Oracle Integration as an MCP Server

Even when you build agentic AI outside Oracle Integration, you can take advantage of the automation investments that you've made by using Oracle Integration as an MCP server. Any agent that works with an MCP server can then use your automation solutions as tools.

Using Oracle Integration as an MCP server allows you bring sophisticated components to adaptive automation by:

  • Leveraging Oracle Integration as an enterprise tool provider for your ecosystem of agentic AI.

  • Reusing your predictable automation components.

 

Agentic AI Platforms

Oracle offers several agentic AI platforms. Build agentic AI using the platform that makes sense for your organization, either in Oracle Integration or somewhere else.

  • Oracle AI Agent Studio

    Part of Fusion Applications. Offers robust capabilities to full-stack developers. See Oracle AI Agent Studio Pattern.

  • Oracle AI Data Platform

    Simplify cataloging, ingesting, and analyzing data for data professionals.

  • Oracle AI Database Private Agent Factory

    Part of Oracle AI Database. Offers robust capabilities to full-stack developers.

  • OCI Generative AI Agents

    Offers robust capabilities to full-stack developers. See OCI Generative AI Agents Pattern.

  • Oracle Integration

    Allows you to build and manage your agentic AI where you build and manage your other automation solutions; offers a low-code, simplified build experience; and supports fast builds. See Learn About Adaptive Automation.

 

Oracle AI Agent Studio Pattern

Oracle AI Agent Studio can use Oracle Integration as an MCP server. In this example, a user submits a purchase requisition (PR) to a portal, and an AI agent updates business systems appropriately.

This pattern is ideal when:

  • You work with Oracle AI Agent Studio or a similar platform.

  • You want to use your AI platform as the primary orchestrator of the automation and use Oracle Integration as a catalog of standard and reusable tools.

  • Your workflows require intelligent validation, enrichment, and adaptive routing.

A business user submits a purchase requisition (PR). Oracle AI Agent Studio validates the PR using the following tools: Tool 1 confirms that all required fields have values with the expected format, and Tool 2 asks an employee to provide values for required fields with missing values. Then, the MCP client for Oracle AI Agent Studio connects to Oracle Integration and its tool catalog, accessing the following tools. (1) An integration that submits the updated PR to the ERP system and awaits acknowledgment. (2) A decision, which uses policies and rules to flag questionable terms. (3) A knowledge base, which finds similar contracts, past legal opinions, and regulator guidance. (4) An integration, which extracts data from the contract, summarizes it, and creates an actionable summary.

 

OCI Generative AI Agents Pattern

OCI Generative AI Agents can use Oracle Integration as an MCP server. In this example, an organization processes vendor invoices by automatically extracting invoice data, validating the payment terms against vendor contracts, and creating invoices in an ERP system.

This pattern is ideal when you need a lightweight, flexible orchestration layer that isn't tied to your AI platform and that uses Oracle Integration wide breadth of technologies.

A business user uploads a vendor invoice to a portal. OCI Generative AI Agents calls the tools in Oracle Integration to complete tasks. Oracle Integration is used as an MCP server, and its tools catalog contains several technologies. First, an integration extracts invoice information and provides a confidence score, Second, an integration that calls a decision determines a discount for an early payment and provides the optimal payment timing for cost savings. Third, an integration with a knowledge base gets the payment terms in a vendor contract from a document repository. Finally, an integration generates an invoice in an ERP system with accounting distributions and payment scheduling.