Hello, you are using an old browser that's unsafe and no longer supported. Please consider updating your browser to a newer version, or downloading a modern browser.
A plain guide to the Microsoft Certified: Azure AI Fundamentals certification as of October 2026, now earned through exam AI-901 after AI-900 retired on June 30, 2026. It covers the two domains and their weights, the hands-on Microsoft Foundry half of the exam, the 45-minute format and passing score, the $99 fee, why it never expires, DoD 8140 status, and the roles it opens the door to.
The Microsoft Certified: Azure AI Fundamentals certification, earned through exam AI-901, is Microsoft's entry-level AI credential. It proves you understand responsible AI, generative models, and the common AI workloads, and that you can build small working solutions with them in Microsoft Foundry.
AI-901 replaced AI-900 when that exam retired on June 30, 2026, and it is a different kind of test. AI-900 asked you to describe AI workloads. AI-901 keeps the concepts, but more than half the weight now sits on doing the work: deploying a model in the Foundry portal, writing system prompts, building a lightweight chat client with the Foundry SDK, testing a single agent, and pulling data out of documents and recordings with Azure Content Understanding. The certification name didn't change, so anyone who earned it through AI-900 keeps it.
There are no formal prerequisites, but Microsoft is clear about what it expects. Its audience profile asks for knowledge of Python coding syntax and programming techniques and familiarity with Azure resources, and the exam page adds that you should be familiar with REST APIs, SDKs, and CLIs. The certification is issued and maintained by Microsoft, and the English skills-measured outline is dated April 15, 2026.
Responsible AI principles, how generative models work, choosing and configuring a model, and recognizing text, speech, vision, agent, and information extraction workloads. 40 to 45%.
Prompts, model deployment, the Foundry SDK, single-agent solutions, speech and vision apps, and Azure Content Understanding. The larger domain at 55 to 60%.
Four things that set Azure AI Fundamentals apart in 2026, starting with how much the exam changed when AI-901 took over from AI-900.
Microsoft kept the name and replaced the exam behind it. AI-900 spread its weight across five describe-the-workload domains. AI-901 cuts that to two and puts 55 to 60 percent on implementing solutions in Microsoft Foundry, from writing prompts and deploying a model to building a chat client with the Foundry SDK and testing a single agent. That is a lot of building for a fundamentals exam.
AI-900 said data science and software engineering experience weren't required. AI-901 asks for knowledge of Python coding syntax and programming techniques. The outline calls its apps lightweight, so this isn't a developer exam, but you should be able to read a short Python sample and say what each line does.
A $99 fee, 45 minutes of exam time, and a credential that never expires or asks for a renewal assessment. Our rundown of the most popular Microsoft certifications in 2026 shows where it sits in the lineup.
AI-901 is not a DoD 8140 credential and does not appear in the Approved Qualifications Matrix V2.1, which lists no Microsoft certifications at all. The current matrix is published at the DoD Cyber Exchange.
Its weight comes from the market instead. Microsoft ties the certification to its AI Engineer job role, and for teams building on Azure it is the quickest way for someone to show they can work in Foundry before they take on an associate exam.
The certification and the AI-901 exam as of October 2026, checked against Microsoft Learn.
Azure AI Fundamentals is the bottom rung of Microsoft's AI path. From here, most people either move up to one of the role-based associate exams Microsoft launched in 2026 or pick up a second fundamentals credential. The three below are the common next steps.
The direct step up on the build side. AI-103 replaced AI-102 on June 30, 2026, and goes much further into Foundry: planning and managing AI resources, building generative apps and multistep reasoning workflows, and developing production-ready agents with multi-agent orchestration. Like other associate certifications, it renews free each year.
The other associate path, for developers who build the cloud services around an AI app. Microsoft describes it as containerized compute, vector-enabled databases, event-driven AI pipelines, serverless functions, secret management, and distributed observability. It took over from AZ-204 when that exam retired on July 31, 2026.
The sideways moves. Azure Fundamentals (AZ-900) fills in the platform basics AI-901 assumes, and Security, Compliance, and Identity Fundamentals (SC-900) covers Microsoft Entra, Defender, and Purview. Both are $99 exams that never expire. See our AZ-900 guide and SC-900 guide.
