Our Model Context Protocol (MCP) server for Infor M3 is designed specifically for developers, ERP power users, and automation teams who aim to enable AI agents to interact securely and contextually with their Infor M3 system through a standardized AI-tool interface.
What is the Model Context Protocol (MCP)?
At its core, the Model Context Protocol (MCP) is an emerging open standard for connecting AI agents with various tools, APIs, and systems. It acts as a crucial bridge, allowing AI clients to interact with real-world services seamlessly. Think of MCP as a universal adapter – it provides a standardized way for AI agents to discover and interact with software applications and services.
This standardization is vital because, without it, developers building agentic workflows would have to create custom integrations for every AI-tool interaction, making complex applications incredibly difficult to build and scale.
MCP empowers AI by helping it go beyond static training data, enabling connections to live data sources and real-time capabilities. It addresses several key pain points in AI integration:
- Static knowledge limitations: AI models often struggle with real-time or new data, but MCP allows them to connect to live data sources, making their responses more current and valuable.
- Tool interoperability: Instead of custom integrations for every external tool (like search engines or databases), MCP provides a standardized interface for easier interaction.
- Scalability: It makes scaling AI systems easier by decoupling models from the specific tools or the data they rely on.
Furthermore, when an AI model or agent calls an MCP server, it can deliver structured outputs that are easier to interpret and use. For instance, instead of a raw list of thousands of objects, an LLM powered by MCP can provide an insightful narrative overview.
The problem
The primary challenge our solution addresses is the lack of a unified, AI-accessible interface layer within Infor M3. Currently, there isn’t a reusable abstraction that allows AI or automation agents to seamlessly connect with Infor ERP data without relying on custom-built API invocation implementations for each scenario.
This absence leads to significant inefficiencies:
- Every new automation use case demands redundant API and logic design.
- There’s no established protocol to define context, intent, and response format for AI interactions.
These limitations hinder the agility and scalability of AI-driven automation within Infor M3 environments.
Fortude’s solution
Our solution resolves these issues by introducing a Model Context Protocol (MCP) server specifically for Infor M3. This server provides a structured, reusable interface definition that empowers AI agents to interact with Infor APIs using context-aware prompts.
Our MCP defines:
- The exposed Infor APIs as tools for AI agents.
- An authentication and permission model for secure access
How it works
Our solution enables seamless and secure interactions between AI agents (such as Fortude’s Charlie) and Infor M3 through a standardized, context-aware interface. Here’s a simplified breakdown of the process:
1.User interaction via MCP host: A user initiates an interaction through an AI Assistant or an IDE that supports MCP. They send a natural language query or command, which the AI Agent then interprets and prepares for MCP execution.
2. MCP request initiation: An MCP-client embedded within the AI Agent sends a structured MCP request (via Server Sent Events – SSE) to Infor M3.
3. MCP server configuration & routing: Infor M3 utilizes pre-defined configurations (server-config) to determine the appropriate Infor M3 ION API to call, the necessary input structure, and field-level mappings to the relevant business context.
4. Secure API communication: The mcp-server securely obtains a valid authentication token via REST-based authentication from Infor ION. This token is then used to access the designated Infor M3 ION API, such as a Purchase Order (PO) or Customer Order (CO).
5. Response handling & delivery: The response received from Infor is parsed, structured, and streamed back to the AI Assistant or IDE using the MCP protocol. The AI Agent then presents this information conversationally to the user, with options for follow-up actions or questions.
6. Infrastructure & hosting: The entire mcp-server logic runs as a lightweight Azure Container App, ensuring scalability, reliability, and isolation from host systems.
Key benefits
Our solution delivers a multitude of advantages for organizations using Infor M3 and embracing AI-driven automation:
- Reusable protocol layer: A single MCP definition can support multiple AI and automation use cases, ensuring consistent and reusable intent and tool mappings.
- AI-ready ERP access: It enables natural language agents to securely and contextually interact with ERP data through structured, well-defined APIs.
- Faster time-to-automation: By reducing the need for bespoke integrations for each new use case, our solution accelerates overall automation efforts.
- Centralized control & governance: MCP offers a centralized interface to manage tool definitions, scopes, and access policies, enhancing governance.
- Interoperability with AI Platforms: Our innovation is natively compatible with leading AI platforms like Claude and Copilot, as well as any other AI platform supporting the MCP standard.
- Extensibility: The solution is future-ready, designed to support write operations, workflow triggers, and human-in-the-loop approvals.
- Decoupled architecture: The solution separates the business logic within the ERP from the AI interaction logic, fostering improved maintainability and modular design.
Use case: See our solution in action
With our solution, managing and interacting with Infor M3 becomes as intuitive as a conversation. Imagine using natural language to query order details, confirm purchase orders, or even set up complex workflows, all while your AI assistant handles the underlying API calls securely and efficiently. This transformation significantly streamlines operations, reduces learning curves, and boosts developer productivity, allowing teams to focus on core business objectives rather than complex configurations.
Stay tuned for more updates on how our solution is transforming ERP interactions.
Talk to our experts
Learn how this solution can support your business.FAQs
MCP servers act as a standardized bridge between AI agents and enterprise systems. Instead of building custom integrations for each tool or API, developers can use MCP to provide a unified interface. This allows AI agents to securely access, interpret, and present live business data, making enterprise workflows more intelligent, scalable, and easier to automate.
It is an MCP server built specifically for Infor M3. It provides a structured, reusable interface that allows AI agents to interact with Infor APIs securely and contextually. By translating natural language queries into structured outputs, itsimplifies ERP access, enabling developers, power users, and automation teams to build AI-driven workflows without repeatedly creating custom integrations.
It connects AI agents to Infor M3 through a context-aware MCP server. When a user makes a request, the AI agent sends it via MCP to Infor APIs. The solution handles authentication, routes the request securely, and streams back structured responses. The AI agent then presents this data conversationally, making ERP interactions intuitive and easy to act upon.
It accelerates automation by reducing custom integration needs and offering a reusable interface for multiple AI use cases. It ensures secure, governed access to ERP data, and supports interoperability with leading AI platforms. Its decoupled, future-ready design enhances maintainability while enabling natural language access to Infor M3, boosting productivity and streamlining enterprise operations.
Subscribe to our blog to know all the things we do
