What you'll learn

Getting an AI agent to work in a demo is easy. Getting it to work reliably, securely, and cost-effectively in an enterprise production environment is a fundamentally different challenge — one that most teams discover only after their proof-of-concept fails to clear the IT, security, or compliance review. In this course, Building Production AI Agents with Microsoft Foundry, you’ll gain the architecture, code patterns, and operational practices needed to deploy AI agents that enterprise environments actually require. First, you’ll explore the Microsoft Foundry platform stack — the Agent Service (built on the OpenAI Responses API), Model Catalog, AI Gateway, and MCP connector ecosystem — and learn to design the production agent architecture that underpins every enterprise deployment, including the new managed Agent Memory for persistent context across sessions. Next, you’ll discover how to orchestrate multi-agent systems using Microsoft Agent Framework’s workflow orchestrations. Finally, you'll learn to tackle complex enterprise workflows, including a live four-agent M&A due diligence pipeline. When you're finished with this course, you'll have the skills and knowledge necessary to extend agent capabilities through two open interoperability protocols: the Model Context Protocol (MCP) for connecting to 1,400+ enterprise tools via Logic Apps, and Agent-to-Agent (A2A) for structured cross-agent communication across runtimes and frameworks.