Microsoft · 2026 Edition
A complete preparation guide written by Microsoft-certified engineers. Covers the exam format,all 5 blueprint domains, a week-by-week study plan, and proven tips for passing first time.
2–4 weeks
Prep time
Beginner
Difficulty
50
Exam questions
700/1000
Pass mark
Exam code
AI-900
Full name
Microsoft Azure AI Fundamentals
Vendor
Microsoft
Duration
60 minutes
Questions
50 items
Passing score
700/1000 (scaled)
Domains covered
5 blueprint domains
Recommended experience
No prerequisites — suitable for beginners to AI and cloud
Typical prep time
2–4 weeks
AI-900 is Microsoft's AI literacy credential. It is valuable for professionals in data, cloud, development, and management roles who need a shared vocabulary for AI projects and Azure AI services.
Job roles this opens
Domain percentage weights are not currently available for this exam. The checklist below is still useful for planning your study.
Week 1
AI Workloads and Responsible AI: principles, use cases, fairness, reliability
Tip: The six Microsoft Responsible AI principles (fairness, reliability/safety, privacy/security, inclusiveness, transparency, accountability) are directly tested. Know each by name and what it means in an AI context.
Week 2
Machine Learning on Azure: supervised vs unsupervised, Azure ML Studio, AutoML
Tip: Know the difference between classification (predict a category), regression (predict a number), and clustering (find groups). Questions give a scenario and ask which ML task type applies.
Week 3
Azure AI Services: Computer Vision, NLP, Document Intelligence, Azure OpenAI
Tip: Azure AI Services are tested by what they do: Computer Vision (analyse images/video), Language (NLP, sentiment), Speech (speech-to-text/text-to-speech), and Azure OpenAI (GPT models). Match the service to the use case.
Week 4
Generative AI: LLMs, prompt engineering, Azure OpenAI, copilots
Tip: Generative AI is a significant addition to AI-900. Know what a large language model (LLM) is, what a foundation model is, and what prompt engineering means. Questions are conceptual — no coding required.
AI-900 is conceptual, not technical. You will not write Python, train models in code, or configure Azure resources — questions test what AI services do and when to use them.
The distinction between AI, machine learning, and deep learning is tested: AI is the broad field, ML is a subset using data-trained models, deep learning is a subset of ML using neural networks.
Know the Azure Machine Learning workspace components at a high level: datasets, experiments, pipelines, models, endpoints. You will not configure them but must identify what each component is used for.
Computer Vision capabilities to know by name: image classification, object detection, optical character recognition (OCR), facial recognition, and spatial analysis. Questions describe an output and ask which capability produced it.
Generative AI on AI-900 covers LLMs, embedding models, image generation, and the concept of grounding responses with your own data (retrieval-augmented generation). These concepts represent a significant portion of the exam.
Apply everything in this guide with adaptive practice questions, detailed answer explanations, and domain analytics.
Deep-dive explanations of the key topics tested on AI-900 — with exam key points and common misconceptions.