Official blueprint
The AI-900 exam covers 5 domains. Each domain has a defined weight — the percentage of exam questions it contributes. Courseiva has 1,029 practice questions mapped across these objectives.
Click any objective to see practice questions for that domain.
This AI-900 domain covers foundational AI concepts: workloads such as computer vision, natural language processing, document intelligence, and generative AI, plus responsible AI principles like fairness, reliability, privacy, inclusiveness, transparency, and accountability. The exam tests these through scenario-based multiple-choice questions asking you to identify the correct workload type or responsible AI principle for a given business situation.
201 practice questions available
This domain covers core machine learning concepts on Azure: regression, classification, clustering, and the Azure Machine Learning workflow. The exam tests whether you can choose the right model type for a scenario, identify Azure Machine Learning components like compute, datasets, and pipelines, and distinguish automated ML from designer. Expect scenario-based questions rather than deep coding.
209 practice questions available
This domain covers Azure's prebuilt and custom computer vision capabilities: image classification, object detection, OCR, facial detection, and spatial analysis. AI-900 tests it through scenario questions asking which Azure AI Vision or Face service feature fits a described workload, and whether a prebuilt or custom model is appropriate. Expect questions on service selection and capability matching rather than implementation detail.
210 practice questions available
Use this page to practise NAT and PAT questions. Understanding the difference between inside local and inside global addresses — and when each NAT type is appropriate — is the fastest way to eliminate wrong answers.
202 practice questions available
This AI-900 domain covers generative AI concepts and how Azure delivers them, chiefly Azure OpenAI Service and Azure AI Foundry. The exam tests your ability to distinguish generative models from discriminative ones, identify responsible AI concerns like hallucination, and match Azure generative features such as prompts, completions, and deployments to described scenarios.
207 practice questions available
Start with the highest-weight domain — that's where most of your exam questions will come from. Work through each objective's practice questions until you can answer at least 85% correctly. Return to lower-weight domains last, and use the full practice test to test cross-domain recall.