AI-900 Microsoft Azure AI Fundamentals Questions and Answers

As outlined in the AI-900 study guide and Microsoft Learn’s “Explore fundamental principles of machine learning” module, clustering is a core example of unsupervised machine learning.

In unsupervised learning, the model is trained on data without labeled outcomes. The goal is to discover patterns or groupings naturally present in the data. Clustering algorithms, such as K-means, DBSCAN, or Hierarchical clustering, analyze similarities among data points and group them into clusters. For example, clustering can group customers by purchasing behavior or segment products by shared characteristics — all without predefined labels.

Supervised learning, by contrast, uses labeled data (input-output pairs) to train a model that predicts outcomes. This includes:

A. Multiclass classification – Predicts more than two categories (e.g., classifying images as dog, cat, or bird).

C. Binary classification – Predicts two categories (e.g., spam vs. not spam).

D. Regression – Predicts continuous numeric values (e.g., price prediction).

Therefore, the only option representing unsupervised learning is clustering, which enables data discovery without predefined labels.

The described app identifies the coordinates of a product within an image of a supermarket shelf. This scenario directly corresponds to the object detection capability of Azure AI Custom Vision. As per the Microsoft Learn module “Train a Custom Vision model”, the object detection project type allows developers to train models that can both detect and locate objects within an image. It returns bounding box coordinates along with predicted labels for each detected item.

In this use case, the app doesn’t just classify what products are present—it needs the position of the product (coordinates). That function distinguishes object detection from classification. Classification simply assigns a label to the entire image, while object detection provides spatial information for multiple items in one image.

The other options are incorrect:

B. Azure AI Computer Vision Read and C. Azure AI Computer Vision OCR are used for extracting text from images, not locating objects.

D. Azure AI Custom Vision classification only categorizes images but cannot determine where objects appear.

Therefore, to build an app that finds and locates products in images, the correct choice is A. Azure AI Custom Vision object detection.

This question is drawn from the Microsoft Azure AI Fundamentals (AI-900) syllabus section “Describe features of natural language processing (NLP) workloads on Azure.” According to the Microsoft Learn materials, Natural Language Processing (NLP) is a branch of artificial intelligence that allows computers to analyze, understand, and generate human language. NLP enables machines to work with text or speech data in a way that extracts meaning, sentiment, and intent.

Microsoft defines NLP as enabling scenarios such as language detection, text classification, key phrase extraction, sentiment analysis, and named entity recognition. The example given—classifying emails as “work-related” or “personal”—is a text classification task, which falls under NLP capabilities. The AI model processes the textual content of emails, identifies linguistic patterns, and categorizes them based on the detected topic or context.

Let’s analyze the other options:

Predict the number of future car rentals → This is a forecasting task, handled by machine learning regression models, not NLP.

Predict which website visitors will make a transaction → This is a classification or prediction problem in machine learning, not NLP, since it deals with behavioral or numerical data rather than language.

Stop a process in a factory when extremely high temperatures are registered → This is an IoT or anomaly detection scenario, focusing on sensor data, not language understanding.

Therefore, only classifying email messages as work-related or personal correctly represents an NLP use case. It illustrates how NLP can analyze written text and make intelligent categorizations—a key capability covered in AI-900’s natural language workloads section.

✅ Yes – Extract key phrases

❌ No – Generate press releases

✅ Yes – Detect sentiment

The Azure AI Language service is a powerful set of natural language processing (NLP) tools within Azure Cognitive Services, designed to analyze, understand, and interpret human language in text form. According to the Microsoft Azure AI Fundamentals (AI-900) study guide and Microsoft Learn documentation, this service includes several capabilities such as key phrase extraction, sentiment analysis, language detection, named entity recognition (NER), and question answering.

Extract key phrases from documents → YesThe Key Phrase Extraction feature identifies the most relevant words or short phrases within a document, helping summarize important topics. This is useful for indexing, summarizing, or organizing content. For instance, from “Azure AI Language helps analyze customer feedback,” it may extract “Azure AI Language” and “customer feedback” as key phrases.

Generate press releases based on user prompts → NoThis functionality falls under generative AI, specifically within Azure OpenAI Service, which uses models such as GPT-4 for text creation. The Azure AI Language service focuses on analyzing and understanding existing text, not generating new content like press releases or articles.

Build a social media feed analyzer to detect sentiment → YesThe Sentiment Analysis capability determines the emotional tone (positive, neutral, negative, or mixed) of text data, making it ideal for analyzing social media posts, reviews, or feedback. Businesses often use this to gauge customer satisfaction or brand reputation.

In summary, the Azure AI Language service analyzes text to extract insights and detect sentiment but does not generate new textual content.

According to the Microsoft Azure AI Fundamentals (AI-900) Official Study Guide and the Microsoft Learn module “Explore fundamental principles of machine learning,” anomaly detection is a specialized machine learning technique used to identify data points, patterns, or events that deviate significantly from normal behavior.

Anomaly detection is widely used for monitoring time-series data and detecting unexpected or rare occurrences that may indicate problems, opportunities, or fraud. For example:

Detecting fraudulent transactions in banking systems.

Identifying equipment malfunctions in industrial IoT applications.

Monitoring network intrusions in cybersecurity.

Detecting unexpected spikes or drops in web traffic or sales.

In Azure, this workload is supported by the Azure AI Anomaly Detector service, which uses statistical and machine learning algorithms to learn from historical data and establish a baseline of normal behavior. When the system detects data points that fall outside expected patterns, it flags them as anomalies.

Let’s evaluate the incorrect options:

A. A machine learning technique that understands written and spoken language → This describes Natural Language Processing (NLP), not anomaly detection.

B. A machine learning technique that classifies objects based on user-supplied images → This refers to image classification, typically using computer vision.

D. A machine learning technique that classifies images based on their contents → Also describes computer vision, not anomaly detection.

Therefore, the correct answer is C, since anomaly detection specifically refers to analyzing data over time and identifying unusual or abnormal patterns that differ from the expected trend.

Yes, Yes, and No.

According to the Microsoft Azure AI Fundamentals (AI-900) official study guide and Microsoft Learn modules under the topic “Describe features of common AI workloads”, conversational AI solutions like chatbots are used to automate and enhance customer interactions. A chatbot is an AI service capable of understanding user inputs (text or voice) and providing appropriate responses, often integrated into websites, mobile apps, or messaging platforms.

A restaurant can use a chatbot to empower customers to make reservations using a website or an app – Yes.This statement is true because conversational AI is designed to handle structured tasks such as booking, scheduling, and information retrieval. Chatbots built with Azure Bot Service can connect to backend systems (like a reservation database) to let customers make or modify reservations through a chat interface. The AI-900 study guide explicitly notes that chatbots can help businesses “automate processes such as booking or reservations” to improve efficiency and customer experience.

