capabilities to the database so Oracle database customers can leave their data where they have it, and benefit from AI automation where their transactional data is," said Mueller. "It's a compelling and unique approach in the industry."

Fusion Agentic Applications

Oracle also announced Fusion Agentic Applications, which are powered by coordinated and specialized AI agents that are focused on business outcomes.

The agentic AI applications are built into Oracle Fusion Cloud Applications. Oracle said Fusion Agentic Applications are differentiated because they are native to transactional systems.

Key points about Fusion Agentic Applications:

• They are aimed at business outcomes and made up of teams of AI agents with roles, expertise and decision authority. Fusion Agentic Applications work autonomously on routine processes and only surface exceptions, tradeoffs and decisions where human judgment can change the outcome.
• Fusion Agentic Applications work across the full Oracle applications suite and maintain shared context.
• The agentic applications work in governed workflows that can be audited.
• There are 22 Fusion Agentic Applications focused on everything from workforce operations, payroll, procurement, seals, and cash flow processes.
• Oracle added that new Fusion Agentic Applications ride along with a new Agentic Applications Builder within Oracle AI Agent Studio.