AI-900: Microsoft Azure AI Fundamentals Exam Practice Test
Are you preparing for the AI-900: Microsoft Azure AI Fundamentals certification exam? This foundational-level exam is designed for candidates who want to demonstrate their knowledge of artificial intelligence (AI) concepts and Microsoft Azure AI services. Whether you’re new to AI or starting your journey in cloud-based AI solutions, the AI-900 certification validates your understanding of fundamental AI principles and Azure’s AI offerings.
What Is the AI-900 Certification Exam?
The AI-900 exam assesses your grasp of basic AI concepts and how they are implemented on Microsoft Azure. It’s ideal for students, developers, and business professionals who want to explore AI and machine learning in a cloud environment without requiring deep technical experience. Passing this certification proves your ability to understand AI workloads and services, making it a strong foundation for further specialized Azure certifications.
Microsoft Azure is one of the world’s leading cloud platforms, and its AI services enable organizations to build intelligent applications that process natural language, interpret images, automate decision-making, and much more. The AI-900 exam ensures you understand these capabilities and can leverage them effectively.
What You Will Learn
By preparing for the AI-900 exam, you will gain a solid understanding of:
Fundamental AI concepts: including machine learning, computer vision, natural language processing, and conversational AI.
Core Azure AI services: such as Azure Cognitive Services, Azure Machine Learning, Azure Bot Service, and Azure Cognitive Search.
Building AI-powered applications: learning how to apply Azure’s AI services to solve real-world problems.
AI workloads and considerations: understanding responsible AI principles, fairness, privacy, and security in AI solutions.
Azure AI tools: familiarizing yourself with Azure Machine Learning Studio, Custom Vision, and Form Recognizer for practical AI model creation.
Key Topics Covered in This Practice Exam
The AI-900 practice test on Examsage.com covers all major exam domains, including:
Describe AI workloads and considerations (15-20%)
Overview of AI, types of AI solutions, responsible AI, and ethical AI considerations.Describe fundamental principles of machine learning on Azure (30-35%)
Types of machine learning models, training and evaluating models, regression, classification, and unsupervised learning.Describe features of computer vision workloads on Azure (15-20%)
Image classification, object detection, OCR, video analysis, and Azure Custom Vision.Describe features of Natural Language Processing (NLP) workloads on Azure (15-20%)
Language understanding, text analytics, translation, and conversational AI.Describe features of conversational AI workloads on Azure (15-20%)
Building chatbots with Azure Bot Service and integrating AI services for natural conversations.
Why Choose Examsage.com for AI-900 Exam Preparation?
At Examsage.com, we offer a comprehensive, up-to-date AI-900 practice exam tailored to the latest Microsoft exam objectives. Our questions are crafted by subject matter experts and designed to mimic the style and difficulty of the official test, ensuring you gain real exam readiness.
Detailed explanations: Every question includes clear, concise explanations to reinforce your learning and help you understand complex concepts.
Realistic practice environment: Our platform simulates the actual exam conditions so you can build confidence and manage your time effectively.
Instant scoring and feedback: Track your progress with immediate results and identify areas needing improvement.
Accessible anywhere: Prepare at your own pace, from any device, anytime.
Whether you’re a beginner or looking to validate your AI fundamentals knowledge on Azure, Examsage.com’s AI-900 practice test is your reliable partner for success. Start your AI certification journey today and open doors to exciting career opportunities in cloud computing and AI development.
Sample Questions and Answers
1. What is the primary purpose of Azure Machine Learning?
A. Store big data
B. Create web applications
C. Train, deploy, and manage machine learning models
D. Monitor virtual machines
Answer: C
Explanation: Azure Machine Learning is a cloud-based platform for building, training, and deploying machine learning models efficiently.
2. Which Microsoft tool allows low-code/no-code model building for beginners?
A. Azure CLI
B. Visual Studio Code
C. Azure Machine Learning Designer
D. Azure DevOps
Answer: C
Explanation: Azure Machine Learning Designer is a drag-and-drop tool in the Azure portal used to build ML models with minimal code.
3. What type of AI is used by a chatbot to understand natural language?
A. Computer Vision
B. Machine Learning
C. Natural Language Processing (NLP)
D. Deep Learning
Answer: C
Explanation: NLP helps machines understand and respond to human language in a meaningful way, which is core to chatbots.
4. Which of the following services enables image classification in Azure?
A. Azure Speech Service
B. Azure Form Recognizer
C. Azure Custom Vision
D. Azure Translator
Answer: C
Explanation: Azure Custom Vision allows users to build, deploy, and improve image classifiers that recognize specific content.
5. Which algorithm type is used in supervised learning?
A. Clustering
B. Regression
C. Association Rules
D. Dimensionality Reduction
Answer: B
Explanation: Regression is a common supervised learning technique used for predicting continuous values.
