The latest release of Oracle integration now includes capabilities to create Agentic AI Agents and publishing MCP Tools.
Executive Summary
- Oracle Integration now generates governed, trusted AI tools for any agent framework using a Model Context Protocol (MCP) server, with a single click.
- Customers can expose their existing integrations and automations as MCP servers to orchestrate complex, multistep processes with unified observability.
- With predictable MCP tools from Oracle Integration, your teams can scale agentic AI initiatives quickly, confidently, and with full visibility.
What if you could give any AI agent trusted, predictable access to business data, actions, and enterprise workflows with one click?
With Oracle Integration’s latest innovations, that’s now a reality for today’s AI leaders.
The era of simple chatbots is over. Forward-looking organizations are investing in agentic AI to combine sophisticated reasoning with safe, auditable actions across enterprise applications, data, and processes. The challenge is that scaling sophisticated agents for real business operations requires more than language skills. It demands deterministic, governed automation and orchestration, trusted integrations, and explainability for every step.
Oracle Integration bridges this gap.
- Now, you can publish any integration project as an enterprise-ready agent tool for any framework using the built-in MCP server generation feature.
- This Model Context Protocol (MCP) standard allows agents, whether built in Langchain, Fusion Agent Studio, Langflow, or elsewhere, to reliably discover, invoke, and describe orchestrated business automations.
These tools are not just simple task automations; they are trusted, auditable endpoints for agents to consistently execute business logic, invoke data flows, and coordinate human workflows, all governed by consistent guardrails.
How Oracle Integration Makes AI Agents Smarter
Smarter with Data:
Oracle Integration offers an ever-expanding library of adapters that connect AI agents to ERP, HCM, SCM, and other essential systems.
Smarter with Action:
Your automations, which orchestrate APIs, robots, business rules, and human-in-the-loop approvals, can now be securely exposed as structured agent tools.
Smarter with Predictability:
Each agentic action executes deterministic, orchestrated flows, giving IT and business users full audit, observability, and policy control.
How To Create MCP Servers and Agentic AI Tools in Oracle Integration
To create an MCP server and AI-ready tools:
- Select any integration project.
- Use the designer to click Create AI Agents > Tool.
- With a single click, you can expose your OIC integrations as AI agent tools.
- Add a description and guidelines to the tool definition. The agent will use guidelines to decide how and when to use the tool.
- Another click and you can expose your project as an MCP Server.
- This makes your AI agent tools discoverable by any third party AI agent framework (for example, Langflow. See images below).
The image above shows an AI agent configured in Langflow using AT Tools published in Oracle Integration and made available via the MCP protocol. This configuration allows the Langflow agent to discover/list the tools, understand what they do, and call them at any time as part of its reasoning and actions.
AI Agents in Oracle Integration
In addition to creating MCP Tools in Oracle Integration, you can also create AI agents. Use Oracle Integration’s new AI Agent Designer to build agents that implement adaptive enterprise workflows accessing your enterprise data and apps. We currently support the ReAct (Reasoning and Acting) agent pattern, breaking complex tasks into a series of step wise, context-aware automations. The agent maintains memory to make smarter, more relevant decisions at each step. We will soon support the Pln pattern as well (and many more in the future).
This approach allows you to chain together thoughts (reasoning), actions (invoking tools), and observations (feedback), so you can automate complex processes while embedding human-in-the-loop governance, as needed.
The image above shows the agent’s reasoning output. In this case, it has just called a tool named ContractSearch and then observes the results and analyzes the data it found before moving on to its next objective.
How To Create AI Agents in Oracle Integration
To create and run an AI agent:
- Create the relevant agent pattern for the agent to follow.
- Create the agent.
Note the usage of the previously defined tools.
- Run the agent.
- Review the result.
Why This Matters for CIOs and IT Leaders
These new MCP and agentic AI capabilities mean:
- Any trusted Oracle Integration project can instantly power secure, explainable AI at enterprise scale.
- Integrations, automations, and audits are now available to every AI agent, whether developed in Oracle Fusion Agent Studio, by partners, or by third-party frameworks.
- Your teams get predictable, governable results—accelerating innovation and reducing risk.
Making Agentic AI Simple
“Our mission is to make advanced AI explainable, governable, and enterprise-ready. With one-click MCP tool generation and agentic AI features, Oracle Integration empowers customers to turn trusted automations into secure, agent-ready tools, driving innovation that scales.”
—Nathan Angstadt, Senior Director, Oracle Integration Product Management
Get Started Now
If you would like to try out Oracle Integration’s agentic AI features today, please join our Accelerate Program.
You can also view a step-by-step tutorial here.