"Microsoft Agent Framework: Building Enterprise-Grade AI Agents and Multi-Agent Workflows on .NET and Python"
Enterprise AI agents are moving from impressive prototypes to systems that must be secure, observable, durable, and governable. This book is for experienced .NET and Python developers, solution architects, platform engineers, and AI practitioners who need to build production-ready agent applications with Microsoft Agent Framework rather than isolated demos or fragile prompt chains.
You will learn how the framework unifies agents, deterministic workflows, tools, MCP integrations, sessions, memory, middleware, Azure hosting, and remote interoperability. The book explains when to use autonomous agents, explicit workflow graphs, multi-agent coordination, or plain code; how to select providers and model clients; how to design reliable tools; and how to manage state, checkpointing, durable execution, and Azure AI Foundry deployments.
Written for advanced readers, the book emphasizes architecture, trade-offs, failure modes, and operational discipline across both .NET and Python. It also addresses preview-version realities, migration from Semantic Kernel and AutoGen concepts, governance boundaries, telemetry, debugging, and production-readiness reviews—giving you a practical mental model for building enterprise-grade agent systems that can survive real-world scale, compliance, and reliability demands.
