Large language models are clever, but without a way to use real systems, they are limited in what they can do. The Model Context Protocol (MCP) is an open standard that gives any AI client (like Copilot, ChatGPT, Claude, etc) a single JSON contract to discover and invoke your APIs, enabling AI clients to both collect information on their own and perform actions within your systems.

An MCP server transforms common integration tasks into AI-accessible tools by exposing them through a simple, standardized interface. It can be run locally or hosted remotely, such as in Azure. MCP makes integrations instantly available to a variety of AI tools, allowing seamless interaction between AI clients and your existing APIs, workflows, or services without changing underlying code.

MCP is quickly becoming the standard way to connect AI with integration workflows. You will also be ready to start using MCP in your projects right away. You do not need to know much about AI, only some basics in REST APIs. After this talk you will have a good understanding of what MCP is, how and when to use it, as well as how to implement your own.