Microsoft Azure AI Fundamentals (AI-900)
Get full access to the updated question bank and pass on your first attempt.
Vendor
Microsoft
Certification
AI & Data
Content
270 Qs
Status
Verified
Updated
9 hours ago
Test the Practice Engine
Experience our real exam environment with free demo questions
Premium Bundle
Complete Success Suite
Save $9 Instantly
-
โFull PDF + Interactive Engine Everything you need to pass
-
โAll Advanced Question Types Drag & Drop, Hotspots, Case Studies
-
โPriority 24/7 Expert Support Direct line to certification leads
-
โ90 Days Free Priority Updates Stay current as exams change
Success Metric
98.4% Pass Rate
Standard Simulation
Practice Engine
One-Time Payment
-
Web-Based (Zero Install)
-
Real Testing Environment Virtual & Practice Modes
-
Interactive Engine Drag & Drop, Hotspots
-
60 Days Free Updates
Compatible with All Devices
Basic Tier
PDF Study Guide
Digital Access
- โ Exam Questions (PDF)
- โ Mobile Friendly
- โ 60 Days Updates
Verified 10-Question Preview
Verified Community
The CertoMetrics Standard.
Recommend the #1 platform for verified Microsoft certification resources.
Success Network
Help a Colleague Succeed.
Invite a peer to get their own updated AI-900 prep kit.
Exam Overview
The Microsoft Azure AI Fundamentals (AI-900) certification is your essential first step into the transformative world of Artificial Intelligence on Microsoft Azure. This credential validates your foundational knowledge of common AI workloads and how to implement them using Azure services. Earning AI-900 demonstrates a crucial understanding of machine learning, computer vision, natural language processing, and conversational AI concepts, alongside responsible AI principles. It equips business professionals, developers, and data enthusiasts with the vocabulary and conceptual framework necessary to engage in AI projects, fostering innovation and driving digital transformation. This certification significantly enhances your professional profile, opening doors to advanced Azure AI roles and positioning you as a valuable asset in today's data-driven economy.
Questions
40-60
Passing Score
700/1000
Duration
60 Minutes
Difficulty
Beginner
Level
Fundamentals
Skills Measured
Career Path
Target Roles
Common Questions
Is the material up to date?
Yes. We update our question bank weekly to match the latest Microsoft standards. You get free updates for 90 days.
What format do I get?
You get instant access to both the **PDF** (for reading) and our **Premium Test Engine** (for exam simulation).
Is there a guarantee?
Absolutely. If you fail the AI-900 exam using our materials, we offer a full money-back guarantee.
When do I get the download?
Instantly. The download link is available in your dashboard immediately after payment is confirmed.
Free Study Guide Samples
Previewing updated AI-900 bank (54 Questions).
Select the answer that correctly completes the sentence.
Hot Area:
โ **Object detection **
Reasoning: Object detection identifies and locates multiple distinct objects (items) within a single image, often by drawing bounding boxes around them. This aligns perfectly with "identify multiple types of items in one image."
To complete the sentence, select the appropriate option in the answer area.
Hot Area:
โ ****object detection **
Reasoning: Object detection identifies and locates multiple objects within an image. It outputs bounding boxes around each detected object, providing both its category (e.g., vehicle) and its precise coordinates or location, matching the question's**
โ
Analysis:
.For each of the following statements, select Yes if the statement is true. Otherwise, select No.
NOTE: Each correct selection is worth one point.
Hot Area:
โ Yes for "Chatbots can support voice input."
Reasoning: Modern chatbot frameworks, like Azure Bot Service, integrate with speech services (e.g., Azure AI Speech) to process voice input via speech-to-text and respond with text-to-speech, enabling voice-driven interactions.
โ No for "A separate chatbot is required for each communication channel."
Reasoning: Chatbot platforms allow a single bot to integrate with multiple communication channels (e.g., web, Teams, Slack) using channel connectors. A separate bot per channel is unnecessary and inefficient.
โ Yes for "Chatbots manage conversation flows by using a combination of natural language and constrained option responses."
Reasoning: Chatbots combine Natural Language Understanding (NLU) to interpret free text and structured input methods (buttons, menus) to guide users through flows, ensuring accuracy and improving user experience.
Select the answer that correctly completes the sentence.
Hot Area:
โ **regression. **
Reasoning: Predicting a continuous numerical value like 'hours of overtime' based on input features (number of orders) is the definition of a regression problem in machine learning.
For each of the following statement, select Yes if the statement is true. Otherwise, select No.
