Business Central MCP Server: Connecting AI Agents to Your ERP
AI assistants are only as useful as the data they can reach. A chatbot that cannot see your orders, stock or customer balances can only give general answers. The Model Context Protocol (MCP) changes that. It is an open standard that lets AI tools connect to business systems in a uniform, secure way. With the Business Central MCP server, your ERP becomes a system that AI agents can work with, under the controls you define.
Key takeaways for leaders
- MCP is an open standard, often described as “USB-C for AI”, for connecting AI assistants to business systems.
- Business Central’s MCP server has been generally available since version 28 (April 2026).
- It allows agents built in Copilot Studio, and other MCP-compliant tools, to read and act on Business Central data.
- Access respects Business Central permissions, so governance is about deciding what each agent is allowed to do.
- Start with read-only, high-value scenarios, such as answering sales or service questions, before you allow write actions.
What Is MCP and Why Does It Matter?
Before MCP, every AI integration needed its own custom connector. MCP defines a common way for an AI client to discover what a system offers, such as customers, items and sales orders, and to call those operations. For Business Central customers, this means:
- Less custom integration work to connect AI tools to your ERP.
- Freedom of choice: Copilot Studio, Microsoft 365 Copilot and other MCP-compliant AI tools can use the same connection.
- Business logic is respected: agents work through Business Central’s APIs and rules, not directly on the database.
From Preview to General Availability
Microsoft introduced the Business Central MCP server as a public preview during 2025 release wave 2. With version 28 (2026 release wave 1), the enhanced MCP server became generally available and also works with Visual Studio Code and other MCP-compliant clients. An MCP server for the Business Central admin center, which lets AI agents help with administration tasks, is in preview. Learn more in Microsoft’s MCP documentation.
Practical Use Cases
- Sales assistant in Teams: “What did customer X order last quarter, and are there any open invoices?”
- Customer service agent: answers order status and delivery questions using live Business Central data.
- Management Q&A: executives ask about revenue, margins or stock directly in Microsoft 365 Copilot.
- Process agents: custom agents that create quotes, update records or prepare documents as part of a wider workflow.
- Developer productivity: AI coding tools that understand your Business Central environment.
Governance: Questions Leaders Should Ask
- Which data may each agent see? Use dedicated permissions and do not reuse administrator accounts.
- Read or write? Start with read-only scenarios and add write actions with human review.
- Who owns each agent? Every agent needs a business owner and a technical owner.
- How is usage monitored? Track which agents access what, and the cost of AI consumption.
- Is it documented? Add AI agents to your IT and data-protection documentation, including UK GDPR records of processing where relevant.
Frequently Asked Questions
What is the Business Central MCP server?
It is a server built into Business Central that implements the open Model Context Protocol, so AI assistants and agents can discover and use Business Central data and operations in a standard, governed way.
Is the Business Central MCP server generally available?
Yes. After a public preview in 2025 release wave 2, the enhanced MCP server became generally available with version 28 (2026 release wave 1).
Can AI agents bypass Business Central security through MCP?
No. Access through the MCP server respects the permissions of the user or agent identity that connects, so you control what each agent can see and do.
MCP makes Business Central part of your company’s AI ecosystem. The winners will be the companies that connect their AI to real business data, with clear governance.
Do you have an AI use case that needs Business Central data? Contact us or book a call with one of our experts. See also our integration services.
Related reading
