Why Study with PlanetCert?
The Latest Questions
Practice questions and exam topics aligned with the current exam objectives.
Detailed Explanations
Go beyond the answer. Master the material with comprehensive learning and professional explanations for every concept.
AI-Powered Insights
Personalized preparation guidance that adapts to your performance and identifies weak spots automatically.
Exam Information
Official specifications published by Microsoft
Exam Format
Registration
Validity
Skills this practice test covers
- Apply Microsoft's six responsible AI principles to a described AI solution.
- Explain how generative AI models work and choose one by its capabilities.
- Match common AI workloads, from sentiment analysis to object detection, to a scenario.
- Deploy a model in the Foundry portal and test prompts in the playground.
- Build lightweight chat and agent clients with the Foundry SDK.
- Extract information from documents, images, audio and video with Content Understanding.
Explore related topics
AI-901 exam content
2 domains · 37 topics · 150 of 399 questions are written for AI-901
Identify AI concepts and capabilities40-45%63 questions
This domain is the concept half of the exam. Candidates describe the six principles of responsible AI and explain how generative models work and are configured. They also recognize workloads such as text analysis, speech, computer vision and information extraction.
Skills tested
- Apply fairness, reliability and safety, privacy and security, inclusiveness, transparency and accountability to a scenario.
- Explain how a generative model reads a prompt as tokens and produces a completion.
- Choose a model by capability and modality, and set parameters such as temperature and max tokens.
- Tell keyword extraction, entity detection, sentiment analysis and summarization apart.
- Separate speech recognition from speech synthesis, and image classification from object detection.
Questions by topic
- Describe considerations for inclusiveness in an AI solution3
- Describe considerations for transparency in an AI solution3
- Describe considerations for fairness in an AI solution2
- Identify appropriate model deployment options and configuration parameters5
- Describe common text analysis techniques5
- Identify features and capabilities of computer vision and image-generation models4
- Identify scenarios for common AI workloads3
- Identify techniques to extract information from text, images, audio, and videos3
- Describe considerations for privacy and security in an AI solution2
- Identify features and capabilities of speech recognition and speech synthesis3
- Describe considerations for reliability and safety in an AI solution2
- Identify an appropriate AI model, based on capabilities3
- Describe how generative AI models work5
- Describe principles of responsible AI7
- Identify AI model components and configurations5
- Identify AI workloads6
- Describe considerations for accountability in an AI solution2
Implement AI solutions by using Microsoft Foundry55-60%87 questions
The larger domain is hands-on work in Microsoft Foundry. Candidates deploy models in the Foundry portal, write system and user prompts, and build lightweight chat and agent clients with the Foundry SDK. They also add speech, vision and Content Understanding features to small apps.
Skills tested
- Deploy a model in the Foundry portal and try it in the playground.
- Write effective system and user prompts for a generative model.
- Build a lightweight chat client in Python with the Foundry SDK.
- Create and test a single-agent solution, then call it from a client application.
- Extract fields from documents, forms, images, audio and video with Azure Content Understanding.
Questions by topic
- Create effective system and user prompts for generative AI models6
- Create a lightweight chat client application by using the Foundry SDK7
- Create and test a single-agent solution in the Foundry portal5
- Deploy a model and interact with it in the Foundry portal5
- Create a lightweight client application for an agent5
- Build a lightweight application that includes text analysis3
- Respond to spoken prompts by using a deployed multimodal model2
- Build a lightweight application by using Azure Speech in Foundry Tools3
- Create new visual outputs by using generative models2
- Interpret visual input in prompts by using a deployed multimodal model3
- Extract information from audio and video by using Content Understanding3
- Extract information from documents and forms by using Azure Content Understanding in Foundry Tools3
- Build a lightweight application that includes vision capabilities2
- Extract information from images by using Content Understanding2
- Build a lightweight application with information extraction capabilities by using Content Understanding2
- Implement generative AI apps and agents by using Foundry9
- Implement AI solutions for information extraction by using Foundry10
- Implement AI solutions with computer vision and image-generation capabilities by using Foundry8
- Implement AI solutions for text and speech by using Foundry6
- Identify appropriate model deployment options and configuration parameters1
Requirements
- No formal prerequisites: anyone can book the exam.
- Recommended: basic Python, including its syntax and common programming techniques.
- Recommended: experience provisioning Azure resources and a conceptual grasp of AI solutions in Azure.
