This course is Part 2 of the eight-part Microsoft AI Fundamentals AI-901 series. It introduces generative AI and agents in Microsoft Azure, building on the AI and Azure foundations covered in Part 1. The focus is understanding the concepts and exploring how Microsoft Foundry supports models and agents. You will begin with what generative AI can produce, including natural language, images, and code. You will explore common applications and limitations, then examine how language models work at an introductory level. Lessons explain transformers, tokens, embeddings, and attention so you can connect the terminology to how a model generates a response. You will also compare large and small language models.
The course then explores prompting and agents. Learn how goals, context, grounding information, and clear expectations improve prompts. Discover how agents combine a model with instructions and tools to retrieve knowledge or perform actions. You will examine retrieval augmented generation (RAG), which supplies relevant information to help ground model responses, and the role of Model Context Protocol (MCP) in connecting agents to tools and data.
In the Azure portion, you will explore Microsoft Foundry, its model catalog, and considerations for choosing and deploying models. Lessons introduce projects, model settings, guardrails, and agent tools. Exercises demonstrate testing prompts and parameters in playgrounds and exploring an agent that uses file search. The final section includes a simulation exercise and a lab exercise, with quiz videos and a summary supporting your review.