A restaurant can use a chatbot to answer inquiries about business hours from a webpage – Yes.This is also true. Chatbots can be trained using QnA Maker (now integrated into Azure AI Language) or Azure Cognitive Services for Language to answer common customer questions. FAQs such as opening hours, menu details, and directions are ideal for chatbot automation, as outlined in the AI-900 modules discussing customer support automation.

A restaurant can use a chatbot to automate responses to customer reviews on an external website – No.This is not a typical chatbot use case taught in AI-900. Chatbots are meant for direct interactions within controlled channels, such as a company’s own website or messaging app. Managing and posting responses to reviews on external platforms (like Yelp or Google Reviews) would involve policy restrictions, authentication issues, and reputational risk. The AI-900 course specifies that responsible AI usage requires maintaining human oversight in public-facing communications that influence brand image.

Microsoft’s Responsible AI principles are the foundation for developing and deploying ethical and trustworthy AI systems. The six key principles are Fairness, Reliability and Safety, Privacy and Security, Inclusiveness, Transparency, and Accountability. Each principle guides specific practices for ensuring AI systems operate responsibly in real-world applications like automated insurance quoting systems.

Transparency – This principle ensures that the AI’s decisions can be understood and explained. Recording the decision-making process and enabling staff to trace how a quote was generated aligns with transparency. It allows stakeholders to interpret the reasoning behind model outputs, ensuring that the AI behaves predictably and ethically.

Privacy and Security – This principle focuses on protecting personal data and ensuring that sensitive information is handled responsibly. Limiting access to customer data only to authorized personnel maintains compliance with privacy laws (like GDPR) and safeguards against misuse. Microsoft emphasizes that AI systems should maintain strict control over data visibility and integrity.

Inclusiveness – This principle ensures that AI systems are accessible to all users, including people with disabilities. By supporting screen readers and assistive technologies, the system ensures equal access to information and services for every customer. Inclusiveness prevents discrimination and promotes accessibility, both of which are central to Microsoft’s Responsible AI strategy.

Thus, the correct mapping of principles is:

Decision process → Transparency

Personal information visibility → Privacy and Security

Accessibility via screen readers → Inclusiveness.

The correct answer is “An embedding.”

In the context of large language models (LLMs) such as GPT-3, GPT-3.5, or GPT-4, an embedding refers to a multi-dimensional numeric vector representation assigned to each word, token, or phrase. According to the Microsoft Azure AI Fundamentals (AI-900) study guide and Microsoft Learn documentation for Azure OpenAI embeddings, embeddings are used to represent textual or semantic meaning in a numerical form that a machine learning model can process mathematically.

Each embedding captures the semantic relationships between words. Words or tokens with similar meanings (for example, “car” and “automobile”) are represented by vectors that are close together in the multi-dimensional space, while unrelated words (like “tree” and “laptop”) are farther apart. This vector representation enables the model to understand context, similarity, and relationships between different pieces of text.

Embeddings are fundamental in tasks such as:

Semantic search: Finding documents or sentences with similar meaning.

Clustering: Grouping related concepts together.

Recommendation systems: Suggesting similar content based on text meaning.

Contextual understanding: Helping generative models produce coherent and context-aware text.

Option review:

Attention: A mechanism used within transformers to focus on relevant parts of input sequences but not a representation of words.

A completion: Refers to the generated text output from a model, not the internal representation.

A transformer: The architecture that powers models like GPT, not the vector representation of tokens.

Therefore, the correct term for a multi-dimensional vector assigned to each word or token in a large language model (LLM) is An embedding, which represents how meaning is numerically encoded and compared within language models.

In the Microsoft Azure AI Fundamentals (AI-900) curriculum, language modeling is described as a core component of Natural Language Processing (NLP) that enables an AI system to understand, interpret, and generate human language in context. The AI-900 Microsoft Learn module “Identify features of Natural Language Processing workloads” explains that language modeling is used to analyze user input and determine intent — essential for conversational systems like chatbots.

In this question, the chatbot must:

Accept customer orders.

Retrieve support documents.

Retrieve order status updates.

All these tasks require the bot to understand user intent and context from text input. This understanding process is driven by language modeling, which predicts meaning and structure within sentences, enabling the system to decide what action to take next.

Microsoft Learn distinguishes between various NLP techniques:

Sentiment analysis detects emotional tone (positive/negative/neutral).

Translation converts text between languages.

Named Entity Recognition (NER) identifies specific entities like names or dates.However, none of these individually allow a system to process commands, requests, or user intents — that capability is part of language modeling, which powers LUIS (Language Understanding Intelligent Service) or the modern Azure Cognitive Service for Language.

Therefore, to build a chatbot that can interpret commands and respond contextually — such as processing orders or retrieving documents — you must use language modeling.

According to the Microsoft Azure AI Fundamentals (AI-900) study guide, the three main types of supervised and unsupervised machine learning models—classification, clustering, and regression—are used for distinct problem types depending on the structure of the data and the prediction goal.

Clustering is an unsupervised learning technique used when the goal is to group items with similar characteristics without predefined labels. In this scenario, “Assign categories to passengers based on demographic data” implies automatically grouping passengers based on patterns such as age, income, or travel frequency, without any prior labeling. This directly maps to clustering, which discovers hidden groupings (for example, segmenting customers into categories like business travelers or vacationers).

Regression is a supervised learning method used to predict continuous numerical values. The scenario “Predict the amount of consumed fuel based on flight distance” is a classic regression problem because the output (fuel consumption) is a continuous variable dependent on another continuous variable (distance). Regression models, such as linear regression, are trained to estimate numeric outputs.

Classification is also a supervised learning approach, but it predicts discrete categories or outcomes. The scenario “Predict whether a passenger will miss their flight based on demographic data” involves a binary decision (missed or not missed), which is typical of classification tasks. These models learn from labeled examples to assign new instances to specific categories.

In summary, Clustering groups similar passengers, Regression predicts continuous numerical outcomes, and Classification determines categorical outcomes. This alignment precisely matches the definitions in Microsoft’s AI-900 learning objectives under “Describe common machine learning types and scenarios.”

The correct answer is B. Optical character recognition (OCR).

OCR is a key capability within the Computer Vision and Document Intelligence services in Azure AI that enables systems to detect and extract printed or handwritten text from images and scanned documents.

When capturing text from images, OCR technology analyzes visual patterns (shapes of letters and numbers) and converts them into machine-readable text. For example, a photo of a receipt, street sign, or printed report can be processed to extract textual content programmatically.

A (Text analysis): Applies to NLP tasks such as sentiment detection or key phrase extraction, not image processing.

C (Image description): Generates captions describing the scene or objects in an image.

D (Object detection): Identifies and locates objects but does not extract text.

