Oracle Bolsters AI Data Platform, Connecting Enterprise AI to Business Context

Oracle unveiled new features of the Oracle AI Data Platform at Oracle AI World held in Las Vegas on the 30th. The goal of the release is to help AI agents jointly understand and use enterprise data and business rules.

Oracle expanded the platform so that, when applying AI agents to multiple tasks in enterprise environments, they can share and use the same data and rules. To do this, Oracle AI Data Platform analyzes database schemas, semantic models, application and API documents, and unstructured documents to generate a business ontology that connects concepts, terms, relationships, and policies. This ontology is a key element for organizing enterprise business information into a structure that AI can understand.

Oracle added business ontology, semantic reasoning, and AI asset governance features. The focus of the additions is to apply data meaning, permissions, and policies consistently when AI agents are deployed across multiple tasks.

The ontology is generated using the AI Discovery Agent, with AI handling the drafting. Business experts then review and revise the draft. It is then finalized based on criteria for shared use across analytics, AI agents, and enterprise applications.

After finalization, the ontology can be updated when business environments or policies change. During this process, the platform monitors how agents use the ontology and suggests items that need to be changed. Enterprise staff review whether the suggested changes are ultimately reflected.

The organized ontology is linked with Oracle Semantic Reasoning Engine. This engine searches for needed information across multiple systems and interprets relationships among data and business rules. It also plays a role in building context for AI models.

The deployment method reuses previous analysis results and business context. This is intended to prevent AI models from repeatedly inferring enterprise data structures and relationships. As a result, repeated computation is reduced, token usage is cut, and operational efficiency is expected to improve.

Oracle assigned the role of connecting the results of ontology and semantic reasoning to real-world business decisions to Deep Insight Agents. These Deep Insight Agents are responsible for analyzing complex business questions, identifying key factors, performing forecasts and scenario analysis, and suggesting next steps.

Enterprises can connect Deep Insight Agents to Oracle Analytics Cloud and also to Oracle Fusion Data Intelligence and their own applications. Oracle has broadened support so that analysis results and recommendations can be applied to data and operations in third-party environments outside the Oracle ecosystem.

In response to the increase in AI agents, Oracle also strengthened management functions. Oracle AI Registry is designed to centrally manage the agents, tools, skills, and MCP servers used by enterprises, and it includes ownership, version, dependency, permission, and lifecycle status as management items for each AI asset.

Oracle will also offer Oracle AI Governance and Control. Based on the AI asset management information in the registry, the feature is structured to monitor agent behavior, quality, performance, usage, and costs.

The feature will allow enterprises to check whether agents comply with policy and are ready for deployment before introduction, and after adoption, it will help detect and respond to unexpected behavior or risks during operation. T.K. Anand, Oracle executive vice president, said AI agents are moving beyond information retrieval and are taking part in business decision-making and execution.

He said that, for enterprises to scale AI agents stably, a context system that reflects how real businesses operate is needed. Registered agents can be searched and used not only in Oracle Agent Hub but also in external AI assistants such as ChatGPT and Claude.

Throughout this process, enterprise-defined security policies and permission frameworks are applied during access and execution. Along with this platform, Oracle also unveiled Oracle AI Data Platform Innovation Service.

Oracle experts support jointly setting customer and business goals. They also support the design and build, validation, deployment, and operations of solutions using the AI Data Platform and OCI.

Source: TECHWORLD · Kim Seung-gi
Original: https://www.epnc.co.kr/news/articleView.html?idxno=409044

References

• https://www.epnc.co.kr/news/articleView.html?idxno=409044