AI-900 Retired: What Changes with the AI-901 Replacement

Microsoft retired AI-900 on June 30, 2026 and replaced it with AI-901 as the exam requirement for Microsoft Certified: Azure AI Fundamentals. Existing AI-900 notes can still support conceptual review, but a new candidate should use the AI-901 study guide as the scope baseline.

The credential name remains Azure AI Fundamentals. The meaningful change is candidate expectation: AI-901 combines conceptual AI knowledge with foundational technical skills, basic Python familiarity, Azure resources, and implementation tasks centered on Microsoft Foundry.

Microsoft's current study guide allocates 40–45% to identifying AI concepts and capabilities and 55–60% to implementing AI solutions by using Microsoft Foundry. This is a clear signal that service recognition alone is not enough.

The concepts area includes responsible AI, model components and configuration, deployment choices, and workload identification across generative and agentic AI, text, speech, vision, and information extraction. Learn to choose an appropriate capability from a requirement rather than memorizing a product list.

The implementation area includes deploying and interacting with models, creating prompts, building a lightweight chat client with the Foundry SDK, creating and testing a single-agent solution, and building a client that interacts with an agent.

It also covers lightweight text and speech applications, multimodal vision and image generation, and information extraction from documents, images, audio, and video. Practice should therefore include small guided builds, even if your goal is a fundamentals credential.

AI-901 explicitly measures fairness, reliability and safety, privacy and security, inclusiveness, transparency, and accountability. Learn what failure each principle is intended to prevent and which control or design decision supports it.

For a scenario, identify the affected people, data, model behavior, and accountable decision maker. A technically working solution may still be weak if it exposes sensitive data, produces unmonitored harmful outcomes, hides limitations, or excludes part of the intended audience.

Week one: review responsible AI, workload types, generative models, agentic AI, and model-selection concepts. Week two: learn Microsoft Foundry navigation, model deployment, prompt structure, and configuration choices.

Week three: use basic Python and the Foundry SDK in small guided applications for chat, agents, text, speech, and vision. Week four: review Content Understanding, mixed scenarios, and the official objective list, then repeat only the decisions and implementation steps you cannot yet explain.

AI-901 fits early-career learners who want an Azure AI foundation and are willing to work with basic Python and Azure resources. It is less suitable as a purely nontechnical survey than the old AI-900 was often perceived to be.

If you already passed AI-900, you do not need to take AI-901 solely because Microsoft replaced the exam. Azure AI Fundamentals is a Fundamentals certification, and Microsoft says Fundamentals certifications do not expire. Check your Microsoft Learn transcript for your credential record before purchasing another exam.