According to the Microsoft Azure AI Fundamentals (AI-900) Official Study Guide and the Microsoft Learn module “Explore natural language processing (NLP) in Azure,” sentiment analysis is a core natural language processing (NLP) workload used to determine the emotional tone or attitude expressed in a piece of text. It helps identify whether a statement, review, or comment conveys a positive, negative, neutral, or mixed sentiment.

In this question, the scenario involves analyzing customer reviews and determining how positive or negative each review is. This directly aligns with sentiment analysis, which evaluates subjective text and quantifies the expressed opinion. In Azure, this workload is implemented through the Azure AI Language service (formerly Text Analytics API), where the Sentiment Analysis feature assigns a sentiment score to text inputs and classifies them accordingly.

For example:

“I love this product!” → Positive sentiment

“It’s okay, but could be better.” → Neutral or mixed sentiment

“I’m disappointed with the service.” → Negative sentiment

Let’s analyze why the other options are incorrect:

A. Language detection: Identifies which language (e.g., English, Spanish, French) the text is written in. It doesn’t measure positivity or negativity.

C. Key phrase extraction: Identifies the main topics or keywords in text (e.g., “battery life,” “customer support”), not the emotion.

D. Entity recognition: Detects and categorizes specific entities such as people, locations, organizations, or dates within the text.

Therefore, based on Microsoft’s AI-900 syllabus and Azure AI Language documentation, the workload that analyzes text to determine positive or negative opinions is Sentiment Analysis (Option B). This capability is widely used in customer feedback analysis, brand monitoring, and social media analytics to understand public perception and improve business decisions.

The correct matches are based on the Microsoft Azure AI Fundamentals (AI-900) Official Study Guide and the Microsoft Learn module “Explore computer vision in Microsoft Azure.” These materials explain that facial recognition tasks can be categorized into four major operations: verification, identification, similarity, and grouping. Each task serves a distinct purpose in facial recognition scenarios.

Verification – “Do two images of a face belong to the same person?”The verification task determines whether two facial images represent the same individual. Azure Face API compares the facial features and returns a confidence score indicating the likelihood that the two faces belong to the same person.

Similarity – “Does this person look like other people?”The similarity task compares a face against a collection of faces to find visually similar individuals. It does not confirm identity but measures how closely two or more faces resemble each other.

Grouping – “Do all the faces belong together?”Grouping organizes a set of unknown faces into clusters based on similar facial features. This is used when identities are not known beforehand, helping discover potential duplicates or visually similar clusters within an image dataset.

Identification – “Who is this person in this group of people?”The identification task is used when the system tries to determine who a specific person is by comparing their face against a known collection (face database or gallery). It returns the identity that best matches the input face.

According to Microsoft’s AI-900 training, these tasks form the basis of Azure Face API’s capabilities. Each helps solve a different type of facial recognition problem—from matching pairs to discovering unknown identities—making them essential components of responsible AI-based vision systems.

This question maps each Azure AI service to its correct capability based on the Microsoft Azure AI Fundamentals (AI-900) syllabus and Microsoft Learn documentation on Azure Cognitive Services.

Classify and label images → Azure AI VisionAzure AI Vision (formerly Computer Vision) provides capabilities to analyze visual content, detect objects, classify images, and extract information from pictures. It includes object detection, image classification, and tagging, which are core vision tasks. This service enables businesses to build solutions that understand visual input, such as identifying products, reading signs, or detecting faces in images.

Generate conversational responses → Azure OpenAI ServiceAzure OpenAI Service integrates powerful large language models such as GPT-3.5 and GPT-4, capable of generating human-like text responses, summarizations, translations, and dialogues. These models are designed for natural language generation (NLG) and conversational AI, making them ideal for chatbots, virtual agents, and intelligent assistants that produce dynamic, context-aware replies.

Convert speech to text in real time → Azure AI SpeechAzure AI Speech provides speech-to-text capabilities (speech recognition) that convert spoken language into written text instantly. It is commonly used in transcription services, voice command systems, and live captioning applications. Additionally, the Speech service supports text-to-speech (speech synthesis) and speech translation, making it versatile for voice-based AI applications.

By understanding each service’s specialization—Vision for visual data, OpenAI for generative text, and Speech for audio processing—you can correctly match the capabilities.

According to the Microsoft Azure AI Fundamentals (AI-900) official study guide and the Microsoft Learn module “Describe features of common AI workloads and considerations”, Microsoft has defined six guiding principles for responsible AI. These principles are intended to ensure that AI systems are developed and deployed in ways that are ethical, transparent, and beneficial to all. The six principles are: Fairness, Reliability and Safety, Privacy and Security, Inclusiveness, Transparency, and Accountability.

Let’s break down the three correct options:

Fairness – Microsoft emphasizes that AI systems should treat all individuals fairly and avoid discrimination against people based on gender, race, age, or other characteristics. Fairness ensures that outcomes and decisions from AI systems are equitable across diverse user groups. In the AI-900 learning materials, fairness is explained as a foundational value that ensures algorithms and models do not introduce or amplify societal bias.

Reliability and Safety – This principle ensures that AI systems function as intended under all expected conditions and that they can handle unexpected inputs safely. Microsoft states that AI should be tested rigorously and validated for reliability before deployment. AI systems must perform consistently and avoid causing harm due to errors or failures.

Inclusiveness – Inclusiveness focuses on empowering everyone and engaging people of all backgrounds. Microsoft’s responsible AI guidance stresses designing AI systems that understand and respect cultural, linguistic, and ability differences to make technology accessible and beneficial to all users.

Options A (knowledgeability), B (decisiveness), and E (opinionatedness) are not part of Microsoft’s Responsible AI principles. These terms do not appear in any Microsoft Learn AI-900 curriculum or official responsible AI documentation.

Thus, based on the verified AI-900 study content and Microsoft’s Responsible AI framework, the correct answer is C. Inclusiveness, D. Fairness, and F. Reliability and Safety.

According to the Microsoft Azure AI Fundamentals (AI-900) official study guide and Microsoft Learn module “Identify features of classification machine learning”, classification is a type of supervised machine learning used when the goal is to predict a categorical outcome. That means the output variable represents discrete labels such as Yes/No, True/False, or Category A/B/C.

In this example, the model is predicting whether a person uses a bicycle (Yes or No) — a binary categorical outcome. The input (distance from home to work) is numeric, but the prediction is a class or category, which makes it a classification problem.

To compare:

A and D (predicting how many cups of coffee or race minutes) involve numeric predictions, which are regression tasks.

B (grouping images by similar colors) involves clustering, an unsupervised learning method used to find natural groupings in data.

Thus, the use case that fits classification is predicting whether someone uses a bicycle, since the answer is categorical.

According to the Microsoft Azure AI Fundamentals (AI-900) official study guide and Microsoft Learn module “Describe features of machine learning on Azure”, a classification model outputs a probability score representing how likely each input belongs to a particular class. To decide whether a prediction is “positive” or “negative,” the model applies a threshold (often defaulted to 0.5). Adjusting this threshold directly affects the balance between false positives and false negatives.