AI-901 sits at the fundamentals tier of Microsoft's AI path. Above it, the 2026 associate exams split the work between building AI apps and agents and building the cloud services around them, and the path then branches into operations, security, and governance.
The AI groundwork
Hands-on role-based exams
Pick your path
Two questions to answer before you commit: can you sit it, and should you spend the time on AI-901 rather than another fundamentals exam or a role-based one. Here's a straight answer to both.
No formal prerequisites of any kind.
There is no prerequisite credential, no experience requirement, and no application. You register through Pearson VUE, at a test center or online, or through Certiport, and sit the exam. If you fail a first attempt, you can retake it after 24 hours, and each attempt after that needs a 14-day wait. Microsoft also offers a free practice assessment and an exam sandbox, so you can see the question formats before test day.
Reading alone won't cover 55 to 60 percent.
Study will carry you through most of the concept domain. The implementation domain is another matter, because Microsoft expects you to deploy a model, write prompts, build a small chat client with the Foundry SDK, and test an agent, and those skills are hard to fake from notes. Get an Azure subscription with access to the Microsoft Foundry portal, work through Microsoft's Introduction to AI in Azure course (AI-901T00), and run the Python samples yourself until you can explain each one. Older AI-900 study materials won't prepare you for this part, since that exam had no implementation domain.
Azure AI Fundamentals is an entry credential, so it opens doors rather than filling senior AI roles on its own. These are the roles where its mix of AI concepts and hands-on Foundry work does the most for a resume.
An entry seat on a team that builds AI features on Azure. AI-901 shows you have deployed a model in Microsoft Foundry, written system prompts, and called a model from a small Python client, which is the kind of work a new developer gets handed first. AI-103 is the credential to aim for next.
Support roles at Microsoft partners and in enterprise IT, where tickets about model deployments, agents, and Foundry projects sit next to the usual Azure issues. AI-901 gives you the vocabulary and enough time in the portal to work those tickets instead of passing every one up the chain.
Technical sales roles at Microsoft partners and resellers. These teams need people who can explain which model fits a use case, what an agent actually does, and where responsible AI comes in, and who can put together a quick demo in Foundry without calling an engineer.
Analysts and project leads who scope AI work for the business. AI-901 helps you tell a realistic request from a wishful one, write requirements that name the right workload, and raise fairness, privacy, and transparency questions before they turn into rework.
Analysts asked to pull usable data out of documents, forms, images, and recordings. The information extraction skills in AI-901, built on Azure Content Understanding, line up closely with that request, and the exam gives you a structured way to learn them.
Admins who end up owning the Azure resources behind AI projects. AI-901 covers what those resources are for and how models are deployed and configured, and it pairs well with AZ-104 for the day-to-day administration side.
The three certifications people weigh against AI-901 most often: the AI-900 exam it replaced, the Azure platform fundamentals sibling, and the AWS equivalent for AI. Here's how they line up.
| Azure AI Fundamentals (AI-901) | Retired AI-900 | Azure Fundamentals (AZ-900) | AWS AI Practitioner (AIF-C01) | |
|---|---|---|---|---|
| Issuer | Microsoft | Microsoft | Microsoft | Amazon Web Services |
| Status | Current exam | Retired June 30, 2026 | Current exam | Current exam |
| Level | Fundamentals | Fundamentals | Fundamentals | Foundational |
| Focus | AI concepts plus hands-on work in Microsoft Foundry | Describing AI workloads and Azure AI services | Cloud concepts and Azure basics | AI, ML, generative AI and the AWS AI services |
| Domains | 2 | 5 | 3 | 5 |
| Coding | Python syntax expected | Not required | Not required | Out of scope |
| Exam | Count not published, 45 min | Count not published, 45 min | Count not published, 45 min | 65 questions, 90 min |
| Passing Score | 700 of 1000 | 700 of 1000 | 700 of 1000 | 700 of 1000 |
| Cost | $99 | No longer offered | $99 | $100 |
| Validity | Does not expire | Does not expire (holders keep it) | Does not expire | 3 years |
| Best For | Early-career builders on Azure AI | Historical reference only | Beginners on Azure | AI literacy on AWS without coding |
Pricing is US list and varies by country and taxes. None of the Microsoft or AWS certifications above appear on the DoD 8140 Matrix, so their recognition is market-driven. AI-900 and AI-901 lead to the same certification, which is why anyone who passed AI-900 keeps the credential. Microsoft fundamentals certifications do not expire, and AWS certifications are valid for three years. For the AWS side, see our AWS AI Practitioner guide.