NOTE: Each correct selection is worth one point.
Hot Area:
โ **for Statement 1 **
Reasoning: Conversational AI encompasses systems like smart devices that interact with users via natural language, understand queries, and provide relevant responses. A smart device answering appointment questions is a direct application of this technology.
โ **for Statement 2 **
Reasoning: Azure Bot Service is a platform specifically designed to build, deploy, and manage intelligent bots. It enables integration of these bots into various channels, including interactive webchat features on company websites.
โ **for Statement 3 **
Reasoning: Automatically generating captions for pre-recorded videos primarily utilizes speech-to-text AI. While an AI capability, it is not conversational AI, which involves interactive dialogue, intent understanding, and response generation, rather than just transcribing spoken audio.
You have an app that identifies birds in images. The app performs the following tasks:
โข Identifies the location of the birds in the image
โข Identifies the species of the birds in the image
Which type of computer vision does each task use? To answer, select the appropriate options in the answer area.
NOTE: Each correct selection is worth one point.
Hot Area:
โ
Reasoning: Object detection locates individual instances of objects within an image by drawing bounding boxes around them. Identifying the "location of the birds" directly corresponds to this capability, providing precise coordinates for each bird found.
โ
Reasoning: Image classification assigns a descriptive label or category to an image or a specific region of an image. Once a bird is located, identifying its "species" (e.g., sparrow, eagle) is a classification task, assigning a specific type to that identified object.
Select the answer that correctly completes the sentence.
Hot Area:
โ
Reasoning: In machine learning, "features" refer to the individual measurable properties or characteristics that serve as input to a model. These are the independent variables used by the model to learn patterns and make predictions.
You need to provide customers with the ability to query the status of orders by using phones, social media, or digital assistants.
What should you use?
Correct Option: C
โ **Azure AI Bot Service **
Reasoning: Azure AI Bot Service enables building, connecting, and managing intelligent conversational bots. It supports integration with various channels like phones, social media, and digital assistants, making it ideal for querying order status conversationally. โ Why the other choices are incorrect:
- Option A is incorrect: Azure AI Document Intelligence focuses on extracting data from documents, not on providing conversational interfaces for real-time queries.
- Option B is incorrect: Azure Machine Learning is a platform for building, training, and deploying ML models. While a bot might use ML, AML itself is not the service for building the conversational interface.
- Option D is incorrect: Azure AI Translator service is specifically for language translation, not for creating interactive conversational agents to query data.
You have an Azure Machine Learning model that uses clinical data to predict whether a patient has a disease.
You clean and transform the clinical data.
You need to ensure that the accuracy of the model can be verified.
What should you do next?
Correct Option: D
โ **Validate the model by using the clinical data. **
Reasoning: After cleaning and transforming data, the most direct step to "ensure that the accuracy of the model can be verified" is to validate it. Validation involves evaluating the model's performance against data not used for training, thereby verifying its accuracy. โ Why the other choices are incorrect:
- Option A is incorrect: Training the model is part of the process, but it does not, by itself, verify accuracy. You need a separate step to evaluate the trained model's performance.
- Option B is incorrect: Splitting the clinical data into two datasets (e.g., training and testing/validation) is a crucial preparatory step for enabling proper validation. However, the question asks for the next action to verify accuracy, which is the act of validation itself, not just the preparation for it.
- Option C is incorrect: Training the model using automated ML is a method of training, not a method for verifying accuracy. It doesn't directly address the requirement to verify the model's performance.
Which three actions improve the quality of responses returned by a generative Al solution that uses GPT-3.5? Each correct answer presents a complete solution.
NOTE: Each correct answer is worth one point.
Correct Option: B,C,E
โ
Reasoning: Adding grounding data (e.g., factual information from external sources like documents or databases) to prompts significantly improves response quality by providing specific, reliable context, thereby reducing hallucinations and enhancing factual accuracy.
โ
Reasoning: Providing additional examples within prompts (few-shot prompting) guides the model to understand the desired output format, style, and content patterns, leading to more consistent, relevant, and higher-quality responses aligned with user intent.
โ
Reasoning: In the context of prompts, "adding training data" means supplying relevant information or context directly in the prompt. This includes specific facts, documents for analysis, or implicit examples, enabling the model to generate more informed and contextually appropriate responses. โ Why the other choices are incorrect:
- Option A is incorrect: Modifying tokenization is a low-level model architecture or fine-tuning aspect, not a direct action a user takes via prompt engineering to improve response quality of a pre-trained GPT-3.5 model.