6. Which Azure service provides pre-trained AI models for text analysis?
A. Azure Bot Service
B. Azure Text Analytics
C. Azure Functions
D. Azure Notebooks
Answer: B
Explanation: Azure Text Analytics is part of Azure Cognitive Services and provides pre-trained models for sentiment analysis, key phrase extraction, etc.
7. What is a key benefit of using pre-trained models in Azure?
A. They are always more accurate
B. They eliminate the need for coding
C. They reduce time-to-deployment
D. They always support all languages
Answer: C
Explanation: Pre-trained models can be used immediately, saving time and resources for deployment.
8. Which task is best suited for computer vision?
A. Text translation
B. Sentiment analysis
C. Image recognition
D. Speech synthesis
Answer: C
Explanation: Computer vision enables machines to analyze and interpret visual data from the world.
9. What is the primary purpose of the Azure Bot Service?
A. Generate images
B. Build conversational agents
C. Train ML models
D. Translate documents
Answer: B
Explanation: Azure Bot Service allows developers to create intelligent, conversational bots across multiple channels.
10. Which metric is best for evaluating classification models?
A. Mean Squared Error
B. R-squared
C. Accuracy
D. Euclidean distance
Answer: C
Explanation: Accuracy measures the proportion of correctly predicted instances and is commonly used for classification problems.
11. What is an example of an unsupervised learning task?
A. Image classification
B. Text sentiment analysis
C. Clustering customers
D. Predicting sales
Answer: C
Explanation: Clustering is an unsupervised learning technique used to group data without predefined labels.
12. What is Azure Cognitive Services?
A. A database for AI models
B. A suite of pre-built AI APIs
C. A data visualization tool
D. A programming language
Answer: B
Explanation: Azure Cognitive Services is a set of APIs that enable developers to add AI capabilities to applications without machine learning expertise.
13. What feature does Azure Speech Service not offer?
A. Speech-to-text
B. Text-to-speech
C. Translation
D. Image recognition
Answer: D
Explanation: Azure Speech Service is focused on audio and language capabilities, not image processing.
14. What does Responsible AI in Azure promote?
A. Faster model training
B. Cost-efficient virtual machines
C. Fairness, privacy, and transparency
D. Real-time gaming
Answer: C
Explanation: Responsible AI ensures AI systems are designed with fairness, inclusivity, accountability, and transparency.
15. Which data type is most commonly used for computer vision?
A. Text
B. Audio
C. Image
D. Video
Answer: C
Explanation: Images are the fundamental data type analyzed in computer vision tasks.
16. Which of the following is NOT a component of machine learning?
A. Data
B. Algorithm
C. Model
D. Browser
Answer: D
Explanation: Browser is not related to the core ML workflow which involves data, algorithms, and models.
17. What is model overfitting?
A. Model works too fast
B. Model memorizes training data too well
C. Model underperforms on training data
D. Model requires fewer parameters
Answer: B
Explanation: Overfitting occurs when a model performs well on training data but poorly on new, unseen data.
18. Which Azure AI tool helps label data?
A. Azure ML SDK
B. Azure Data Factory
C. Azure ML Data Labeling
D. Azure Blob Storage
Answer: C
Explanation: Azure Machine Learning Data Labeling is used for annotating datasets used in model training.
19. What’s a benefit of using cloud-based AI solutions?
A. Limited scalability
B. High manual configuration
C. Easy deployment and scalability
D. Only available on Windows OS
Answer: C
Explanation: Cloud-based AI provides flexible scalability and easy access to infrastructure and services.
20. What is inferencing in machine learning?
A. Training the model
B. Tuning hyperparameters
C. Making predictions with a trained model
D. Selecting training data
Answer: C
Explanation: Inferencing is the process of using a trained model to make predictions on new data.
21. Which task would Azure Form Recognizer be best used for?
A. Face detection
B. Reading scanned invoices
C. Translating text
D. Object detection
Answer: B
Explanation: Azure Form Recognizer extracts structured data from forms, receipts, and documents.
22. What’s a key characteristic of reinforcement learning?
A. Pre-labeled data
B. Rewards and penalties
C. Clustering data
D. Real-time graphics
Answer: B
Explanation: Reinforcement learning uses rewards and penalties to learn optimal actions in an environment.
23. Which Azure service allows language translation?
A. Azure Translate
B. Azure Speech Studio
C. Azure Vision API
D. Azure Face API
Answer: A
Explanation: Azure Translator (formerly Microsoft Translator) is used for real-time text translation between languages.
24. What is the output of a classification model?
A. A number
B. A cluster
C. A category label
D. A time series
Answer: C
Explanation: Classification models output discrete labels like “spam” or “not spam.”
