AI-901

AI-901 Azure AI Fundamentals Certification Guide

What the AI-901 Microsoft Azure AI Fundamentals exam covers, how it replaced AI-900, who it suits, and how to prepare, with the official study guide.

By Code Nexus team · 5 min read · Published 2026-09-25 · Updated 2026-09-25 · Reviewed by Code Nexus team

Official exam page

Most AI exams are either purely conceptual or aimed at experienced developers. AI-901 Microsoft Azure AI Fundamentals sits in between: it is a fundamentals exam, but one that expects you to describe how to build small AI applications with Microsoft Foundry. This guide covers what changed from its predecessor, what is tested, and how to prepare. All facts about the exam come from Microsoft, starting with the official study guide.

What is the AI-901 certification?

Microsoft describes the candidate as someone at the beginning of their career in AI solution development. You should have conceptual knowledge of AI solutions in Azure and the foundational technical skills to work with them, knowledge of Python coding syntax and programming techniques, and familiarity with Azure resources.

It is the exam that earns the Microsoft Certified: Azure AI Fundamentals certification. Microsoft's AI-900 page is explicit about the change: the AI-900 exam was retired on June 30, 2026 and replaced by AI-901, and candidates must now pass AI-901 to earn the certification. Microsoft did not retire the certification itself, only changed the exam behind it.

For someone heading toward AI engineering, this is the first exam that names the tools directly: generative models, agents, text, speech, vision and information extraction.

What the exam looks like

• Passing score: 700, on a scaled score.
• Skills outline: applies as of April 15, 2026.
• Languages: the exam page lists English, Arabic, Chinese (Simplified and Traditional), French, German, Indonesian, Italian, Japanese, Korean, Portuguese (Brazil), Russian and Spanish.
• Price: based on the country or region where the exam is proctored.
• Question count and time: not stated on the pages we read, so confirm when you book.
• Practice assessment: Microsoft points to a free practice assessment on AI Skills Navigator, which requires you to be signed in.

Microsoft also notes that most questions cover generally available features, that the exam may include commonly used preview features, and that you should be familiar with REST APIs, SDKs and command-line tools.

What AI-901 teaches

The outline has two parts.

The second area is the larger, so more than half the marks are about how you would build something, even though the exam is described as fundamentals. Candidates who only memorise definitions will struggle.

What people underestimate

• The six responsible AI principles. They are easy to list and easy to confuse when applied to a scenario.
• Model configuration. Deployment options and parameters, and how they change behaviour.
• Agents. Creating and testing a single agent, and building a small client for it.
• Information extraction. Content Understanding for documents, images, audio and video.

Who AI-901 is for

• Aspiring AI developers who have some Python and want a structured start.
• Career changers moving into AI from IT, data or software roles.
• Students who want a recognised credential in AI.
• Cloud learners who have completed a fundamentals exam such as AZ-900 and want the AI equivalent.

The benefits

• Current tooling. The outline is built around Microsoft Foundry, generative AI and agents, not older machine learning topics.
• A wide survey. Text, speech, vision, extraction and agents in one exam gives you a map of the field.
• Responsible AI from day one. The principles are examined directly, which is a habit worth building early.
• A stepping stone. It leads naturally to AI-103, where you build and operate these systems.

Is there demand for the skill?

We could not find a reliable, current, independent figure for demand for AI-901 specifically, so we do not quote one. The skills it covers, such as prompting, grounding, agents and responsible AI, are the same ones described in the AI-103 study guide for the role of an Azure AI engineer. Whether a certificate is required for a given job varies, and we have not found evidence that employers ask for AI-901 by name. Treat it as structured proof of learning.

How Code Nexus prepares you for AI-901

Five episodes cover the exam, taught through a company that is deciding where AI belongs in its product.

• Episode 33, Do We Need AI for That? covers AI concepts, when AI is the right tool, and responsible AI.
• Episode 34, The Model Audition covers reading a model card and choosing and configuring a model in a Foundry simulation.
• Episode 35, The Caption Heard Around the Office covers speech and language.
• Episode 36, Give the Feed Eyes covers vision, alt text and structured extraction.
• Episode 37, 8-BIT Needs Boundaries covers building a single, constrained agent and responsible AI, and is the readiness check.

The episodes use a simulated Foundry environment that teaches the workflow and names the real tools the exam uses. It is not your own Azure subscription, so pair it with Microsoft's free learning material and hands-on time in the real Foundry portal. See the curriculum.

Code Nexus prepares you for the exam. Microsoft delivers it and issues the certification.

A study plan

1. Read the current outline on the study guide and ignore material for the retired AI-900.
2. Learn the six responsible AI principles with one example of each.
3. Write prompts for a model in a free playground, and change one parameter at a time.
4. Build one tiny app: a chat client, then a small agent, using the SDK.
5. Try each workload once: text analysis, speech, an image, and a document extraction.
6. Take the practice assessment and study every miss.

What to do next

If you prefer AWS, the parallel exam is AIF-C01, which is conceptual rather than hands-on. If you want to build, continue to AI-103. The whole route is laid out in the certification roadmap, and the reason we teach through tasks is in why practice beats passive study.

Ready to start? Create a free account and open Episode 33.

Code Nexus is not affiliated with Microsoft. Microsoft Azure AI Fundamentals and its badge are trademarks of Microsoft Corporation. Code Nexus prepares you for the exam; Microsoft issues the certification.