- Option D is incorrect: Modifying system messages primarily influences the AI's general persona, tone, and overall behavioral constraints. While important for alignment, it doesn't directly inject specific factual content or examples into a particular query to improve the immediate response's content quality in the same way grounding data or in-context examples do.
Match the Azure OpenAl large language model (LLM) process to the appropriate task.
To answer, drag the appropriate process from the column on the left to its task on the right. Each process may be used once, more than once, or not at all.
NOTE: Each correct match is worth one point.
Select and Place:

Premium Solution Locked
Unlock all 270 answers & explanations
Which parameter should you configure to produce more verbose responses from a chat solution that uses the Azure OpenAl GPT-3.5 model?
Premium Solution Locked
Unlock all 270 answers & explanations
Which parameter should you configure to produce a more diverse range of tokens in the responses from a chat solution that uses the Azure OpenAl GPT-3.5 model?
Premium Solution Locked
Unlock all 270 answers & explanations
You need to count the number of animals in a photograph.
Which type of computer vision should you use?
Premium Solution Locked
Unlock all 270 answers & explanations
For each of the following statements, select Yes if the statement is true. Otherwise, select No.
NOTE: Each correct selection is worth one point.
Hot Area:
Premium Solution Locked
Unlock all 270 answers & explanations
To complete the sentence, select the appropriate option in the answer area.
Hot Area:
Premium Solution Locked
Unlock all 270 answers & explanations
You need to predict the animal population of an area.
Which Azure Machine Learning type should you use?
Premium Solution Locked
Unlock all 270 answers & explanations
To complete the sentence, select the appropriate option in the answer area.
Hot Area:
Premium Solution Locked
Unlock all 270 answers & explanations
Select the answer that correctly completes the sentence.
Hot Area:
Premium Solution Locked
Unlock all 270 answers & explanations
Select the answer that correctly completes the sentence.
Hot Area:
Premium Solution Locked
Unlock all 270 answers & explanations
Select the answer that correctly completes the sentence.
Hot Area:
Premium Solution Locked
Unlock all 270 answers & explanations
Select the answer that correctly completes the sentence.
Hot Area:
Premium Solution Locked
Unlock all 270 answers & explanations
A company employs a team of customer service agents to provide telephone and email support to customers. The company develops a webchat bot to provide automated answers to common customer queries.
Which business benefit should the company expect as a result of creating the webchat bot solution?
Premium Solution Locked
Unlock all 270 answers & explanations
Select the answer that correctly completes the sentence.
Hot Area:
Premium Solution Locked
Unlock all 270 answers & explanations
For a machine learning progress, how should you split data for training and evaluation?
Premium Solution Locked
Unlock all 270 answers & explanations
You have an Azure Machine Learning model that generates a large quantity of false negatives.
You need to reduce the number of false negatives without re-training the model
What should you do?
Premium Solution Locked
Unlock all 270 answers & explanations
You are developing a model to predict events by using classification.
You have a confusion matrix for the model scored on test data as shown in the following exhibit.

Use the drop-down menus to select the answer choice that completes each statement based on the information presented in the graphic.
NOTE: Each correct selection is worth one point.
Hot Area:
Premium Solution Locked
Unlock all 270 answers & explanations
Select the answer that correctly completes the sentence.
Premium Solution Locked
Unlock all 270 answers & explanations
You build a machine learning model by using the automated machine learning user interface (UI). You need to ensure that the model meets the Microsoft transparency principle for responsible AI. What should you do?
Premium Solution Locked
Unlock all 270 answers & explanations
What is a form of unsupervised machine learning?
Premium Solution Locked
Unlock all 270 answers & explanations
For each of the following statements, select Yes if the statement is true. Otherwise, select No.
NOTE: Each correct selection is worth one point.
Hot Area:
Premium Solution Locked
Unlock all 270 answers & explanations
Select the answer that correctly completes the sentence.
Premium Solution Locked
Unlock all 270 answers & explanations
To complete the sentence, select the appropriate option in the answer area.
Hot Area:
Premium Solution Locked
Unlock all 270 answers & explanations
Select the answer that correctly completes the sentence.
Premium Solution Locked
Unlock all 270 answers & explanations
Match the types of AI workloads to the appropriate scenarios.
To answer, drag the appropriate workload type from the column on the left to its scenario on the right. Each workload type may be used once, more than once, or not at all.
NOTE: Each correct selection is worth one point.
