AI-901 Study Guide
A curated learning path for the AI-901 exam: the best documentation, videos, blog posts and tutorials for every exam objective.A curated learning path for the AI-901 exam: the best documentation, videos, blog posts and tutorials for every exam objective, in the order worth studying them.
Microsoft Certified: Azure AI Fundamentals
Exam at a Glance
AI-901 measures whether you understand what AI can do and can start building with it on Azure: common AI workloads and responsible AI, choosing, deploying and prompting generative models, and using Microsoft Foundry for chat, agents, text, speech, vision and information extraction. More than half of it is about implementing solutions, and it expects you to read Python.
| AI-901 | Microsoft Azure AI Fundamentals |
|---|---|
| Level | Beginner (fundamentals exam) |
| Exam length | 45 minutes |
| Questions | Typically 40 to 60; the number and format mix can vary |
| Passing score | 700 out of 1000 |
| Question formats | Multiple choice, multiple select, drag and drop, build list, hot area |
| Microsoft Learn access | Not available on fundamentals exams |
| Skills measured version | April 15, 2026 |
| Renewal | Not needed: fundamentals certifications do not expire |
What the Exam Covers
The exam has two domains, and the weights are deliberately lopsided: implementing solutions with Microsoft Foundry carries 55 to 60 percent of the marks. Plan your time accordingly: the concepts domain is the easier read, but the Foundry domain is where you need hands-on time.
Identify AI concepts and capabilities — 40-45%
The six responsible AI principles, how generative AI models work, and choosing a model by its capabilities, deployment options and configuration parameters. Then recognizing the common AI workloads: generative and agentic AI, text analysis (keyword extraction, entity detection, sentiment analysis, summarization), speech recognition and synthesis, computer vision and image generation, and extracting information from text, images, audio and video.
Implement AI solutions by using Microsoft Foundry — 55-60%
Writing system and user prompts, deploying a model in the Foundry portal, and building lightweight chat and single-agent clients with the Foundry SDK. Then the workloads in practice: text analysis and speech with multimodal models and Azure Speech in Foundry Tools, interpreting and generating images, and extracting information from documents, forms, images, audio and video with Azure Content Understanding in Foundry Tools.
The full bullet list is in Microsoft's study guide, linked in Step 1 below.
How This Maps to CertiAce Practice Modules
The CertiAce question bank has two practice modules named after the two official learning paths, one per exam domain:
| CertiAce practice module | What you will drill there |
|---|---|
| AI concepts for developers and technology professionals | Responsible AI, generative models, choosing and configuring models, recognizing AI workloads |
| Get started with AI applications and agents on Azure | Prompts, Foundry SDK chat clients, single agents, Azure Language and Speech, multimodal models, Content Understanding |
Where AI-901 Fits

| Certification | What it validates | When to take it |
|---|---|---|
| AZ-900: Azure Fundamentals | Cloud concepts, core Azure services, management and governance | Optional companion, before or alongside AI-901 if Azure itself is new to you |
| AI-901: Azure AI Fundamentals (this exam) | AI concepts and responsible AI, plus building first solutions with Microsoft Foundry | You are here |
| AI-103: Azure AI Apps and Agents Developer Associate | Designing, developing and deploying AI apps and agents with Python and Microsoft Foundry | The natural next step: it deepens exactly the Foundry skills AI-901 introduces |
| AI-200: Azure AI Cloud Developer Associate | Back-end services, scalable architectures and the full development lifecycle of AI solutions on Azure | The next step if you lean toward cloud development rather than AI engineering |
AI-901 has no prerequisites: the certification page lists none, and you do not need AZ-900 first. Nor is AI-901 required by anything above it: Azure AI Apps and Agents Developer Associate (AI-103), Azure AI Cloud Developer Associate (AI-200) and Machine Learning Operations Engineer Associate (AI-300) all have no certification prerequisites either, so you can go straight to whichever fits your direction once you are ready. What AI-901 gives you is the shared vocabulary those exams assume: models and deployments, prompts, agents, Foundry Tools, and the responsible AI principles.