Certification Matchups: the published facts, strengths and weak spots of 33 certifications, side by side.
As a Microsoft Authorized Training Partner, our Azure AI Fundamentals boot camp covers both AI-901 domains in a single instructor-led day, in person or live online, with guided hands-on labs in Microsoft Foundry and official Microsoft courseware. Your $99 exam voucher and a free retake are included, so you can sit the exam while the material is fresh.
Where AI-901 sits among this year's AI certifications, how the 2026 Microsoft retirements redrew the path, and what to expect on exam day.
A side by side look at the AI certifications worth knowing this year, from Microsoft AI-901 and AI-103 to the AWS AI Practitioner and the governance and security credentials, with the kind of professional each one suits.
The Microsoft exams employers and training managers ask for most this year, what each one costs, and the 2026 retirements, AI-900 to AI-901 among them, that put older lists out of date.
The full Microsoft map, from the fundamentals tier AI-901 belongs to through the associate and expert levels, with the AI and data track redrawn after the June 2026 retirements.
From MCSE to role-based exams to the 2026 rebuild around Copilot, agents, and Microsoft Foundry, and what a faster retirement clock means when you plan your next exam.
Whether the Azure platform fundamentals exam is worth taking first, and how the AI path from AI-901 to AI-103 fits into the wider Azure lineup.
What the six Azure Copilot agents announced at Ignite 2025 do, how they respect RBAC and compliance guardrails, and what they mean for the Azure certifications.
What happens at check-in, what ends a session early, and how retake waits compare across vendors, including the Microsoft 24-hour and 14-day rules.
The AI-901 exam is organized into two domains, each carrying its own weight. Implementing solutions in Microsoft Foundry is the heavier one at 55 to 60 percent, and it is where the hands-on Python and portal work lives. Click either domain for what it covers.
The concept half of the exam, in three parts. Responsible AI comes first, with the six considerations Microsoft tests by name: fairness, reliability and safety, privacy and security, inclusiveness, transparency, and accountability. Next is how generative AI models work, how to pick a model based on what it can do, and which deployment options and configuration parameters apply. The last part is recognizing workloads, which means matching a scenario to generative or agentic AI, text analysis (keyword extraction, entity detection, sentiment analysis, summarization), speech recognition and synthesis, computer vision and image generation, or information extraction from text, images, audio, and video.
The larger, hands-on half, split across four areas. For generative AI apps and agents, you write system and user prompts, deploy a model and work with it in the Foundry portal, build a lightweight chat client with the Foundry SDK, and create and test a single-agent solution plus a small client for it. For text and speech, you build a text analysis app, answer spoken prompts with a deployed multimodal model, and use Azure Speech in Foundry Tools. For vision, you interpret visual input, generate new images, and build a small vision app. Information extraction closes it out, pulling data from documents, forms, images, audio, and video with Azure Content Understanding in Foundry Tools.
Domains and weights reflect the Microsoft AI-901 skills-measured outline dated April 15, 2026. Localized versions of the exam are updated about eight weeks after the English version, and if AI-901 isn't offered in your preferred language you can request an extra 30 minutes. Microsoft refreshes exams periodically, so confirm the current outline on Microsoft Learn before scheduling.
The questions candidates ask most often when researching the Microsoft Azure AI Fundamentals certification.
The Microsoft Certified: Azure AI Fundamentals certification, earned by passing exam AI-901, is Microsoft's entry-level AI credential. It tests whether you understand responsible AI, how generative models work, and the common AI workloads, and whether you can build small solutions with them in Microsoft Foundry. Microsoft writes it for people at the beginning of a career in AI solution development.
