Microsoft Reactor

Discover how to create MCP (Model Context Protocol) tools for AI assistants like GitHub Copilot by building an intelligent code-snippet service on Azure Functions. Learn to implement durable AI agents using Microsoft Agent Framework, orchestrate multi-agent workflows with Durable Functions, and enable semantic search using Azure Cosmos DB vector search powered by OpenAI embeddings.

This session demonstrates how to combine serverless architecture, advanced orchestration, and enterprise-ready AI patterns for scalable, secure solutions.

Technologies Used

• Azure Functions
• Azure Cosmos DB
• Azure OpenAI Service
• Microsoft Agent Framework
• Model Context Protocol (MCP)
• GitHub Copilot

🧠 What You’ll Learn:

• Start small: Build a single agent that answers questions and triggers actions using Azure AI Agent Service and Azure Functions.
• Make it smarter: Add business context and connect to your data with MCP and Azure Cosmos DB.
• Scale up: Orchestrate multiple agents for complex workflows across teams.
• Deploy with confidence: Secure, monitor, and optimize costs on a Flex Consumption plan.