Discover how AI agents combine language models with tools, knowledge, and memory to carry out tasks. Intro to Foundry Agents AI, Part 8, introduces agent concepts and explores the Microsoft tools available for building agent-based solutions. This course is Part 8 of 8, the final part of the complete Microsoft AI Fundamentals AI-901 class. It builds on the earlier parts of the series and brings your learning journey together with Foundry agents, automation, concluding review lessons, and optional exam preparation using the 4 Quarter Learning System (4QLS).
Begin by comparing AI agents with traditional automation. Learn how agents can handle open-ended, multistep, and ongoing tasks through a cycle of perceiving, reasoning, acting, and observing. Explore the ReAct pattern, short-term and long-term memory, tool connections, and the role of knowledge in supporting useful agent responses. You will also examine how an orchestrator can coordinate specialist agents in a multi-agent system. The Foundry Capability Stack section covers model selection, agent building, and connecting tools, knowledge, memory, and orchestration. It introduces developer workflows, quality evaluation, safety testing, governance, and Foundry IQ.
Next, explore the Microsoft Agent Development Spectrum. The curriculum includes Agent Builder in Microsoft 365 Copilot, a Copilot demonstration, Microsoft Copilot Studio, the Microsoft 365 Agents SDK, Microsoft Foundry Agent Service, and Microsoft Agent Framework. A decision matrix helps you compare development approaches, while lab and simulation lessons reinforce the material.