A false positive occurs when the model incorrectly predicts a positive outcome (for example, predicting that a patient has a disease when they do not).

A false negative occurs when the model fails to predict a true positive (for example, predicting that a patient does not have a disease when they actually do).

To reduce false positives, you must make the model less likely to classify borderline cases as positive. This is done by increasing the decision threshold, thereby favoring false negatives (because the model will only classify a case as positive when the prediction confidence is very high). In other words, by moving the threshold upward, you tighten the model’s standard for what qualifies as a “positive” prediction, reducing incorrect positives.

Let’s review why other options are incorrect:

A. Include test data in training data: This contaminates your dataset and causes overfitting, which leads to unreliable performance metrics.

B. Increase the number of training iterations: This may improve learning but doesn’t specifically target false positives.

C. Modify the threshold in favor of false positives: That would increase, not reduce, false positives.

Therefore, the correct step to reduce false positives is to adjust the threshold in favor of false negatives, making the model more conservative when labeling a case as positive — hence, Answer: D.

According to Microsoft’s Responsible AI Principles, transparency refers to ensuring that AI systems are understandable and that their decisions or predictions can be explained clearly to users. When building an AI-based loan approval app, documenting the reasons for approving or rejecting a loan and making this information available to the applicant ensures that the system’s decision-making process is transparent and easily interpretable.

Transparency in AI involves making the model’s inputs, features, and reasoning process visible and comprehensible to both developers and users. For a loan approval application, this could mean showing which factors—such as income, credit score, or debt ratio—influenced the outcome. Microsoft emphasizes that users should be aware when AI is making decisions that affect them and should have access to an explanation of how those decisions were made.

Fairness (A) ensures that AI systems treat all individuals equitably without bias.

Inclusiveness (B) focuses on accessibility and ensuring AI benefits all groups.

Accountability (D) ensures that humans remain responsible for the outcomes of AI systems.

While accountability supports ethical oversight, the specific act of explaining and documenting the AI’s decision process aligns directly with transparency.

Therefore, the verified answer is C. Transparency.

These answers are derived from the Microsoft Azure AI Fundamentals (AI-900) Official Study Guide and the Microsoft Learn module “Explore computer vision in Microsoft Azure.” The Azure Face service, part of Azure Cognitive Services, provides advanced facial recognition capabilities including detection, verification, identification, grouping, and similarity analysis.

Let’s analyze each statement:

“The Face service can be used to group all the employees who have similar facial characteristics.” → YesThe Face service supports a grouping function that automatically organizes a collection of unknown faces into groups based on visual similarity. It doesn’t require labeled data; instead, it identifies clusters of similar-looking faces. This is particularly useful when building or validating datasets of people.

“The Face service will be more accurate if you provide more sample photos of each employee from different angles.” → YesAccording to Microsoft documentation, model accuracy improves when you provide multiple high-quality images of each person under different conditions—such as varying lighting, poses, and angles. This diversity allows the service to better learn unique facial characteristics and improves recognition reliability, especially for identification and verification tasks.

“If an employee is wearing sunglasses, the Face service will always fail to recognize the employee.” → NoThis is incorrect. While occlusions (like sunglasses or hats) can reduce accuracy, the service may still recognize the person depending on how much of the face remains visible. Microsoft Learn explicitly notes that partial occlusion affects recognition confidence but does not guarantee failure.

In conclusion, the Face service can group similar faces (Yes), become more accurate with diverse samples (Yes), and still recognize partially covered faces though with lower confidence (No). These principles align directly with the Face API’s core functions and AI-900 learning objectives regarding computer vision and responsible AI-based facial recognition.

According to the Microsoft Learn documentation for Azure AI Document Intelligence (formerly Form Recognizer) and the AI-900 official study materials, the prebuilt receipt model in Azure AI Document Intelligence supports analyzing image and PDF files up to a maximum size of 10 MB per document.

Azure AI Document Intelligence is a cloud-based service that applies advanced optical character recognition (OCR) and machine learning to extract structured information from documents such as receipts, invoices, identity cards, and business forms. The prebuilt receipt model is specifically designed to extract key data fields from retail receipts—such as merchant name, transaction date, items purchased, subtotal, tax, and total—without requiring users to build or train a custom model.

As per Microsoft’s service limits, the input file for the prebuilt models (including receipts, invoices, business cards, and identity documents) must:

Not exceed 10 MB in file size.

Not exceed 17 x 17 inches (43 x 43 cm) in physical dimensions.

Be in a supported image or document format such as JPG, PNG, BMP, TIFF, or PDF.

Let’s examine why other options are incorrect:

A. 5 MB → Too small; the service allows up to 10 MB.

C. 50 MB and D. 100 MB → Exceed the official maximum file size supported by Azure AI Document Intelligence.

Therefore, when using the prebuilt receipt model, you must ensure that the input file is 10 MB or smaller to be successfully processed by the service.

QnA Maker supports automatic extraction of question-and-answer pairs from structured files such as PDF, Microsoft Word, or Excel documents, as well as from public webpages. This makes PDF the correct file format for populating a knowledge base.

Other options are invalid:

B. PPTX – Not supported.

C. XML – Not a recognized input for QnA Maker.

D. ZIP – Used for packaging, not Q & A content.

According to the Microsoft Azure AI Fundamentals (AI-900) official study guide and the Microsoft Learn module “Identify features of computer vision workloads on Azure”, computer vision models can perform different types of image analysis depending on the goal of the task. The main types include image classification, object detection, and semantic segmentation. Each method analyzes images at a different level of granularity.

Image Classification → Separate images of polar bears and brown bearsImage classification assigns an entire image to a specific category or label. The model analyzes the image as a whole and determines which predefined class it belongs to. For example, in this case, the model would look at the features of each image and decide whether it shows a polar bear or a brown bear. The Microsoft Learn materials define classification as “assigning an image to a specific category.”

Object Detection → Determine the location of a bear in a photoObject detection identifies where objects appear within an image by drawing bounding boxes around them. This type of model not only classifies what object is present but also provides its location. Microsoft Learn explains that object detection “detects and locates individual objects within an image.” For instance, the model can detect a bear in a forest scene and highlight its position.

Semantic Segmentation → Determine which pixels in an image are part of a bearSemantic segmentation is the most detailed form of image analysis. It classifies each pixel in an image according to the object it belongs to. In this scenario, the model identifies every pixel corresponding to the bear’s body. The AI-900 content defines this as “classifying every pixel in an image into a category.”

To summarize:

Image classification → Categorizes entire images.

Object detection → Locates and labels objects within images.

