Glossary

What is MCP (Model Context Protocol)?

MCP, the Model Context Protocol, is an open standard for connecting AI tools to other systems and data, so an AI assistant or agent can look things up and take actions in them through one common interface.

In plain English.

On its own, an AI model can only work with what you type or paste into it. To be useful in a business, it needs to reach your systems: your orders, products, customers and documents. Before MCP, every AI tool needed its own custom connection to every system. MCP gives them a common language. A system offers an MCP server, describing the information and actions it makes available, and any AI tool that supports MCP can connect to it and use them.

It was introduced by Anthropic, the company behind Claude, in 2024, and is now supported by many AI tools and platforms. A useful comparison is a USB port: one standard plug, so each new device doesn't need its own cable.

An example. A sales director wants to know which trade customers ordered less this quarter than last, and what they stopped buying. With an MCP server on the customer portal's data, they can ask that in plain English from an AI tool such as Claude and get the answer back, without waiting for someone to build a report.

Why it matters

  • Your data where your people already work. Staff can ask questions of your systems from the AI tools they already use.
  • Built once, used by many. One MCP server can serve several AI tools and AI agents, rather than a separate integration for each.
  • A clear boundary. The server decides exactly what the AI can see and do, so you control what each tool has access to.

Doing it well

An MCP server is a door into your systems, so it's designed like any other integration: only the information and actions each job needs, respecting who is allowed to see what, with a record of what was asked and done. Because AI tools act on what they read, connections also need protecting from prompt injection.

How we do it

Our WebPOS AI Framework can include an MCP server for your data, so you and your team can ask questions of your portal's data in plain English, from AI tools such as Claude, and get answers and insight back. If your portal or platform is already built on WebPOS, a data intelligence MCP server is something we can build quickly.

Related

Related terms and services.

  • AI agent

    AI that plans and carries out a task in steps, often through MCP.

  • Prompt injection

    The attack every AI connection to your systems needs protecting from.

  • WebPOS AI Framework

    Our portal framework, with data intelligence through MCP.

  • Integration

    Connecting your storefront, portal and AI to the systems you run.