About the AI-901 exam
AI-901 checks that a beginner can explain core AI ideas and build small AI apps and agents with Microsoft Foundry on Azure. Its first domain covers responsible AI, generative models and common AI workloads. The second, larger domain is hands-on: deploying models, writing prompts and adding speech, vision and information extraction to apps. Passing earns Microsoft Certified: Azure AI Fundamentals, which does not expire.
Most questions in the PlanetCert practice test sit in the Foundry domain, from system prompts to client apps for agents. Flashcard Mode suits the concept domain's lists: the six responsible AI principles and the text analysis techniques. When two tools both seem to fit, such as Azure Speech and a multimodal model for a spoken prompt, the explanation says which one the scenario calls for. Brainy AI Assistant can show what temperature or max tokens changes in a reply, or how Content Understanding pulls fields from a form.
How to earn the certification
- Earns
- Microsoft Certified: Azure AI FundamentalsFundamentals
The certification does not expire.
- After
- AI-102 Designing and Implementing a Microsoft Azure AI Solution
The associate exam for AI engineers who design and build Azure AI solutions.
- DP-100 Designing and Implementing a Data Science Solution on Azure
The associate exam for data scientists who train and deploy machine learning models on Azure.
- AI-102 Designing and Implementing a Microsoft Azure AI Solution
How to prepare for AI-901
- 1
Open a Microsoft Foundry project
Use an Azure subscription to create a Foundry project, then deploy one chat model and one multimodal model. The larger domain assumes these steps are routine.
- 2
Change one setting at a time
In the Foundry portal playground, write a system prompt and a user prompt, then change temperature and max tokens. Note how each change alters the reply.
- 3
Write a small Python client
Build a short chat client with the Foundry SDK that sends a prompt and prints the reply. Then create a single agent in the portal and call it from a second client.
- 4
Give every workload an example
Write one real scenario for each workload, from entity detection to image generation, and one for each responsible AI principle. Drill both lists in Flashcard Mode.
- 5
End with a timed attempt
Take the AI-901 practice test in Timed Mode after the hands-on work. For each miss, ask Brainy AI Assistant why the chosen model, tool or setting did not fit.
Start a timed AI-901run
Microsoft offers three learning paths: Introduction to AI in Azure, AI concepts for developers and technology professionals, and Get started with AI applications and agents on Azure. Its Study guide for Exam AI-901 lists the skills measured.
Who AI-901 is for
- People beginning a career in AI solution development on Azure.
- Developers with basic Python who want to build a first AI app or agent in Microsoft Foundry.
- Students who plan to continue to an associate Azure AI or data science exam.
Typical job titles: Junior AI developer · Software developer · Cloud support engineer · Technical consultant
AI-901 glossary
17 terms from the exam objectives and the vendor's official material, each defined in a sentence.
- Agentic AI
- AI that plans steps and uses tools to reach a goal, instead of only answering a single prompt.
- Azure Content Understanding
- A Foundry tool that extracts structured information from documents, forms, images, audio and video.
- Azure Speech
- The Azure service for speech recognition and speech synthesis, used in apps through Foundry Tools.
- Entity detection
- A text analysis technique that finds named items in text, such as people, places and organizations.
- Foundry portal playground
- The test area in the Foundry portal where a deployed model can be tried with prompts before any code is written.
- Foundry SDK
- The software development kit that lets code, such as a Python chat client, call models and agents deployed in Foundry.
- Foundry Tools
- Ready-made AI services in Foundry, such as Azure Speech and Azure Content Understanding, that apps can call.
- Inclusiveness
- The responsible AI principle that an AI system should work well for people of every ability and background.
- Max tokens
- A setting that limits how many tokens a model may produce in one response.
- Microsoft Foundry
- Microsoft's platform on Azure for building, deploying and running AI models, apps and agents.
- Multimodal model
- A model that accepts more than one kind of input, such as text and images, in the same prompt.
- Object detection
- A computer vision technique that finds objects in an image and marks where each one is.
- Sentiment analysis
- A text analysis technique that judges whether a text expresses a positive, negative or neutral opinion.
- Speech synthesis (text-to-speech)
- Turning written text into spoken audio, for example to read an answer aloud.
- System prompt
- Instructions given to a model before the conversation starts. They set its role, tone and limits.
- Temperature
- A model setting that controls randomness. Lower values give more predictable answers, higher values more varied ones.
- Tokens
- The small pieces of text, such as words or parts of words, that a language model reads and writes.