Semantic segmentation → Labels each pixel for precise object boundaries.

https://nanonets.com/blog/how-to-do-semantic-segmentation-using-deep-learning/

Object detection is similar to tagging, but the API returns the bounding box coordinates (in pixels) for each object found. For example, if an image contains a dog, cat and person, the Detect operation will list those objects together with their coordinates in the image. You can use this functionality to process the relationships between the objects in an image. It also lets you determine whether there are multiple instances of the same tag in an image.

The Detect API applies tags based on the objects or living things identified in the image. There is currently no formal relationship between the tagging taxonomy and the object detection taxonomy. At a conceptual level, the Detect API only finds objects and living things, while the Tag API can also include contextual terms like " indoor " , which can ' t be localized with bounding boxes.

A transformer model is the foundational architecture behind many modern natural language processing systems such as GPT and BERT. It processes text data through multiple key stages. According to the Microsoft Azure AI Fundamentals (AI-900) curriculum and Microsoft Learn materials, the major stages of a transformer-based large language model are tokenization, embedding calculation, and next token prediction.

Tokenization (C) – The first step converts raw text into smaller units called tokens (words, subwords, or characters). This process allows the model to handle text in a structured numerical form rather than as raw language.

Embedding Calculation (B) – After tokenization, the tokens are mapped into high-dimensional numeric vectors, known as embeddings. These embeddings capture semantic relationships between words and phrases so that the model can understand context and meaning.

Next Token Prediction (D) – This stage is the heart of transformer operation, where the model predicts the next likely token in a sequence based on prior tokens. Repeated next-token predictions enable text generation, summarization, or translation.

Options A (object detection) and E (anonymization) are incorrect because they relate to vision and privacy workflows, not language modeling.

Box 1: Entity recognition

the Named Entity Recognition module in Machine Learning Studio (classic), to identify the names of things, such as people, companies, or locations in a column of text.

Named entity recognition is an important area of research in machine learning and natural language processing (NLP), because it can be used to answer many real-world questions, such as:

Which companies were mentioned in a news article?

Does a tweet contain the name of a person? Does the tweet also provide his current location?

Were specified products mentioned in complaints or reviews?

Box 2: Sentiment Analysis

The Text Analytics API ' s Sentiment Analysis feature provides two ways for detecting positive and negative sentiment. If you send a Sentiment Analysis request, the API will return sentiment labels (such as " negative " , " neutral " and " positive " ) and confidence scores at the sentence and document-level.

Microsoft’s Responsible AI principles are central to the AI-900 curriculum and consist of six key tenets:

Fairness – AI systems should treat all people fairly.

Reliability and safety – AI systems should perform reliably and safely.

Privacy and security – AI systems should be secure and respect user privacy.

Inclusiveness – AI systems should empower everyone.

Transparency – AI systems should be understandable.

Accountability – People should be accountable for AI outcomes.

The statement “AI systems should be secure and respect privacy” reflects the Privacy and Security principle, which ensures AI solutions protect personal data and operate within compliance frameworks. Microsoft’s responsible AI framework emphasizes building trust by safeguarding sensitive data used in AI applications.

The other options do not align with official responsible AI principles; for example, AI systems need not “be in the public domain,” nor are they meant to prioritize developers’ interests or expose personal details. Hence, the correct and Microsoft-verified answer is C. AI systems should be secure and respect privacy.

The correct answers are based on the Microsoft Azure AI Fundamentals (AI-900) Official Study Guide and the Microsoft Learn module “Explore fundamental principles of machine learning.”

In supervised machine learning, data is typically divided into three main subsets:

Training set – used to train the model, i.e., to teach the algorithm the patterns and relationships between input features and output labels.

Validation set – used to evaluate the model during training to tune hyperparameters and prevent overfitting.

Test set – used after training to assess the final model’s performance on unseen data.

Let’s analyze each statement in light of these definitions:

“A validation set includes the set of input examples that will be used to train a model.” → NoThis is incorrect because the training set, not the validation set, contains the input examples used for model training. The validation set is separate from the training data to ensure unbiased evaluation.

“A validation set can be used to determine how well a model predicts labels.” → YesThis is correct. The validation set helps assess how effectively the model generalizes during training. It measures performance and helps tune model parameters for optimal results.

“A validation set can be used to verify that all the training data was used to train the model.” → NoThis is false. The validation set is not used to verify the completeness of training data usage. It exists independently to evaluate the model’s performance during training cycles.

According to Microsoft Learn, using a validation set helps ensure that a model generalizes well and avoids overfitting to the training data. It plays a crucial role in refining and optimizing models before final testing.

In this scenario, analyzing vehicle images and measuring the distance between them requires first detecting each vehicle’s position in the image. Object detection models can locate and identify multiple objects (such as cars, trucks, or motorcycles) by assigning bounding boxes. Once detected, their coordinates can be used to calculate distances or spacing.

Image classification only assigns a single label per image, not per object. Facial recognition is human-focused, and OCR deals with text extraction. Thus, object detection is the correct model type for this task.

The Azure OpenAI DALL-E model is a generative image model designed to create original images from textual descriptions (prompts). According to the Microsoft Learn documentation and the AI-900 study guide, DALL-E’s primary function is text-to-image generation—it converts creative or descriptive text input into visually relevant imagery.

“Generate captions for uploaded images” → NoDALL-E cannot create image captions. Captioning an image (describing what’s in an uploaded image) is a vision analysis task, not an image generation task. That functionality belongs to Azure AI Vision, which can analyze and describe images, detect objects, and generate captions automatically.

“Reliably generate technically accurate diagrams” → NoWhile DALL-E can create visually appealing artwork or conceptual sketches, it is not designed for producing precise or technically correct diagrams, such as engineering schematics or architectural blueprints. The model’s generative process emphasizes creativity and visual diversity rather than factual or geometric accuracy. Thus, it cannot be relied upon for professional technical outputs.

“Generate decorative images to enhance learning materials” → YesThis is one of DALL-E’s strongest use cases. It can generate decorative, conceptual, or illustrative images to enhance presentations, educational materials, and marketing content. It enables educators and designers to quickly produce unique visuals aligned with specific themes or topics, enhancing engagement and creativity.

Yes, Yes, and No.

According to the Microsoft Azure AI Fundamentals (AI-900) official study materials and the Microsoft Learn module “Identify features of natural language processing (NLP) workloads on Azure”, the Azure Translator service is a cloud-based AI service within Azure Cognitive Services that provides real-time text translation across multiple languages.

“You can use the Translator service to translate text between languages.” – Yes.This is the core function of the Translator service. It takes text as input in one language and returns it in another using advanced neural machine translation models. This aligns with the AI-900 learning objective: “Describe the capabilities of Azure Cognitive Services for language”, which specifically names Azure Translator as the service used to perform automatic text translation. The service supports over 100 languages and dialects, offering both single-sentence and document-level translations.

