If you were planning to take AI-900, the exam you will actually book is AI-901. Microsoft's AI-900 exam page now says: "The AI-900 exam was retired on June 30, 2026, and has been replaced by AI-901."
The good news is that the certification did not change. AI-901 leads to the same Microsoft Certified: Azure AI Fundamentals credential. What changed is the exam, and it changed a lot. AI-900 mostly asked you to describe Azure AI services. AI-901 asks you to use them.
Everything below comes from Microsoft's official AI-901 study guide (skills measured as of April 15, 2026), the AI-901 exam page, and the AI-900 study guide for comparison.
The exam at a glance
- Duration: 45 minutes, the standard length for Microsoft Fundamentals exams, per Microsoft's exam duration page. Allow 65 minutes of seat time.
- Passing score: 700 (scaled)
- Question count: Microsoft does not publish one
- Price: US$99 in the United States. The price varies by country or region.
- Languages: 13, including English, Japanese, Chinese (Simplified and Traditional), Korean, German, French, Spanish, and Portuguese (Brazil)
- Expiry: none, because Fundamentals certifications do not expire
- Practice assessment: free, on Microsoft's AI Skills Navigator (linked from the study guide)
The biggest change: who the exam is for
AI-900 was written for "both technical and non-technical backgrounds" and said data science and software engineering experience were not required.
AI-901's audience profile is different. It targets people "at the beginning of your career in AI solution development" and says you need "knowledge of Python coding syntax and programming techniques" and familiarity with Azure resources. The study guide also says you should be familiar with REST APIs, SDKs, and CLIs.
So AI-901 is still a Fundamentals exam, but it is no longer a no-code exam. If you are in sales, management, or another non-technical role, plan for extra time with basic Python.
Old outline vs new outline
AI-900 had five skill areas of roughly equal weight:
| AI-900 area | Weight |
|---|
| Describe AI workloads and considerations | 15-20% |
| Describe fundamental principles of machine learning on Azure | 15-20% |
| Describe features of computer vision workloads on Azure | 15-20% |
| Describe features of NLP workloads on Azure | 15-20% |
| Describe features of generative AI workloads on Azure | 20-25% |
AI-901 has two functional groups:
| AI-901 functional group | Weight |
|---|
| Identify AI concepts and capabilities | 40-45% |
| Implement AI solutions by using Microsoft Foundry | 55-60% |
Most of the exam now sits in the "Implement" group.
What was removed
The whole machine learning fundamentals area is gone. AI-900 tested regression, classification, and clustering scenarios, features and labels, training and validation data, deep learning and the Transformer architecture, automated machine learning, and Azure Machine Learning capabilities. None of those appear as skills in the AI-901 outline.
If your AI-900 prep leaned on Azure Machine Learning designer or automated ML, you can drop that material.
What stayed
Responsible AI is unchanged. The six bullets (fairness, reliability and safety, privacy and security, inclusiveness, transparency, and accountability) are worded exactly as they were in AI-900.
The workload families are still there. You still need to recognize text analysis, speech, computer vision, and generative AI scenarios. AI-901 adds agentic AI and information extraction to the list, and it names specific text analysis techniques: keyword extraction, entity detection, sentiment analysis, and summarization.
What was added
The concepts group adds a new skill, Identify AI model components and configurations:
- how generative AI models work
- choosing an appropriate model based on its capabilities
- choosing model deployment options and configuration parameters
Then there is the new implementation group, where every skill is something you build in Microsoft Foundry:
- Generative AI apps and agents: write effective system and user prompts, deploy a model and use it in the Foundry portal, build a lightweight chat client with the Foundry SDK, and create and test a single-agent solution plus a client app for it
- Text and speech: build a text analysis app, respond to spoken prompts with a deployed multimodal model, and build an app with Azure Speech in Foundry Tools
- Vision and image generation: interpret images in prompts with a multimodal model, create new visual outputs with generative models, and build an app with vision capabilities
- Information extraction: use Azure Content Understanding in Foundry Tools to pull information from documents and forms, images, audio, and video, and build a small app around it
AI-900 did have a generative AI area, but it asked you to describe Azure AI Foundry, Azure OpenAI, and the model catalog. AI-901 asks you to deploy, prompt, and build.
If you were studying for AI-900
Here is how I would adjust:
- Keep your responsible AI notes and your workload-recognition practice. They transfer directly.
- Drop the machine learning chapter (regression, classification, clustering, Azure Machine Learning, automated ML).
- Add a Foundry project. Deploy a chat model, test it in the playground, and change the system message and temperature to see what happens.
- Write a short Python chat client that keeps conversation history. Then build a single agent in the portal, give it a tool, and call it from code.
- Run real content through the services: sentiment and PII detection on text, speech to text and text to speech, an image and an audio clip sent to a multimodal model, and a document, photo, and short video through Content Understanding.
- Check that any course or practice test you buy says AI-901. Material written for AI-900 will miss more than half of this exam.
How to practice
Start with Microsoft's free practice assessment to get a baseline. Our Azure AI Fundamentals (AI-901) study guide breaks down both functional groups with notes and a four-week plan. The AI-901 practice exam has 40 original scenario questions written to the April 2026 outline, with an explanation for every answer choice and a timed mode.
If you are new to Azure itself, Azure Fundamentals (AZ-900) is still a useful first step, and the full list of Azure exams is on our Azure certifications page.
The short version: the credential is the same, but the exam now expects you to build small AI apps. If you can deploy a model, prompt it, give an agent a tool, and pull fields out of a document with Content Understanding, you are studying the right things.