AI-901 Exam Syllabus 2026 — Official Domains & Objectives

The AI-901 exam covers two domains: (01) Identify AI Concepts and Responsibilities (40–45%), which covers responsible AI principles, AI model components and configurations, and identifying the right AI workload for a scenario; and (02) Implement AI Solutions by Using Microsoft Foundry (55–60%), which covers building generative AI apps, agents, text and speech solutions, computer vision applications, and information extraction using Microsoft AI Foundry.

AI-901 replaces AI-900 (Microsoft Azure AI Fundamentals), which retires on 30 June 2026. The most significant difference is Domain 02, which is built entirely around Microsoft AI Foundry — a platform that did not exist in the AI-900 era. AI-901 also introduces agentic AI as a topic, reflects the consolidation of Azure AI Services into the Foundry environment, and replaces Azure Document Intelligence with Azure Content Understanding in Foundry Tools. The responsible AI principles from AI-900 carry over into Domain 01.

Microsoft AI Foundry is Microsoft's unified platform for building, deploying, and managing AI applications on Azure. It consolidates model deployment, the Azure AI Services ecosystem, prompt engineering tools, and agent development into a single environment. Domain 02 of AI-901 is structured entirely around Foundry — candidates are expected to know how to use the Foundry portal to deploy models, create agents, and build lightweight AI applications across text, speech, vision, and information extraction workloads.

Domain 01 tests all six of Microsoft's Responsible AI principles: fairness, reliability and safety, privacy and security, inclusiveness, transparency, and accountability. Questions typically present a scenario describing an AI system issue or design decision and ask which principle is most relevant, or what consideration the team should apply. Candidates should be able to distinguish between principles that may seem similar — particularly transparency (explainability of decisions) versus accountability (human ownership of outcomes).

Azure Content Understanding is the service within Microsoft AI Foundry Tools used to extract structured information from documents, forms, images, audio, and video. It is the successor to Azure AI Document Intelligence in the context of the Foundry platform. AI-901 tests Content Understanding across all four media types as a unified capability within Foundry Tools.

The official study guide notes that candidates should have familiarity with Python coding syntax and programming techniques. This reflects the inclusion of lightweight application development tasks in Domain 02 — such as building a chat client using the Foundry SDK or a client application for an agent. Candidates do not need to be experienced developers, but basic familiarity with code structure and Python syntax is expected for the application-building objectives.

Agentic AI refers to AI systems that can autonomously take actions, use tools, and work toward goals — going beyond simple question-and-answer interactions. It appears on AI-901 in both domains: Domain 01 lists agentic AI as one of the common AI workload types candidates must be able to identify, and Domain 02 includes creating and testing a single-agent solution in the Foundry portal as a testable implementation skill.

Starting with Domain 01 builds the conceptual vocabulary — responsible AI principles, workload types, and model selection — that provides useful context when studying the implementation topics in Domain 02. Within Domain 02, starting with the generative AI and agents section establishes the Foundry environment before moving into the more specific workloads of text and speech, computer vision, and information extraction.