“You can use the Translator service to detect the language of a given text.” – Yes.This statement is also true. The Translator service automatically detects the source language if it is not specified in the request. This feature is documented in the Azure Translator API, where the system identifies the input language before performing translation. The AI-900 exam content emphasizes this as one of the Translator service’s built-in capabilities—language detection for untagged text.

“You can use the Translator service to transcribe audible speech into text.” – No.This is not a function of Translator. Transcription (converting speech to text) is a speech AI workload, handled by the Azure Speech Service, not Translator. The Speech-to-Text capability in Azure Cognitive Services processes spoken audio input and returns the text transcription. The Translator service only works with text input, not direct audio.

Therefore, based on official AI-900 guidance, the verified configuration is:

✅ Yes – for text translation

✅ Yes – for language detection

❌ No – for speech transcription.

This aligns precisely with the AI-900 learning outcomes describing Text Translation and Language Detection as Translator capabilities, and Speech Transcription as part of the separate Speech service.

According to the Microsoft Azure AI Fundamentals (AI-900) official study guide and the Microsoft Learn module “Describe features of common AI workloads”, clustering is a type of unsupervised machine learning used to group data points that share similar characteristics. In unsupervised learning, the data provided to the model does not have predefined labels or outcomes. Instead, the algorithm identifies inherent patterns or groupings within the dataset based on similarities in input features.

In this scenario, the task is to identify groups of people who have similar purchasing habits. There is no predefined label such as “buyer type” or “purchase category.” The goal is to discover hidden patterns—such as grouping customers by spending behavior, preferred products, or frequency of purchases. This is precisely what clustering algorithms are designed to do.

Clustering is commonly used in:

Customer segmentation for marketing analytics.

Market basket analysis to find associations in purchasing patterns.

Recommender systems to identify similar user profiles.

Anomaly detection when outliers deviate from natural clusters.

Typical algorithms for clustering include K-means, Hierarchical clustering, and DBSCAN. These models analyze multidimensional data to form clusters that maximize intra-group similarity and minimize inter-group similarity.

By contrast:

Classification (A) is a supervised learning method that predicts a categorical label (e.g., whether a customer will churn or not). It requires labeled training data.

Regression (B) is used to predict continuous numeric values (e.g., sales revenue, temperature).

Since the question focuses on discovering groups of similar customers without prior labels, the correct type of machine learning is Clustering.

Clustering.

According to the Microsoft Azure AI Fundamentals (AI-900) official study guide and Microsoft Learn module “Identify features of common machine learning types”, clustering is an unsupervised machine learning technique used to group data points into distinct segments or clusters based on shared characteristics. Unlike supervised learning (classification or regression), clustering works with unlabeled data, discovering natural groupings without predefined outcomes.

In this question, Recency, Frequency, and Monetary (RFM) values are common marketing metrics used to evaluate customer behavior:

Recency – how recently a customer made a purchase.

Frequency – how often they make purchases.

Monetary – how much money they spend.

Using RFM analysis, a company can segment its customers into groups such as “loyal,” “occasional,” or “at-risk” buyers. This segmentation process does not rely on predefined labels but rather discovers patterns within the data — which is the defining characteristic of clustering.

In the AI-900 context, clustering is described as a method that “groups items with similar features so that items in the same group are more similar to each other than to those in other groups.” Common algorithms used include K-Means, Hierarchical Clustering, and DBSCAN, all available within Azure Machine Learning Designer and other Azure ML environments.

To clarify the incorrect options:

Classification is supervised learning used to predict discrete categories (e.g., yes/no, spam/not spam).

Regression predicts continuous numeric values (e.g., house prices).

Regularization is a model optimization technique, not a type of machine learning.

Therefore, when businesses use RFM values to identify customer segments without labeled outcomes, this is an application of unsupervised learning through clustering.

According to the Microsoft Azure AI Fundamentals (AI-900) Official Study Guide and the Microsoft Learn module “Explore computer vision in Microsoft Azure,” computer vision is a field of artificial intelligence that enables computers to interpret and understand visual information from the world — such as images or videos.

In this scenario, the task is to count the number of animals in an area based on a video feed. This requires the system to:

Detect the presence of animals in each frame of the video (object detection).

Track and count them across multiple frames as they move.

These are classic computer vision tasks, as they involve analyzing visual inputs (video or image data) and identifying objects (in this case, animals). Azure provides services such as Azure Computer Vision, Custom Vision, and Video Indexer, which can perform object detection, counting, and activity recognition using AI models trained on visual datasets.

Why the other options are incorrect:

Forecasting: Involves predicting future values based on historical data (e.g., predicting sales or weather), not analyzing video feeds.

Knowledge mining: Focuses on extracting insights from large text-based document repositories, not images or videos.

Anomaly detection: Identifies unusual patterns in numeric or time-series data, not visual objects.

Therefore, identifying and counting animals in video footage falls under computer vision, since it uses AI to visually detect, classify, and quantify objects in real-time or recorded feeds.

According to the Microsoft Azure AI Fundamentals (AI-900) official study guide and the Microsoft Learn module “Identify features of common machine learning types”, a regression model is a type of supervised machine learning model that is used to predict continuous numeric values based on one or more input variables (features).

In supervised learning, models are trained using labeled data, where each input record has a known target value (label). For regression specifically, the label represents a numeric quantity — such as price, age, temperature, or sales figures. The model learns to approximate a function that maps input variables to numeric outputs.

For example:

Predicting house prices based on size, location, and number of rooms.

Predicting monthly sales revenue from marketing spend and seasonality.

Forecasting temperature based on historical weather data.

In Azure Machine Learning Designer and AutoML, when building a regression model, the label column must therefore be of numeric data type (integer or float). If the label were categorical (for example, “yes/no” or “approved/denied”), the appropriate model type would be classification, not regression.

To contrast with other options:

Boolean – used in binary classification problems (true/false outcomes).

Datetime – used for time series forecasting, not standard regression labels.

Text – used as input features for NLP models, not as regression targets.

Hence, when configuring a regression task in Azure, ensuring the target variable (label) is numeric is a fundamental requirement. The model’s performance metrics—such as Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and R²—also rely on numeric computations.

According to the Microsoft Azure AI Fundamentals (AI-900) learning content, Language Understanding Intelligent Service (LUIS) is part of Azure Cognitive Services used to interpret the meaning or intent behind a user’s input in natural language. When a user says, “Call me back later,” the system must recognize that the user intends for a call to be scheduled or delayed—this is not just about translating or analyzing text but understanding intent and relevant entities.

LUIS allows developers to train models to identify intents (such as ScheduleCall, CancelMeeting, etc.) and extract key entities (like names, times, or actions) from text inputs. It is typically integrated with conversational agents such as Azure Bot Service, enabling more natural, human-like interactions.

Other options do not fit the scenario:

Translator Text (A) translates text between languages but does not interpret meaning.

