Oracle’s AI Agent Studio launches marketplace and expands large language model support for Fusion Applications

Oracle announces major updates to its AI Agent Studio for Fusion Applications, including a new AI Agent Marketplace, enhanced support for leading LLMs, and advanced workflow tools to accelerate enterprise AI integration.

Oracle has unveiled significant updates to its AI Agent Studio for Fusion Applications, introducing an AI Agent Marketplace alongside enhanced support for leading large language models (LLMs) and expanded resources aimed at accelerating enterprise AI adoption. This announcement marks a major step in Oracle’s ongoing strategy to embed AI deeply into its cloud applications, offering businesses new ways to automate workflows while addressing industry-specific challenges.

The newly launched Oracle Fusion Applications AI Agent Marketplace provides customers with access to a broad range of Oracle-validated, partner-built AI agents. These agents are seamlessly integrated within Oracle Fusion Applications, enabling businesses to deploy automation capabilities without leaving their existing workflows. According to Oracle’s announcement, the marketplace offers agent templates created by a diverse group of systems integrators and independent software vendors, covering key business functions such as finance, human resources, supply chain, and customer experience. Notable contributors include Apex IT, Grant Thornton, Huron, IBM Consulting, and Infosys, with additional templates available to joint customers from Accenture, Deloitte, KPMG, and PwC.

An example of the marketplace’s practical applications includes IBM’s Smart Sales Order Entry Assistant, which uses natural language prompts to streamline the sales process by reducing errors and manual steps involved in order creation and validation. KPMG’s Purchase Order Item Price History agent provides procurement teams with instant access to historical purchasing data, assisting in negotiation strategies by offering insights on previous suppliers, purchase dates, and average prices. These exemplars demonstrate the potential for AI agents to enhance operational efficiency and decision-making in specific industry contexts.

The update to Oracle’s AI Agent Studio also brings robust new features for agent management and workflow enhancements. These include prompt libraries and lifecycle management tools that allow users to store and manage prompts centrally, alongside a topics management function that ensures consistency across agents with overlapping domains. The studio’s no-code interface enables customers to build customised agents from scratch, modify existing agents, and integrate third-party relationships via MCA and A2A protocols, according to Natalia Rachelson, Group Vice President, Cloud Applications Development at Oracle.

Significantly, Oracle’s AI Agent Studio now supports retrieval-augmented generation (RAG), empowering agents to incorporate and analyse information from multiple formats such as documents, images, and tables. Agents can also perform RAG over external content repositories like SharePoint, broadening their data access capabilities and enabling richer, contextually aware outputs.

Further workflow improvements focus on the execution and oversight of business-critical processes, including deterministic execution settings, chaining multiple agent workflows to tackle complex tasks, adding agent nodes to workflows, and integrating human-in-the-loop controls for crucial interventions. These features enhance operational control and security, essential for enterprises scaling AI deployments.

Oracle has expanded the studio’s compatibility with a wide range of LLMs, including OpenAI, Anthropic, Cohere, Google, Meta, and xAI. This multi-model support allows businesses to select models best suited to their specific needs. Rachelson highlighted Oracle’s privileged market position due to its ownership of Oracle Cloud Infrastructure (OCI), which is purpose-built for AI. This infrastructure gives Oracle direct access to key LLMs, enabling rapid tuning and deployment without delays typically caused by third-party token allotments. She emphasised that Oracle’s AI capabilities are natively built within Fusion Applications, rather than assembled through acquisitions or external integrations.

Additional technical enhancements include Model Context Protocol (MCP) integration to extend agent connectivity with external systems, cross-agent collaboration through A2A agent cards, and a secure credential store to manage API and third-party service authentications. The platform’s observability has also been improved with a new monitoring dashboard that offers real-time performance data, systematic evaluations, workflow analysis, troubleshooting, and token usage monitoring to help manage the costs of premium LLM usage.

Oracle’s AI Agent Studio exemplifies the company’s broader vision of AI augmentation within enterprise software, providing customers and partners with powerful, customizable AI tools that integrate smoothly into existing workflows. This initiative aligns with Oracle’s focus on enabling an AI-driven workforce that can enhance productivity and innovation across multiple business domains.

📌 Reference Map:

Source: Noah Wire Services