Select and Place:
Premium Solution Locked
Unlock all 270 answers & explanations
Select the answer that correctly completes the sentence.
Premium Solution Locked
Unlock all 270 answers & explanations
You are designing an AI system that empowers everyone, including people who have hearing, visual, and other impairments. This is an example of which Microsoft guiding principle for responsible AI?
Premium Solution Locked
Unlock all 270 answers & explanations
You have the following apps:
โข App1: Uses a set of images of tumors to identify whether the tumors are benign or malignant and suggest a treatment
โข App2: Uses images from cameras to track individual livestock as they move around a farm
โข App3: Identifies brands in photographs of billboards
What does each app use? To answer, select the appropriate options in the answer area.
NOTE: Each correct selection is worth one point.โ

Correct Answer:
Premium Solution Locked
Unlock all 270 answers & explanations
Match the Microsoft guiding principles for responsible AI to the appropriate descriptions.
To answer, drag the appropriate principle from the column on the left to its description on the right. Each principle may be used once, more than once, or not at all.
NOTE: Each correct selection is worth one point.
Select and Place:
Premium Solution Locked
Unlock all 270 answers & explanations
Select the answer that correctly completes the sentence.

Correct Answer:
Premium Solution Locked
Unlock all 270 answers & explanations
To complete the sentence, select the appropriate option in the answer area.
Hot Area:
Premium Solution Locked
Unlock all 270 answers & explanations
You plan to develop an Al application that will read the license plates of motor vehicles by using Azure Al Foundry,
What should you use to develop the application?
Premium Solution Locked
Unlock all 270 answers & explanations
You are building an AI system.
Which task should you include to ensure that the service meets the Microsoft transparency principle for responsible AI?
Premium Solution Locked
Unlock all 270 answers & explanations
Match the AI workload to the appropriate task.
To answer, drag the appropriate AI workload from the column on the left to its task on the right- Each workload may be used once,
more than once, or not at all.
NOTE: Each correct match is worth one point.
Premium Solution Locked
Unlock all 270 answers & explanations
Match the types of AI workloads to the appropriate scenarios.
To answer, drag the appropriate workload type from the column on the left to its scenario on the right. Each workload type may be used once, more than once, or not at all.
NOTE: Each correct selection is worth one point.
Select and Place:
Premium Solution Locked
Unlock all 270 answers & explanations
Select the answer that correctly completes the sentence.
Premium Solution Locked
Unlock all 270 answers & explanations
Your company is exploring the use of voice recognition technologies in its smart home devices. The company wants to identify any barriers that might unintentionally leave out specific user groups.
This an example of which Microsoft guiding principle for responsible AI?
Premium Solution Locked
Unlock all 270 answers & explanations
You have an Azure subscription that uses Azure OpenAl in Foundry Models.
You need to create an original image of a rural scene to use on a website,
What should you do?
Premium Solution Locked
Unlock all 270 answers & explanations
What are three Microsoft guiding principles for responsible AI? Each correct answer presents a complete solution.
NOTE: Each correct selection is worth one point.
Premium Solution Locked
Unlock all 270 answers & explanations
To complete the sentence, select the appropriate option in the answer area.
Hot Area:
Premium Solution Locked
Unlock all 270 answers & explanations
To complete the sentence, select the appropriate option in the answer area.
Hot Area:
Premium Solution Locked
Unlock all 270 answers & explanations
For each of the following statements, select Yes if the statement is true. Otherwise, select No.
NOTE: Each correct selection is worth one point.
Premium Solution Locked
Unlock all 270 answers & explanations
You run a charity event that involves posting photos of people wearing sunglasses on Twitter. You need to ensure that you only retweet photos that meet the following requirements:


What should you use to analyze the images?
Premium Solution Locked
Unlock all 270 answers & explanations
When you design an AI system to assess whether loans should be approved, the factors used to make the decision should be explainable. This is an example of which Microsoft guiding principle for responsible AI?
Premium Solution Locked
Unlock all 270 answers & explanations
Full Question Bank Locked
You have reached the end of the free study guide preview. Upgrade now to unlock all 270 questions and the full simulation engine.
Certification Path
Related Certifications
Customer Reviews
Global Community Feedback
David M.
"The practice engine is incredible. It feels exactly like the real testing environment and helped me build so much confidence."
Sarah J.
"The PDF is very well organized and the explanations for the answers are actually helpful, not just random text."
Michael C.
"I was skeptical, but the content is high quality and definitely worth the price. I passed on my first try!"