Microsoft retired exam AI-900 on June 30, 2026, and replaced it with AI-901. The certification itself continues under the same name, Microsoft Certified: Azure AI Fundamentals, and AI-901 is now the exam that earns it. If you earned the certification through AI-900, it stays valid on your transcript, because retiring an exam doesn't revoke certifications people already hold.
AI-900 had five domains about describing AI workloads, and Microsoft said data science and software engineering experience weren't required. AI-901 has two domains, and the larger one, at 55 to 60 percent, covers implementing solutions in Microsoft Foundry. Microsoft now expects knowledge of Python coding syntax and familiarity with Azure resources. The standalone machine learning domain is gone, and agents, prompt writing, and information extraction with Azure Content Understanding are in.
There are no formal prerequisites, so anyone can register and sit AI-901. Microsoft does set expectations, though: knowledge of Python coding syntax and programming techniques, familiarity with Azure resources, and some comfort with REST APIs, SDKs, and CLIs. That's a real change from AI-900, which was written for people with no coding background.
The two domains are Identify AI concepts and capabilities (40 to 45 percent) and Implement AI solutions by using Microsoft Foundry (55 to 60 percent). The first covers responsible AI, model selection and configuration, and recognizing AI workloads. The second covers prompts, model deployment, the Foundry SDK, agents, speech and vision apps, and information extraction.
Microsoft does not publish a question count for AI-901, and the number can vary by session. Microsoft gives fundamentals exams 45 minutes of exam time and 65 minutes of seat time, which covers the instructions and the candidate agreement. With more than half the exam built around Foundry tasks, pacing matters more than the word fundamentals suggests.
You need 700 or greater on a scale of 1 to 1000. Microsoft uses scaled scoring, so 700 may not equal 70 percent of the points. The passing line reflects the difficulty of the questions you get, which means an easier set of questions needs more points to pass and a harder set needs fewer.
The AI-901 exam costs $99 USD in the United States, and the price varies by country. It's delivered through Pearson VUE, at a test center or online, or through Certiport. Training Camp's Azure AI Fundamentals boot camp includes the exam voucher in the program fee, along with a free retake.
No. Microsoft certifications don't appear on the DoD 8140 Approved Qualifications Matrix V2.1, so AI-901 won't satisfy a cyber-coded position requirement on its own. Its value in government work is practical instead, for teams that build AI solutions on Azure and need people who know how models and agents are deployed. See the full DoD 8140 work role paths.
No. Microsoft fundamentals certifications, including Azure AI Fundamentals, do not expire and have no renewal assessment. Associate and expert certifications work differently, since they expire each year and renew through a free online assessment. If you earned Azure AI Fundamentals through AI-900 before it retired, that certification still stands.
On its own, AI-901 opens entry-level and AI-adjacent roles rather than senior engineering ones: junior AI developer, cloud support engineer, AI pre-sales or solution specialist, business analyst on AI projects, data analyst working on document and media extraction, and cloud or systems administrator supporting AI workloads. For AI development roles, AI-103 is the associate credential to work toward next.
Only if the cloud itself is new to you. AI-901 assumes you're familiar with Azure resources, and AZ-900 is the $99, 45-minute exam that teaches subscriptions, resource groups, regions, and the core services. If you already work in Azure, skip it, since AZ-900 isn't a prerequisite for AI-901 or anything else. The right call depends on where you start.
It asks more than AI-900 did, mostly because of the Foundry domain. The concept questions reward study, but 55 to 60 percent of the exam covers tasks like deploying a model, writing prompts, building a small chat client with the Foundry SDK, and testing an agent. If you can read Python and have spent real time in the Foundry portal, you're in a far better spot than someone who has only read about it. Microsoft's free practice assessment is a good way to find the gaps.
For people early in a technical career on Azure, yes. It's a $99 exam that takes 45 minutes, never expires, and shows you can deploy and call a model in Microsoft Foundry, not just define one. It also sets you up for AI-103 and AI-200. It's a weaker fit for non-technical staff who want AI literacy without code, and for experienced developers, who'll get more from AI-103 directly.
Deciding whether AI-901 or an associate exam fits your starting point, planning the Microsoft AI path, or arranging AI training for a team: tell us where you are and we'll help you map out the right next step.