25. Why is labeled data important in supervised learning?
A. It reduces processing power
B. It’s used for testing only
C. It trains the model to recognize patterns
D. It eliminates training time
Answer: C
Explanation: Labeled data provides known outputs for the model to learn from and adjust predictions accordingly.
26. What does the AI term ‘bias’ refer to?
A. Faster processing
B. Prediction delays
C. Systematic error in model output
D. Data overuse
Answer: C
Explanation: Bias in AI refers to systematic errors in prediction caused by flawed data or design.
27. Which Azure feature supports model versioning?
A. Azure Notebooks
B. Azure Machine Learning Workspaces
C. Azure Marketplace
D. Azure Event Hubs
Answer: B
Explanation: Azure Machine Learning Workspaces support model tracking, versioning, and management.
28. Which type of AI system mimics human vision?
A. Computer Vision
B. Natural Language Processing
C. Predictive Analytics
D. Anomaly Detection
Answer: A
Explanation: Computer Vision is designed to replicate how humans see and interpret visual information.
29. Which visual interface is used to build AI workflows in Azure?
A. Azure CLI
B. Azure Machine Learning Designer
C. PowerShell
D. Azure Data Explorer
Answer: B
Explanation: Azure ML Designer provides a GUI for building AI workflows using drag-and-drop modules.
30. What does sentiment analysis determine in a text?
A. Length of text
B. Language used
C. Emotional tone
D. Number of characters
Answer: C
Explanation: Sentiment analysis detects whether text expresses positive, negative, or neutral emotions.
31. What is the purpose of the ‘training dataset’ in machine learning?
A. To validate model performance
B. To test prediction accuracy
C. To build and train the model
D. To store backup data
Answer: C
Explanation: The training dataset is used to teach the model to identify patterns and make predictions.
32. Which type of learning involves labeled data?
A. Unsupervised Learning
B. Reinforcement Learning
C. Semi-supervised Learning
D. Supervised Learning
Answer: D
Explanation: Supervised learning requires labeled datasets where the desired output is known for each input.
33. What is one of the main functions of Azure Cognitive Services?
A. Building custom operating systems
B. Providing AI capabilities via APIs
C. Managing virtual machines
D. Encrypting cloud storage
Answer: B
Explanation: Azure Cognitive Services offers APIs that bring AI capabilities to applications without requiring machine learning expertise.
34. Which Azure service helps you automate and orchestrate data workflows for AI pipelines?
A. Azure Synapse Analytics
B. Azure Data Lake
C. Azure Data Factory
D. Azure Cosmos DB
Answer: C
Explanation: Azure Data Factory is used to build and automate ETL workflows necessary for AI/ML pipelines.
35. Which feature of Azure Machine Learning supports automated machine learning (AutoML)?
A. Azure DevOps
B. Designer
C. Automated ML
D. Azure SDK
Answer: C
Explanation: Automated ML simplifies the process of selecting models and tuning hyperparameters automatically.
36. Which Azure Cognitive Service is designed to detect faces and facial attributes?
A. Azure Form Recognizer
B. Azure Face API
C. Azure Text Analytics
D. Azure Language Studio
Answer: B
Explanation: Azure Face API detects faces, identifies people, and evaluates facial attributes such as age and emotion.
37. What is the role of the validation dataset?
A. To correct training errors
B. To evaluate training speed
C. To tune model hyperparameters
D. To replace test data
Answer: C
Explanation: A validation dataset helps fine-tune a model by evaluating its performance during training.
38. What is transfer learning in AI?
A. Transferring data between models
B. Applying an existing model to a new but similar problem
C. Moving models between data centers
D. Encrypting model outputs
Answer: B
Explanation: Transfer learning allows developers to use a pre-trained model and adapt it to a similar but new task.
39. What does the term “model interpretability” mean?
A. The model runs faster
B. The model uses fewer parameters
C. The ability to understand how a model makes decisions
D. The model is less accurate
Answer: C
Explanation: Interpretability refers to how easily humans can understand the decisions or predictions made by a model.
40. Which task can Azure Computer Vision NOT do natively?
A. Read text in images
B. Recognize objects
C. Generate SQL queries
D. Analyze image content
Answer: C
Explanation: Azure Computer Vision cannot generate SQL queries but can read text, recognize objects, and describe images.
41. Which of these best defines artificial intelligence (AI)?
A. Using statistics to predict stock prices
B. A set of rules for software development
C. Machines mimicking human cognitive functions
D. Programming robotic arms
Answer: C
Explanation: AI refers to machines replicating human behaviors such as learning, reasoning, and decision-making.
42. What does “F1 Score” evaluate in model performance?
A. Model speed
B. Trade-off between precision and recall
C. Training time
D. Data size
Answer: B
Explanation: F1 Score balances precision and recall, especially useful for imbalanced datasets.
