AI-901 Cheat Sheet.
The night-before summary, built like the exam.
Exam Logistics at a Glance
| Detail | Value |
|---|---|
| Passing score | 700 (on a scale of 1-1,000) |
| Job role alignment | AI Engineer |
| Required exams to earn certification | 1 |
| Content domains | 2 |
| Language | English |
| Delivery | Online proctored (OnVUE) or local test center, via Pearson VUE |
| Certification validity | 1 year |
| Retake wait (after 1st fail) | 24 hours |
| Max exams schedulable at once | 2 |
| Advance scheduling window | Up to 90 days out |
| Extra time for non-native English speakers | +30 minutes |
| Guessing penalty | None — incorrect answers are not deducted |
| Price | Varies by country/region of proctoring |
Quick Notes on the Table
- The 700 passing score sits on a 1,000-point scale, so don't confuse it with a raw percentage — it's a scaled score.
- Because there's no penalty for wrong answers, always select an answer for every question rather than leaving one blank.
- A failed first attempt requires only a 24-hour wait before retaking; policy tightens after multiple fails, but that detail isn't part of this cheat sheet's source facts.
Official Content Domains
- Identify AI concepts and capabilities — roughly 40-45% of the exam
- Implement AI solutions by using Microsoft Foundry — roughly 55-60% of the exam
- Foundry implementation is the heavier domain by weight — prioritize hands-on familiarity with Foundry workflows over pure definitions.
- Only two domains total, but the second one carries the majority of scored content, so don't split study time 50/50.
Key Terms and Concepts to Memorize
Domain 1: AI Concepts and Capabilities
- AI workload categories: generative and agentic AI, text analysis, speech, and computer vision.
- Responsible AI principles are explicitly tested within this domain — know the pillars (fairness, reliability/safety, privacy/security, inclusiveness, transparency, accountability) even though exact wording may vary by source.
- Understand the difference between narrow task-specific AI workloads (e.g., classification, object detection) versus generative/agentic workloads that produce novel content or take autonomous actions.
Domain 2: Implementing with Microsoft Foundry
- Generative AI apps and agents — building and orchestrating agentic behavior.
- Text and speech solutions — recognition, synthesis, and language understanding pipelines.
- Computer vision and image generation — both analyzing existing images and generating new ones.
- Information extraction — pulling structured data out of unstructured documents/content.
- Microsoft Foundry is the unifying platform tested throughout this domain — expect scenario questions asking which Foundry capability fits a given business need.
Common Gotchas and Traps
- Don't confuse the 700 passing score with a percentage — it's a scaled score out of 1,000, not \"70%\".
- The exam is aligned to the AI Engineer role, so questions lean toward implementation and application, not just theory — expect \"which service/feature would you use\" scenario questions, especially in the Foundry domain.
- Because the Foundry domain (55-60%) outweighs the concepts domain (40-45%), under-studying Foundry-specific implementation details is the single biggest risk to failing.
- Responsible AI shows up inside the concepts domain, not as its own separate domain — don't skip it thinking it's a minor add-on.
- Remember there's no penalty for guessing: never leave a question unanswered, even under time pressure.
- Pricing is region-dependent — don't assume a flat fee when budgeting or registering; check the current price for your proctoring location before payment.
- Only one exam is required for this certification credential — don't assume you need to stack multiple exams.
Night-Before Checklist
- Confirm your delivery method (online OnVUE proctoring vs. local test center) and double-check the appointment time and time zone.
- If testing online, verify your ID, workspace, and system requirements meet OnVUE proctoring rules in advance.
- Review the two domain weightings one more time — skim Foundry implementation topics last since they carry the larger 55-60% share.
- Do a final pass on responsible AI principles — an easy scoring area if memorized cleanly.
- Remind yourself: answer every question, since there's no penalty for incorrect guesses.
- Know your retake policy going in — a 24-hour wait applies if you need a first retake, so don't panic if you don't pass on the first try.
- Get a full night's sleep — scenario-based questions on Foundry implementation reward clear thinking over memorized trivia.
- Arrive (or log in) early to handle any check-in delays without cutting into your exam time.
Frequently asked questions
What score do I need to pass the Azure AI Fundamentals (AI-901) exam?
You need a score of 700 or greater to pass AI-901. Scores are reported on a scale of 1 to 1,000, so 700 is a scaled score — it does not mean you need to answer 70% of questions correctly. There is also no penalty for guessing: no points are deducted for incorrect answers, so you should answer every question even if you have to make an educated guess.
What topics does the AI-901 exam cover, and how are they weighted?
AI-901 assesses two skill areas: Identify AI concepts and capabilities (40-45%) and Implement AI solutions by using Microsoft Foundry (55-60%). In the concepts area, you must describe principles of responsible AI, identify AI model components and configurations, and identify AI workloads — including generative and agentic AI, text analysis, speech, and computer vision. In the Foundry area, you implement generative AI apps and agents, text and speech solutions, computer vision and image-generation capabilities, and information extraction. Because the Foundry implementation domain carries the majority of the exam weight, hands-on practice with Microsoft Foundry should be the core of your study plan rather than concept memorization alone.
How do I schedule the AI-901 exam, and can I take it online?
AI-901 is scheduled through the exam delivery provider Pearson VUE from the exam or certification details page. You can choose to take it online with online proctored (OnVUE) delivery or in person at a local test center — both formats deliver the same exam, so pick whichever environment suits you. You can schedule certification exams no more than 90 days in advance, and you can have a maximum of two Microsoft Certification exams scheduled at a time through Pearson VUE. The exam price is based on the country or region in which the exam is proctored, so check the exam page for your local fee before registering.
How many exams do I need to earn the certification, and what if I fail my first attempt?
The Azure AI Fundamentals certification requires passing just one exam: AI-901. If you don't pass on your first try, you can retake it 24 hours after the first attempt, so a failed attempt costs you very little time. The exam is offered in English; if your preferred language is unavailable, you can request an additional 30 minutes to complete the exam. The certification is aligned to the AI Engineer job role and targets candidates at the beginning of their career in AI solution development, so it's an appropriate first credential even without prior professional AI experience.
Sources
- 1.Exam AI-901: Microsoft Azure AI Fundamentals — Microsoft (accessed Jul 18, 2026)
- 2.Study guide for Exam AI-901: Microsoft Azure AI Fundamentals — Microsoft (accessed Jul 18, 2026)
- 3.Register and schedule an exam (Pearson VUE) — Microsoft (accessed Jul 18, 2026)
- 4.Exam scoring and score reports — Microsoft (accessed Jul 18, 2026)
- 5.Microsoft Certified: Azure AI Fundamentals — Microsoft (accessed Jul 18, 2026)