Before You Start
The audience profile is clear: conceptual knowledge of AI on Azure plus the foundational technical skills to work with it. Anything unfamiliar in this table is where to start:
| Area | You should be comfortable with |
|---|---|
| Python basics | Reading simple Python: variables, functions, loops, importing a package, calling a method on a client object; SDK snippets appear in the questions |
| Azure resources | Creating a resource in the Azure portal, what a resource group and a region are, and where a resource's endpoint and key live |
| Cloud concepts | Cloud storage, cloud compute and cloud-based authentication and authorization, at the level the official course recommends (what AZ-900 teaches) |
| APIs and SDKs | What a REST API, an SDK and a CLI are, and the idea of calling a service with an endpoint plus a credential |
| Computing and math basics | General computing concepts and the light math the learning paths assume |
Commonly Confused Pairs
These pairs tend to show up together as the two plausible answers to one question:
| Looks similar | The difference that decides the answer |
|---|---|
| Fairness or inclusiveness | Fairness is about outputs that don't discriminate, for example an admissions system that evaluates every application without bias from irrelevant demographic factors. Inclusiveness is about not excluding users, for example a speech agent that also generates text captions for people with a hearing impairment. |
| Transparency or accountability | Transparency makes users aware of how the system works and of its limitations, for example a bank disclosing that its loan app uses AI and describing the data it was trained on. Accountability means the people and organizations behind an AI solution answer for it, by defining and applying a framework of governance. |
| System message or temperature | The system message sets the model's role, behavior and rules, such as answering only questions about one topic. Temperature, on models that support it, controls how random the output is: higher values give more varied responses, lower values more focused ones. |
| Fast or batch transcription | Fast transcription returns the transcript of a recorded file synchronously and quickly, as for a voicemail. Batch transcription processes large volumes of stored audio asynchronously, as for an archive of recorded calls. Live audio, such as captions during a meeting, needs real-time transcription. |
| Image analyzer or document analyzer | Azure Content Understanding image analyzers extract insights from what a picture shows, such as defects on a product or items on a shelf. They aren't optimized for analysis that rests mainly on extracted text, so reading the fields of a photographed receipt or form calls for document field extraction instead. |
Step-by-Step Study Plan

Treat these estimates as planning guidance and adjust them to your starting knowledge and weekly study hours:
| Your starting point | Suggested prep time |
|---|---|
| You already work with Azure AI services or Microsoft Foundry | 1 to 3 weeks |
| You know Azure, but are new to AI workloads | 3 to 6 weeks |
| New to both Azure and Python | 6 to 8 weeks |
Step 1: Read the Official Study Guide
Read the skills list once and mark every bullet you could not explain to a colleague. The current skills list (April 15, 2026) says Microsoft Foundry, and calls the individual services Azure Speech, Azure Language and Azure Content Understanding "in Foundry Tools". Older materials use the previous names (Azure AI Foundry, Azure AI services), so cross-check older product names and objectives against the current study guide.
Step 2: Schedule Your Exam
Choose a realistic target date after reviewing the skills list and your available study time, then work backwards from it with the prep-time table above.
Certification and Exam Details Page
Step 3: Work Through the Official Learning Path
Complete the two official learning paths on Microsoft Learn: the first covers the concepts domain, the second the Foundry domain, and they are the same two modules you will find in CertiAce practice. Flag anything that needs hands-on practice for Step 5.
AI Concepts for Developers and Technology Professionals
Get Started with AI Applications and Agents on Azure
Step 4: Watch a Full Video Course
Microsoft Learn publishes its own AI-901 course series on YouTube: 14 episodes that mirror the two learning paths, pairing a concepts episode with a get-started-in-Azure episode for each workload. Watch it end to end, pausing to replicate the Foundry actions in your own environment, and later rewatch the episodes behind your weak topics.