43. What is a real-world use case of Azure Language Understanding (LUIS)?
A. Translating images
B. Forecasting weather
C. Understanding user intent in a chatbot
D. Generating music
Answer: C
Explanation: LUIS is used to interpret natural language input, identifying intent and entities—ideal for chatbot applications.
44. What is a model’s “hyperparameter”?
A. A fixed value set before training
B. An output value from training
C. A part of input data
D. A GPU setting
Answer: A
Explanation: Hyperparameters are predefined settings such as learning rate or batch size that influence training performance.
45. What kind of data does Azure Form Recognizer work with?
A. Audio
B. Unstructured documents like forms and receipts
C. Video files
D. Graph databases
Answer: B
Explanation: Form Recognizer extracts structured data from unstructured or semi-structured documents like forms.
46. Which Azure AI service supports QnA-style knowledge bases?
A. Azure Bot Framework
B. Azure Question Answering
C. Azure ML Studio
D. Azure Metrics Advisor
Answer: B
Explanation: Azure Question Answering helps you build FAQ-style bots by extracting answers from unstructured content.
47. What is data preprocessing?
A. Storing data in the cloud
B. Cleaning and transforming raw data
C. Converting models to Python
D. Making predictions
Answer: B
Explanation: Data preprocessing involves cleaning, transforming, and organizing data for analysis or model training.
48. In Azure, what is the “workspace” used for in Machine Learning?
A. Creating storage accounts
B. Launching VMs
C. Organizing and managing ML resources
D. Monitoring GPU usage
Answer: C
Explanation: An ML workspace in Azure provides a centralized place to manage experiments, datasets, compute targets, and models.
49. What does “model deployment” mean in AI?
A. Creating a dataset
B. Generating synthetic data
C. Publishing a model for real-time or batch predictions
D. Labeling data
Answer: C
Explanation: Deployment means making the trained model available for use in real-world applications.
50. What type of AI is used for language translation?
A. Reinforcement Learning
B. Regression
C. Natural Language Processing (NLP)
D. Clustering
Answer: C
Explanation: NLP is the field of AI that deals with analyzing and generating human language, including translation.
51. Which Azure Cognitive Service is ideal for recognizing handwriting?
A. Custom Vision
B. Computer Vision Read API
C. Text Analytics
D. Translator
Answer: B
Explanation: The Read API in Azure Computer Vision can detect and extract printed and handwritten text from images.
52. What’s the result of “classification” in machine learning?
A. Time estimates
B. Group membership
C. Quantity prediction
D. Pattern frequency
Answer: B
Explanation: Classification predicts a label or category for an input, such as spam or not spam.
53. Which Azure service helps identify anomalies in time series data?
A. Azure Face API
B. Azure Anomaly Detector
C. Azure Maps
D. Azure Translator
Answer: B
Explanation: Azure Anomaly Detector helps detect irregular patterns in time series data automatically.
54. What is a use case of image segmentation?
A. Separating voice signals
B. Translating languages
C. Identifying object boundaries in images
D. Grouping customer data
Answer: C
Explanation: Image segmentation divides an image into regions or objects, useful in medical imaging, autonomous driving, etc.
55. Which term refers to training a model on new data while retaining past learning?
A. Overfitting
B. Transfer Learning
C. Online Learning
D. Feature Engineering
Answer: C
Explanation: Online learning allows a model to learn incrementally as new data arrives, useful for dynamic environments.
56. Which Azure service enables real-time voice translation?
A. Azure Speech Translation
B. Azure Bot Service
C. Azure Custom Vision
D. Azure Data Explorer
Answer: A
Explanation: Azure Speech Translation offers real-time audio translation from one language to another.
57. Which type of learning uses feedback to improve performance?
A. Reinforcement Learning
B. Supervised Learning
C. Semi-supervised Learning
D. Unsupervised Learning
Answer: A
Explanation: Reinforcement learning improves model performance using feedback through rewards or penalties.
58. What defines “precision” in model evaluation?
A. Ratio of true positives to total predicted positives
B. Ratio of true positives to total negatives
C. Total correct predictions divided by total samples
D. Percentage of false positives
Answer: A
Explanation: Precision measures how many of the model’s positive predictions were actually correct.
59. What is a “pipeline” in Azure Machine Learning?
A. A data stream from IoT
B. A model deployment tool
C. An automated sequence of data processing and model training steps
D. A communication channel
Answer: C
Explanation: A pipeline organizes and automates tasks in the machine learning lifecycle, from preprocessing to deployment.
60. What is the key advantage of Azure’s AI-as-a-Service model?
A. On-premises model hosting
B. Complex setup required
C. Pre-built, scalable, and easy-to-integrate AI features
D. Requires deep ML knowledge
Answer: C
Explanation: AI-as-a-Service on Azure offers pre-built, scalable AI capabilities that are easy to integrate without deep ML expertise.
