Artificial Intelligence is shifting from isolated systems into a connected web of tools and data. The Model Context Protocol (MCP) sits at the centre of this web, a shared language that lets AI systems communicate directly with existing software and services.
Introduced in November 2024, MCP provides a standard way for AI models to connect to tools, data sources, and online systems safely and consistently. Think of it as the digital equivalent of the USB connector, any computer can talk to a mouse, keyboard, or phone with one familiar plug. MCP does the same for AI: it gives models a “plug” to access information or perform tasks using approved connections.
For businesses, this means less isolation between your systems and AI assistants. Instead of copying data around or building custom integrations, your AI tools can securely interact with the software you already use.
This standard is already gaining traction among major business platforms. Xero and NetSuite, for example, both leverage MCP to allow AI agents to interact directly with financial and business data. Strict access control and using your existing permissions lets this happen safely and securely.
As more software adopts MCP, expect AI assistants to become less about conversation and more about action.
This is the beginning of AI systems working with your tools, not around them. The model context protocol is simply the language they speak to get it done.
Next month we will explore the common large language models you can connect to and their recognised strength.
If you’d like to talk about how this could benefit your business, we’d be happy to help.
Michael Graham is a Partner specialising in business systems, operational workflows, and cost accounting. Direct, honest, and analytical, he helps clients optimise their systems, improve operational efficiency, and make informed financial and strategic decisions across diverse industries.