AI-901: Introduction to AI in Azure
Step 5: Get Hands-On Practice
More than half of AI-901 is about implementing AI solutions with Microsoft Foundry, so time in the Foundry portal is the highest-value preparation. Microsoft's official exercises are free and self-paced: seven Foundry labs that need an Azure subscription (a free account is enough), plus browser-based concept exercises that need no subscription. Make sure you personally:
- Deploy a model in the Foundry portal, refine your system and user prompts, then chat with it from a lightweight Foundry SDK client
- Create and test a single agent in the portal, then call it from a lightweight client
- Run text analysis and speech recognition or synthesis with Azure Language and Azure Speech in Foundry Tools
- Send an image to a multimodal model, and generate a new image with a generative model
- Extract fields from a document, an image and an audio or video file with Azure Content Understanding
Get started with AI apps and agents in Azure
Step 6: Benchmark Your Knowledge
Use CertiAce to benchmark your readiness module by module. The practice modules map to the exam domains as shown above, so use the module mapping and the topics behind your missed questions to identify the skills-measured areas to revisit. Microsoft also offers a free official practice assessment for AI-901 on AI Skills Navigator (sign in to launch it). It is a useful extra check, but it tends to be somewhat easier than the real exam. CertiAce questions are written to sit closer to the real exam's difficulty. Watch the questions that separate similar workloads, such as image analysis versus image generation, or speech recognition versus speech synthesis. If a module stays weak, read the explanations on the questions you missed, then retry that module.
Official Practice Assessment on AI Skills Navigator
Step 7: Take the Exam
The day before, review only your weak topics. Many AI-901 questions come down to matching a one-line scenario to a workload or a service: creating new content points to generative AI, interpreting existing content points to analysis, and structured fields out of a document point to Content Understanding.
- This is a closed-book exam: unlike the role-based exams, Microsoft Learn is NOT available during fundamentals exams.
- You can take unscheduled breaks, but the clock keeps running and you cannot return to questions you already saw.
- Try the exam sandbox beforehand so the question formats and interface hold no surprises: Exam Sandbox
Additional Learning Resources
| Resource | Type | Why it is useful |
|---|---|---|
| AI-901: Introduction to AI in Azure | YouTube playlist | Microsoft Learn's official 14-episode course series, one concepts and one get-started episode per workload |
| AI-901 Microsoft Azure AI Fundamentals Study Cram | YouTube video | John Savill's one-hour whiteboard cram of the whole syllabus, recorded after the April 2026 skills update |
| Official Practice Assessment (AI Skills Navigator) | Practice | Microsoft's free official practice assessment for AI-901 (sign in to launch it) |
| Get started with AI apps and agents in Azure | Hands-on labs | The seven official Foundry exercises behind the course: portal, agents, text, speech, vision, Content Understanding, Foundry IQ |
| Introduction to AI Concepts | Hands-on labs | Browser-based versions of the same exercises that run a model locally (no Azure subscription needed) |
| Microsoft Foundry documentation | Docs hub | The portal, models, deployments, agents and the SDK the Foundry domain is built on |
| What are Foundry Tools? | Docs | The map of the individual services and their current names. Settle any naming confusion here |
| Azure Content Understanding documentation | Docs | The information-extraction service the exam names explicitly: documents, images, audio and video |
| Azure Speech documentation | Docs | Speech recognition and speech synthesis capabilities and quickstarts |
| Azure Language documentation | Docs | Keyword extraction, entity detection, sentiment analysis and summarization |
| AI Show | Video show | Short Microsoft AI episodes, good for passive learning between study sessions |
| Microsoft Foundry community hub | Community | Microsoft's official Foundry blog and announcements. Watch it for naming and feature changes |
Realistic AI-901 exam-style questions with instant feedback and detailed explanations.
Practice AI-901 now