Microsoft Azure AI Fundamentals Study Guide

You're in the right place. This is the written companion to the AI-900 practice exams, a complete walk through everything the Microsoft Azure AI Fundamentals exam covers, gathered in one place so you can learn the material and not just drill questions.

AI-900 is a fundamentals exam, and knowing what it actually rewards changes how you study. It is not about building models by hand. It is about reading a scenario, recognizing what kind of AI workload it describes, and matching it to the right Azure service, plus knowing Microsoft's responsible-AI principles. The exam spans AI workloads and responsible AI, machine learning fundamentals (supervised versus unsupervised, classification versus regression, and the train, evaluate, and inference loop), computer vision, natural language processing, and generative AI on Azure. Most questions come down to a single decision. Is this classification or regression, given a numeric label versus a category? Do you reach for Azure AI Vision, Custom Vision, or Document Intelligence to read a receipt versus train your own labels versus pull named fields from a form? Get the input-and-output signature right and the answer usually follows.

The guide follows the official domains at the real exam weighting. Each chapter builds the mental model in plain language, separates the look-alike options that the exam loves to test with comparison tables and decision trees, and closes with a cheat sheet you can review the night before. Start at the top for the full path, or pick a domain from the list beside this page and jump straight to what you need.