Skip to content
All projects
ConceptAI Agents

ERP MCP Server

A concept for exposing ERP capabilities to LLM agents as governed tools through the Model Context Protocol.

Overview

An architecture for enterprise AI integration in which ERP functions are published as MCP tools. Agents and LLM clients call those tools instead of touching databases directly, with authentication and control in between.

Problem

Connecting LLMs to ERP systems ad hoc leads to fragile integrations and unclear access boundaries, and giving an agent direct database access is unacceptable in an enterprise.

Solution

Wrap selected ERP operations as MCP tools behind an authenticated server. The agent discovers and calls tools; the server enforces permissions, calls existing enterprise APIs and returns structured results.

Architecture

ERP MCP Server architecture, top to bottom
  1. 01

    Client

    • LLM
    • AI agent
  2. 02

    Protocol

    • MCP client
    • Tool discovery
  3. 03

    MCP server

    • Tool definitions
    • Auth & permissions
  4. 04

    Enterprise APIs

    • ERP APIs (ASP.NET Core)
    • Google Workspace APIs
  5. 05

    Data

    • SQL Server
  • Existing ERP APIs remain the system of record; the MCP layer is an adapter, not a rewrite.
  • Authentication and authorization are enforced at the tool boundary.
  • The same pattern extends to Google Workspace and other enterprise APIs.

Key features

  • ERP operations exposed as callable tools
  • Permission checks at the tool boundary
  • Reusable pattern for other enterprise APIs

Engineering challenges

  • Designing tool granularity so agents stay effective without over-broad access.
  • Authenticating agent-initiated calls against enterprise identity.
  • Keeping automated, write-capable actions controlled and auditable.

Results

No measured outcomes are published for this project. I only list results that have been verified.