AI-900 Exam Overview
Prepare for the Microsoft AI-900 certification exam
with our comprehensive study guide. This study material contains 222 practice questions
sourced from real exams and expert-verified for accuracy. Each question includes the correct answer
and a detailed explanation to help you understand the material thoroughly.
The AI-900 exam — Microsoft Azure AI Fundamentals — is offered by Microsoft.
Passing this exam earns you the Microsoft Certified: Azure AI Fundamentals credential,
an industry-recognized certification that validates your expertise.
Our study materials were last updated on 2026-02-10 to reflect the
most recent exam objectives and content.
About the Microsoft Certified: Azure AI Fundamentals
The Microsoft Certified: Azure AI Fundamentals is awarded by Microsoft
to professionals who demonstrate competence in the skills measured by the AI-900 exam.
According to the
official Microsoft certification page,
this certification validates your ability to work with the technologies covered in the exam objectives.
According to the
Global Knowledge IT Skills and Salary Report,
certified IT professionals earn 15-25% more than their non-certified peers.
Certifications from Microsoft are among the most recognized credentials in the IT industry,
with strong demand across enterprise organizations worldwide.
Free Sample — 15 Practice Questions
Preview 15 of 222 questions from the AI-900 exam.
Try before you buy — purchase the full study guide for all 222 questions with answers and explanations.
Question 1
HOTSPOT
-
Select the answer that correctly completes the sentence.
Show Answer
Correct Answer: natural language processing
Explanation:
Interactive answering of user-entered questions involves understanding and generating human language, which is the core function of natural language processing.
Question 2
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?
A. an Azure Machine Learning model
B. the Translator service
C. a Form Recognizer model
D. Azure Bot Service
Show Answer
Correct Answer: D
Explanation:
Azure Bot Service is designed to build conversational interfaces that let customers query information through multiple channels such as phone (via voice integration), social media platforms, web chat, and digital assistants. The other options focus on translation, document processing, or custom ML models and do not provide a multi-channel conversational interface.
Question 3
HOTSPOT
-
Select the answer that correctly completes the sentence.
Show Answer
Correct Answer: optical character recognition (OCR)
Explanation:
OCR converts scanned images of text, such as newspapers, into machine-readable digital text.
Question 4
HOTSPOT
-
Select the answer that correctly completes the sentence.
Show Answer
Correct Answer: regression
Explanation:
The task predicts a numerical value (hours of overtime) from an input variable (number of orders), which is a regression problem.
Question 5
HOTSPOT
-
You have an app that identifies birds in images. The app performs the following tasks:
• Identifies the location of the birds in the image
• 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.
Show Answer
Correct Answer: Locate the birds: Object detection
Identify the species of the birds: Image classification
Explanation:
Object detection finds and localizes birds within the image using bounding boxes. Image classification determines the species by assigning a label to the detected bird image.
Question 6
HOTSPOT
-
Select the answer that correctly completes the sentence.
Show Answer
Correct Answer: features
Explanation:
In machine learning, input variables provided to a model are called features; labels are outputs, instances are rows/samples, and functions define transformations.
Question 7
You plan to build a conversational AI solution that can be surfaced in Microsoft Teams, Microsoft Cortana, and Amazon Alexa.
Which service should you use?
A. Azure Bot Service
B. Azure Cognitive Search
C. Speech
D. Language service
Show Answer
Correct Answer: A
Explanation:
Azure Bot Service is designed to build conversational bots and expose them across multiple channels. It provides built-in connectors for Microsoft Teams and Cortana, and can integrate with voice assistants like Amazon Alexa via supported channels and adapters. The other options provide supporting capabilities (search, speech-to-text, language understanding) but do not by themselves handle multi-channel conversational bot deployment.
Question 8
HOTSPOT -
Select the answer that correctly completes the sentence.
Show Answer
Correct Answer: Named Entity Recognition (NER)
Explanation:
NER identifies and extracts structured entities such as dates, quantities, and locations from unstructured text.
Question 9
You are developing a chatbot solution in Azure.
Which service should you use to determine a user’s intent?
A. Translator
B. Language
C. Azure Cognitive Search
D. Speech
Show Answer
Correct Answer: B
Explanation:
To determine a user’s intent in an Azure-based chatbot, you use the Azure Language service (formerly LUIS – Language Understanding). This service analyzes natural language text to identify user intents and extract relevant entities. The other options handle translation, search, or speech processing, not intent recognition.
Question 11
Predicting agricultural yields based on weather conditions and soil quality measurements is an example of which type of machine learning model?
A. classification
B. regression
C. clustering
Show Answer
Correct Answer: B
Explanation:
Predicting agricultural yield involves estimating a continuous numerical value (e.g., tons per hectare) from input variables like weather and soil measurements. Models that predict continuous outcomes are regression models, not classification (discrete labels) or clustering (unsupervised grouping).
Question 12
DRAG DROP
-
Match the machine learning models to the appropriate descriptions.
To answer, drag the appropriate model from the column on the left to its description on the right. Each model may be used once, more than once, or not at all.
NOTE: Each correct match is worth one point.
Show Answer
Correct Answer: Regression → A supervised machine learning model used to predict numeric values.
Classification → A supervised machine learning model used to predict categories.
Clustering → An unsupervised machine learning model used to group similar entities based on features.
Explanation:
Regression predicts continuous numerical outputs, classification predicts discrete class labels using labeled data, and clustering groups unlabeled data based on similarity.
Question 13
HOTSPOT
-
Select the answer that correctly completes the sentence.
Show Answer
Correct Answer: an anomaly detection workload.
Explanation:
It involves identifying unusual or abnormal patterns in temperature data from a machine, which is the goal of anomaly detection.
Question 14
Which Azure Cognitive Services service can be used to identify documents that contain sensitive information?
A. Custom Vision
B. Conversational Language Understanding
C. Form Recognizer
Show Answer
Correct Answer: C
Explanation:
Form Recognizer (now called Azure AI Document Intelligence) is designed to analyze documents, extract text, key-value pairs, and structured data, and can be used to identify sensitive information such as PII within documents. The other options focus on images (Custom Vision) or conversational text and intents (Conversational Language Understanding), not document analysis.
Question 15
You have a solution that reads manuscripts in different languages and categorizes the manuscripts based on topic.
Which types of natural language processing (NLP) workloads does the solution use?
A. speech recognition and entity recognition
B. speech recognition and language modeling
C. translation and key phrase extraction
D. translation and sentiment analysis
Show Answer
Correct Answer: C
Explanation:
The solution processes manuscripts (text) in different languages, which requires translation. Categorizing manuscripts by topic involves extracting key themes or topics from the text, which aligns with key phrase extraction. Speech recognition is not applicable because the input is not audio, and sentiment analysis is not required for topic categorization.
Question 16
You have an app that identifies the coordinates of a product in an image of a supermarket shelf.
Which service does the app use?
A. Custom Vision classification
B. Custom Vision object detection
C. Computer Vision Read
D. Computer Vision optical character recognition (OCR)
Show Answer
Correct Answer: B
Explanation:
The app needs to locate a product within an image and return its coordinates. This requires identifying objects and providing bounding boxes. Custom Vision object detection is designed for this purpose, returning the object class, confidence score, and bounding box coordinates. Classification does not return location data, and Computer Vision Read/OCR focus on text, not product objects.