Text Analytics (B) performs sentiment analysis, key phrase extraction, and named entity recognition, but it doesn’t identify intent.

Speech (C) converts spoken language to text or text to speech but doesn’t interpret what the words mean.

Therefore, for understanding user intent such as “Call me back later,” the correct AI service is D. Language Understanding (LUIS).

Features

The correct completion of the sentence is:

“In a machine learning model, the data that is used as inputs are called features.”

According to the Microsoft Azure AI Fundamentals (AI-900) official study materials and Microsoft Learn module “Identify features of common machine learning types,” the term feature refers to an individual measurable property or characteristic of the data that is used by a machine learning model to make predictions or decisions.

In supervised and unsupervised learning, features serve as the inputs to the model. They are the variables that represent the information the algorithm learns from. For example, if a model predicts the price of a house, the features might include square footage, number of bedrooms, location, and age of the property. These features feed into the model so that it can learn the relationship between inputs and outputs.

Microsoft Learn further defines these key concepts:

Features: Input variables (independent variables) used by the model to learn patterns.

Labels: The desired output or target variable that the model is trained to predict (e.g., price, category).

Instances: Individual rows or data records within the dataset (each instance has multiple features).

Functions: Algorithms or mathematical operations used during training and prediction — not data inputs.

Therefore, among the provided options — features, functions, labels, instances — only features accurately describe the data elements used as inputs for training or inference in a machine learning model.

In summary, within the AI-900 learning context:

Features = inputs to the model.

Labels = outputs for supervised learning.

Instances = examples (rows) of data.

 “Generative AI enables software applications to generate new content, such as language dialogs and images.” — YES

This statement is true. According to the Microsoft Azure AI Fundamentals (AI-900) study guide and Microsoft Learn documentation, Generative AI refers to systems capable of creating new content such as text, audio, images, video, and code. Models like GPT, DALL·E, and Codex use deep learning to generate human-like responses, natural conversations, or creative media. This is a key differentiator between generative and discriminative AI — generative AI produces new data, while discriminative AI categorizes or analyzes existing data.

 “The difference between a large language model (LLM) and a small language model (SLM) is the number of variables in the model.” — YES

This statement is true. The primary distinction between an LLM and an SLM lies in the scale of parameters (variables) within the neural network. LLMs contain billions or even trillions of parameters, which enable them to capture complex linguistic patterns and perform broader tasks. SLMs have fewer parameters, making them faster but less capable of handling complex, context-rich tasks.

 “Generative AI is a type of supervised learning.” — NO

This statement is false. Generative AI models are typically trained using unsupervised or self-supervised learning methods. They learn by predicting missing or next elements in large text or image datasets rather than relying on labeled input-output pairs, which are used in supervised learning.

“features.”

According to the Microsoft Azure AI Fundamentals (AI-900) official study guide and Microsoft Learn module “Describe fundamental principles of machine learning on Azure,” in a machine learning model, the data used as inputs are known as features, while the data that represents the output or target prediction is known as the label.

Features are measurable attributes or properties of the data used by a model to learn patterns and make predictions. They are also referred to as independent variables because they influence the result that the model tries to predict. For example, in a machine learning model that predicts house prices:

Features might include square footage, location, and number of bedrooms, while

The label would be the house price (the value being predicted).

In the context of Azure Machine Learning, during model training, features are passed into the algorithm as input variables (X-values), and the label is the corresponding output (Y-value). The model then learns the relationship between the features and the label.

Let’s review the incorrect options:

Functions: These are mathematical operations or relationships used inside algorithms, not the input data itself.

Labels: These are the outputs or results that the model predicts, not the inputs.

Instances: These refer to individual data records or rows in the dataset, not the input fields themselves.

Hence, in any supervised or unsupervised learning process, the input data (independent variables) are called features, and the model uses them to predict labels (dependent variables).

According to the Microsoft Azure AI Fundamentals (AI-900) official study materials and Microsoft Learn documentation, Azure AI Custom Vision is specifically designed for training custom image classification and object detection models using images that a company provides. In this scenario, the company wants to identify competitor products from images captured in retail stores — a classic use case for custom image classification or object detection, depending on whether you are labeling entire images or identifying multiple items within an image.

Azure AI Custom Vision allows users to:

Upload their own labeled training images.

Train a model that learns to recognize specific objects (in this case, competitor products).

Evaluate, iterate, and deploy the model as an API endpoint for real-time inference.

This fits perfectly with the requirement that the solution “must be trained on images provided by your company.” The key phrase here indicates the need for a custom-trained model rather than a prebuilt one.

The other options are not suitable for this scenario:

B. Azure AI Computer Vision provides prebuilt models for general-purpose image understanding (e.g., detecting common objects, reading text, describing scenes). It is not intended for training on custom datasets.

C. Face service is limited to detecting and recognizing human faces; it cannot be trained to identify products.

D. Azure AI Document Intelligence (formerly Form Recognizer) is focused on extracting structured data from documents and forms, not analyzing retail images.

Therefore, per Microsoft’s official AI-900 training content, when a solution must be trained on custom company images to recognize specific products, the appropriate service is Azure AI Custom Vision.

According to the Microsoft Learn module “Build a QnA Maker knowledge base”, QnA Maker allows developers to create bots that answer user queries based on documents like FAQs or manuals. To make a bot more natural and conversational, Microsoft provides a “chit-chat” feature — a prebuilt, professionally written set of responses to common conversational phrases such as greetings (“Hello”), small talk (“How are you?”), and polite phrases (“Thank you”).

Adding chit-chat improves user experience by making the bot sound friendlier and more human-like. It doesn’t alter the main Q & A logic but enhances the bot’s tone and responsiveness.

The other options are not correct:

A. Increase the confidence threshold makes the bot more selective in responses but doesn’t add new conversational features.

B. Enable active learning improves knowledge base accuracy over time through user feedback.

C. Create multi-turn questions adds conversational flow for related topics but doesn’t add greetings or casual dialogue.

Thus, to make the bot more personable, the correct action is to Add chit-chat.

Box 1: Select Columns in Dataset

For Columns to be cleaned, choose the columns that contain the missing values you want to change. You can choose multiple columns, but you must use the same replacement method in all selected columns.

Example:

The task is to build a machine learning model in Azure Machine Learning designer to predict automobile prices, which is a regression problem since the output (price) is a continuous numeric value. The pipeline must follow the logical data preparation, training, and evaluation flow as outlined in the Microsoft Azure AI Fundamentals (AI-900) study guide and Microsoft Learn module “Create a machine learning model with Azure Machine Learning designer.”

Here’s the correct sequence and reasoning:

Select Columns in Dataset:The first step after loading the raw automobile dataset is to choose the relevant columns that will be used as features (inputs) and the label (output). This module ensures that only necessary fields (for example, horsepower, engine size, mileage, etc.) are used to train the model while excluding irrelevant columns like vehicle ID or serial number.

