Study guide · Exam overview
The retirement date, passing score, and five content domains of Microsoft Certified: Azure AI Fundamentals, plus the ideas the May 2, 2025 study guide asked you to describe.
Exam version notice: Exam AI-900 retired on June 30, 2026, at 11:59 PM Central Standard Time. This overview follows the Microsoft study guide for that exam, skills measured as of May 2, 2025. New candidates earn the same certification by passing AI-901, which tests implementation with Microsoft Foundry and is not covered here. If you already passed AI-900, the fundamentals credential does not expire. Items tagged Extra are useful context that the study guide does not name as a measured skill. General Azure concepts are on the AZ-900 overview.
Microsoft Certified: Azure AI Fundamentals checks whether you can describe common AI workloads, the six responsible AI principles, basic machine learning, and the Azure services used for vision, language, speech, and generative AI. It is not a role-based exam. The May 2, 2025 guide does not ask you to write Python or to deploy a model from memory.
Microsoft described the candidate as someone with a technical or non-technical background. Data science and software engineering experience were not required. Awareness of basic cloud concepts and of client-server applications was useful. The credential can prepare you for Azure Data Scientist Associate or Azure AI Engineer Associate, but it is not a prerequisite for either. You may have been eligible for ACE college credit. Check the current offer with Microsoft.
Figures below come from the AI-900 study guide, the certification page, and Microsoft's fundamentals exam-duration page, reviewed in October 2026.
| Specification | AI-900 |
|---|---|
| Status | Retired on June 30, 2026, at 11:59 PM Central Standard Time. Replaced by AI-901 for new candidates. |
| Level | Beginner. No prerequisite certification. Data science and software engineering experience were not required. |
| Length of test | Fundamentals exams allow 45 minutes of exam time. Seat time, including instructions and the candidate agreement, is 65 minutes. |
| Number of questions | Microsoft did not publish a fixed count. Most certification exams contain about 40 to 60 questions. A practice length is not the live length. |
| Passing score | 700 or greater on a scale of 1 to 1,000. A scaled score is not the percentage of questions answered correctly. |
| Recommended experience | Familiarity with the AI-900 learning material. Basic cloud concepts and client-server applications help. |
| Exam price | Price depended on the country or region where the exam was proctored. |
| Validity | Microsoft fundamentals certifications do not expire. Passing AI-900 already earned the certification. Confirm the current policy with Microsoft. |
| What comes next | AI-901 is the current exam for this certification. It expects basic Python and implementation in Microsoft Foundry. This path does not teach those tasks. |
A blank answer earns no credit. Microsoft exams do not subtract extra points for a wrong answer, so answer every question. If an exam is not offered in your preferred language, you can request an additional 30 minutes. Request accommodations before you sit if you need them.
These ranges are the share of scored content in the study guide dated May 2, 2025. The ranges overlap, so they do not add to 100. The 200 practice questions stay inside each range and give generative AI the largest share: 19%, 19%, 19%, 19%, and 24%.
Name the workload, then name the responsible AI principle the scenario is testing.
Predict a number, a category, or a group. Then say how Azure Machine Learning trains and deploys.
Separate a label for the whole image from a box around an object, and from text read in the image.
Text understanding and speech are different services, even when the sentence looks similar.
The largest domain. A generative model creates new content. A classifier only labels what already exists.
The study guide names workloads and services more often than API parameters. The same map is in the core notes and flashcards.
| Need | Concept to recognize | Remember |
|---|---|---|
| Label an entire photo | Image classification | One label for the image. Object detection also returns where the thing is. |
| Read words in a scan | Optical character recognition | Azure AI Vision can read text. Sentiment is a language task, not a vision task. |
| Find a face, not a name | Facial detection | Detection locates a face. Analysis describes attributes. The skill list does not ask you to identify a person. |
| Predict a price | Regression | A numeric label. Classification predicts a category. Clustering has no labels. |
| Check a model on unseen rows | Validation data | Training data is what the model learns from. Scoring training rows again hides overfitting. |
| Main points of a paragraph | Key phrase extraction | Phrases, not people and places. Named entities are a different language task. |
| Spoken audio to text | Speech recognition | Azure AI Speech. Speech synthesis goes the other way, from text to audio. |
| Write a new answer | Generative AI | Creates content. A classifier only chooses among labels you already defined. |
| OpenAI models in Azure | Azure OpenAI | Chat, embeddings, and image generation, with Azure identity and content filters. |
| Try models from many providers | Azure AI Foundry model catalog | Not limited to OpenAI models. Foundry is the place you explore and build. |
| Extra Harmful text and images | Azure AI Content Safety | Not named in the May 2, 2025 skill list. The guide does ask about responsible AI for generative output. |
| Extra Chat orchestration | Azure Bot Service | Listed among documentation links, not as a measured skill. |
Microsoft says fundamentals exams can include the item types shown in the exam sandbox, and it does not publish the live mix in advance. This site's practice questions use multiple choice, multiple response, matching, and ordering. Matching and ordering are extra practice for the distinctions in the study guide. They are not a claim about the live item mix.
