Oracle Expands AI Agent Studio for Fusion Applications with New Marketplace, LLMs, and Vast Partner Network

• New AI Agent Marketplace delivers partner-built AI agents directly within Oracle Fusion Cloud Applications to streamline operations and unlock new levels of efficiency
• Support for OpenAI, Anthropic, Cohere, Google, Meta, and xAI in Oracle AI Agent Studio for Fusion Applications will enhance flexibility and improve AI adoption
• More than 32,000 certified experts trained in Oracle AI Agent Studio will help customers optimize AI

Oracle announced the latest updates to Oracle AI Agent Studio for Fusion Applications, a comprehensive platform for building, testing, and deploying AI agents and agent teams across the enterprise. The latest updates expand the Oracle Fusion Cloud Applications AI ecosystem and include a new AI Agent Marketplace, extended LLM support and agent-building resources, and a vast network of Oracle-certified AI agent experts.

“Organizations are grappling with rising business complexity and the urgent need to accelerate AI adoption,” said Chris Leone, executive vice president of Applications Development, Oracle. “By building a comprehensive AI ecosystem centered around Oracle Fusion Applications, we’re giving customers the flexibility to address complex challenges swiftly, securely, and confidently. The AI Agent Marketplace and other enhancements to the AI Agent Studio enable our customers to supplement the embedded AI agents in Fusion Applications with validated, industry-specific capabilities from our growing AI ecosystem of systems integrators and independent software vendors.”

New AI Agent Marketplace Taps Partner Expertise to Accelerate Enterprise AI Adoption

The new Oracle Fusion Applications AI Agent Marketplace enables customers to easily deploy Oracle-validated, partner-built AI agents within Oracle Fusion Applications. Marketplace seeded partner templates are embedded in Oracle AI Agent Studio to help customers accelerate automation, boost productivity, and address complex, industry-specific business challenges.

Unlike other AI agent marketplaces, Oracle AI Agent Marketplace is embedded natively within Oracle Fusion Applications, allowing customers to access, test, and deploy third-party AI agents directly within existing workflows. Customers can install and manage validated agent templates, created by certified Oracle PartnerNetwork members, alongside Oracle pre-built agents in a unified experience. To learn more, please visit Oracle AI Agent Marketplace.

New LLMs and Enhancements Support Multi-Step Agentic Workflows in AI Agent Studio

Oracle AI Agent Studio enables customers and partners to select the best-performing LLM for their business needs, with support for OpenAI, Anthropic, Cohere, Google, Meta, and xAI. Additional updates to Oracle AI Agent Studio include new:

Integration & Extensibility Capabilities

• MCP support: Enables users to extend agent capabilities with third-party data and tools through seamless integration with external systems via the Model Context Protocol (MCP), an AI industry standard.
• A2A agent cards: Enables cross-agent collaboration with third-party agents by allowing agents to communicate and pass context between each other via standardized connectors.
• Credential store: Controls to help ensure that AI agents can safely access external services, such as APIs, without exposing sensitive data by enabling users to securely manage API keys and authentication tokens within the Oracle AI Agent Studio.

Observability and Evaluation Capabilities

• Monitoring dashboard: Helps users detect and address AI agent issues quickly by providing real-time visibility into agent performance, such as data around sessions, latency, error rates, and token usage.
• Agent performance evaluation: Improves AI agent effectiveness by providing systematic testing and measurement of how well agents perform in quality, correctness, and safety metrics compared to set goals.
• Agent tracing: Helps users debug AI agents and optimize performance by capturing detailed execution data around agent workflows.
• Performance metrics: Help users improve agent effectiveness and track efficiency over time by monitoring key indicators like correctness, latency, API errors, and token consumption.
• Token usage: Makes costs more predictable for customers by measuring token consumption for premium LLMs.

Prompt Management Capabilities

• Prompt libraries and lifecycle management: Help users manage agents across different parts of the lifecycle—from authoring and testing to version maintenance—by storing all prompts and agentic use cases in a central store.
• Topics management: Helps improve agent consistency by storing all topics being used across AI agents in a central store, increasing visibility into which capabilities are available to specific agents and providing the same prompt boundaries for agents operating in similar domains.

AI Agents Capabilities

• Expanded Agent templates: Help users quickly configure and deploy AI agents by delivering predefined blueprints for common agent use cases within the Oracle AI Agent Studio.
• Agent builder assistant: Enables AI agents in the Oracle AI Agent Studio to create an agent from scratch, defining topics, prompts, and tools based on high level direction from a user.
• AI Agent Studio FAQ agent: Provides additional support for agent builders with a Q&A assistant agent that answers questions using natural language about tasks or projects they are working on within Oracle AI Agent Studio, such as configuring agent templates, building agents from scratch, publishing agents into production, or evaluating agent performance.

Multimodal and Retrieval-Augmented Generation Capabilities

• Multimodal RAG: Improves agent Q&A performance by using RAG to incorporate more than just text – documents, images, tables, etc. – in the agent’s analysis.
• RAG over external sources: Enhances agent performance by using RAG over documents stored in external content repositories such as SharePoint.

Workflow Agents and Nodes Capabilities

• Deterministic execution: Enables consistent and predictable results for business-critical processes by allowing users to set prescriptive agentic workflows with pre-defined outcomes for specific types of AI agents.
• Chaining workflows: Help improve the outcome of deterministic agent execution for multi-steps tasks by connecting multiple workflows together.
• Agent node: Enables users to complete more complex processes with their agent by adding additional agents to a workflow when a more dynamic action is required, such as making decisions or interpreting context.
• Human-in-the-loop: Helps users balance automation with oversight and control by incorporating human review and approvals into workflows.

Trusted Network of Certified Experts

Over 32,000 certified experts have completed rigorous training on how to build the most effective agents within Oracle AI Agent Studio, enabling customers to leverage the highest performing agents and optimize AI across their workflows.

This network of experts will further expand the breadth of Oracle AI Agent Marketplace by adding new expert-built agents and agent templates to help customers maximize their AI potential. This growing AI ecosystem of experts ensures customers have access to validated, secure, and trusted AI agents ready to transform business processes.

Analyst Support

“In the current enterprise AI-arms race, Oracle has proven itself to be a steadfast competitor. With the launch of AI Agent Marketplace, Oracle is raising the bar once again,” said Mickey North Rizza, Group Vice-President Enterprise Software, IDC. “With the new marketplace offering an ever-expanding range of partner-built AI agents natively supported in Fusion Applications, Oracle customers will be uniquely able to accelerate AI adoption. This continued innovation underscores Oracle’s commitment to delivering tangible value and productivity gains to its customers in today’s fiercely competitive AI landscape.”

“Enterprise application suites that embed AI capabilities and offer flexible agent development environments are taking a clear lead in the market,” said Holger Mueller, vice president and principal analyst at Constellation Research. “Expanding these ecosystems with accessible marketplaces for partner-built AI agents represents a natural evolution, making it easier for enterprises to adopt and scale AI-driven automation.”

Source: Oracle media announcement