AI agents want tools, and developers want a simple way to deliver them without managing infrastructure. This session shows how Azure Functions can be used to build a remote MCP server and expose capabilities through the MCP Tool trigger, giving teams a serverless way to make business logic available to MCP clients and agentic applications. Using C# (and not only) and Azure Functions, the session walks through how a function becomes a tool, how discovery and invocation work, and where this approach fits in real-world architectures.
The focus is practical: architecture, code, demos, and design choices that matter when moving from experimentation to production. Attendees will leave with a clear understanding of when Azure Functions is a good fit for MCP workloads, how to structure tools for reuse and scale, which implementation details deserve attention early, and which pitfalls to avoid when exposing serverless functions to AI-driven workflows. The session is aimed at developers and architects who want a concrete, modern pattern for building MCP-enabled solutions on Azure.