A typical stem asks which workload, principle, technique, or service fits a constraint. Distractors are nearby ideas that solve a different problem. Regression is not classification. Azure AI Face is not Azure AI Vision. Key phrase extraction is not entity recognition. A generative model is not a classifier. Fairness is not privacy.
There is no published extra penalty for a wrong guess. A blank answer earns nothing. On a timed practice exam, answer every item, then use the review to see which domain actually needs work.
A practical sequence for the retired objectives. Use it to learn the ideas, or to review a credential you already earned. Book AI-901 only after you study that exam's own guide.
Say whether the job is vision, language, document processing, or generating new content. Then attach one responsible AI principle.
Regression predicts a number. Classification predicts a category. Clustering groups unlabeled rows. Features are inputs. The label is the target.
Classification, detection, OCR, and faces. Key phrases, entities, sentiment, speech to text, text to speech, and translation.
Azure OpenAI hosts OpenAI models. The model catalog lists more providers. Azure AI Foundry is where you explore, test, and build. Watch for hallucinations.
Use the four decks until you can answer before the card flips. Revisit misses the next day.
Work through original questions, review every explanation, and return to the notes on domains below 75 percent. Aim for a steady 85 percent. Use 45 minutes as the pace of a fundamentals exam.
The overview, core notes, flashcards, and practice exams are all published.
Retirement date, the 700 passing score, domain weights, and a study plan.
Reference notes for workloads, machine learning, vision, language, and generative AI.
40 recall cards in four decks, plus memory notes for principles, models, and services.
200 original questions with custom exams, explanations, and a score report by domain.
No. Microsoft retired exam AI-900 on June 30, 2026, at 11:59 PM Central Standard Time. After that date, new candidates earn Microsoft Certified: Azure AI Fundamentals by passing AI-901. This path covers the retired AI-900 objectives only.
A score of 700 or greater was required to pass. Microsoft reports scores on a scale of 1 to 1,000. A scaled score is not the percentage of questions you answered correctly.
Microsoft lists 45 minutes of exam time for fundamentals exams, and 65 minutes of seat time. The certification page did not publish a fixed question count. Do not treat a practice length as the live exam length.
The study guide dated May 2, 2025 measures five areas: AI workloads and considerations (15–20%), machine learning on Azure (15–20%), computer vision (15–20%), natural language processing (15–20%), and generative AI (20–25%).
Microsoft fundamentals certifications do not expire. Passing AI-900 already earned Microsoft Certified: Azure AI Fundamentals. Confirm the current policy with Microsoft before you rely on it. AI-901 is the exam for candidates who still need to earn the certification.
No. The practice questions on this site are original. They are not Microsoft exam items, and a practice score does not predict a result.
Service names change. The May 2, 2025 guide says Azure AI Foundry. Later AI-901 material uses the name Microsoft Foundry for a broader, hands-on exam. Use the current Microsoft study guide for whichever exam you still need to take.
Useful companions while you study.
Generate strong random passwords with custom length, characters, symbols, and security options.
Convert structured data between JSON and YAML with formatting and validation.
Encode text and files to Base64 or decode Base64 data with UTF-8 support.
Compare two text blocks and highlight added, removed, and changed content.
NodnWebTools provides general informational, educational, and convenience resources. Calculations, conversions, estimates, and learning materials may contain errors or become outdated. Financial, tax, medical, legal, and travel information is not professional advice. Verify important results and current requirements with qualified professionals or authoritative sources. Protect sensitive files and personal information, review each tool’s privacy limitations, and use only content you are authorized to process. You are responsible for how you use and share results. Study resources are independent and do not guarantee exam success or imply certification-provider endorsement. Microsoft, Azure, and related marks are trademarks of Microsoft Corporation. NodnWebTools is not affiliated with, endorsed by, or sponsored by Microsoft.
Nodn