Split Data:Next, the cleaned and filtered dataset must be split into two subsets: training data and testing data (often 70/30 or 80/20). This allows the model to be trained on one portion and evaluated on the other to measure predictive accuracy.

Linear Regression:Since automobile price prediction is a numeric prediction task, the appropriate learning algorithm is Linear Regression. This supervised algorithm learns relationships between numeric features and the target (price).

Finally, the workflow connects the training data and Linear Regression module to the Train Model module, which outputs a trained regression model. The trained model is then linked to the Score Model module to compare predicted vs. actual prices.

This pipeline fully aligns with Microsoft’s recommended process for regression in Azure ML Designer.

Conversational AI workloads involve human-like dialogue with AI systems.

A: A smart assistant (e.g., smart speaker) uses voice-based conversational AI.

B: A knowledge-based chatbot interacts with users via natural language.Options C and D describe automation/IoT workloads, not conversational AI.

✅ Final Answer (Q110): A and B

According to the Microsoft Azure AI Fundamentals (AI-900) official study materials and Microsoft Learn documentation on GitHub Copilot, the GitHub Copilot extension for Microsoft Visual Studio Code (VS Code) is powered by the OpenAI Codex model, a specialized descendant of the GPT (Generative Pre-trained Transformer) family of models. The OpenAI Codex model is specifically fine-tuned for programming tasks—it understands and generates code in multiple programming languages such as Python, JavaScript, C#, and more.

GitHub Copilot functions as an AI-powered coding assistant that suggests complete lines or blocks of code, helps write comments, and can even generate functions based on natural language prompts entered by the developer. When a user types a comment like “// sort a list of numbers in ascending order,” Copilot uses the Codex model to understand the intent and generate an appropriate code implementation.

The integration with Visual Studio Code allows developers to work efficiently without needing to switch between documentation and their coding environment. GitHub Copilot leverages context-aware suggestions, meaning it understands the surrounding code, function definitions, and variable names to provide relevant completions.

The other options are incorrect because:

GitHub source control manages code repositories but doesn’t use AI models.

IntelliSense is a built-in VS Code feature for code completion but doesn’t rely on OpenAI models.

Microsoft 365 Copilot uses large language models (like GPT-4) for productivity tools such as Word, Excel, and Outlook—not for code generation.

Therefore, the correct choice that uses the OpenAI Codex model in Visual Studio Code is GitHub Copilot, aligning with Microsoft’s AI-900 learning objectives under “Identify common use cases for Azure OpenAI and GitHub Copilot.”

According to the Microsoft Azure AI Fundamentals (AI-900) official study guide and Microsoft Learn module “Describe features of common AI workloads”, there are three primary supervised and unsupervised machine learning types: Regression, Classification, and Clustering. Each type of learning addresses a different kind of problem depending on the data and desired prediction output.

Regression – Regression models are used to predict numeric, continuous values. The study guide specifies that “regression predicts a number.” In the scenario “Predict how many minutes late a flight will arrive based on the amount of snowfall,” the output (minutes late) is a continuous numeric value. Therefore, this is a regression problem. Regression algorithms like linear regression or decision tree regression estimate relationships between variables and predict measurable quantities.

Clustering – Clustering falls under unsupervised learning, where the model identifies natural groupings or patterns in unlabeled data. The official AI-900 training material states that “clustering is used to find groups or segments of data that share similar characteristics.” The scenario “Segment customers into different groups to support a marketing department” fits this description because the goal is to group customers based on behavior or demographics without predefined labels. Thus, it is a clustering problem.

Classification – Classification is a supervised learning method used to predict discrete categories or labels. The AI-900 content defines classification as “predicting which category an item belongs to.” The scenario “Predict whether a student will complete a university course” requires a yes/no (binary) outcome, which is a classic classification problem. Examples include logistic regression, decision trees, or neural networks trained for categorical prediction.

In summary:

Regression → Predicts continuous numeric outcomes.

Clustering → Groups data by similarities without predefined labels.

Classification → Predicts discrete or categorical outcomes.

Hence, the correct and verified mappings based on the official AI-900 study material are:

Regression → Flight delay prediction

Clustering → Customer segmentation

Classification → Course completion prediction

Privacy and security.

According to Microsoft’s Responsible AI Principles, implementing filters to block harmful or inappropriate content in a Generative AI chat solution demonstrates a commitment to the Privacy and Security principle. This principle ensures that AI systems are designed and operated in a way that protects users, their data, and society from harm.

When a chat system uses Generative AI models (like Azure OpenAI’s GPT-based services), there is a risk that the model might produce unsafe, offensive, or sensitive content. Microsoft addresses this through content filters and safety systems, which automatically detect and block violent, hate-based, or sexually explicit outputs. This is part of responsible deployment practices to ensure that user interactions remain safe, private, and compliant with ethical standards.

Implementing these filters aligns with the Privacy and Security principle because it:

Protects users from exposure to harmful or abusive content.

Ensures that conversations are safeguarded against malicious or unsafe use.

Upholds user trust by maintaining a safe digital environment for all participants.

Let’s briefly clarify why the other options are incorrect:

Fairness deals with ensuring unbiased treatment and equitable outcomes in AI decisions.

Transparency focuses on explaining how AI systems make decisions.

Accountability refers to human oversight and responsibility for AI actions.

Thus, content filtering mechanisms are explicitly an example of Privacy and Security, as they protect users and data from harm or misuse while maintaining ethical AI behavior.

Therefore, the verified correct answer is Privacy and security.

In a Language Understanding (LUIS) application, an utterance represents an example of what a user might say to the bot. According to Microsoft Learn – “Build a Language Understanding app”, an utterance is a sample phrase that helps train the LUIS model to recognize user intent.

In the given example — “Which act is playing on the main stage?” — the statement is an utterance that a user might say to find out about show schedules. LUIS uses utterances like this to identify the intent (the user’s goal, e.g., GetShowInfo) and to extract any entities (e.g., main stage) that provide additional details for fulfilling the request.

To clarify the other elements:

Intent: The overall purpose or action (e.g., “FindShowDetails”).

Entity: Specific information in the utterance (e.g., “main stage”).

Domain: A general subject area (e.g., entertainment, events).

Thus, “Which act is playing on the main stage?” is an utterance used to train the LUIS model to understand natural language input.

According to the Microsoft Learn module “Describe features of Azure Machine Learning” and the AI-900 study guide, Azure Machine Learning designer supports extending workflows through custom code modules written in Python and R.

Python is the most commonly used language for AI and machine learning due to its extensive library support (such as TensorFlow, Scikit-learn, and PyTorch).

R is widely used for statistical computing and data visualization, making it valuable for analytical workloads.

In Azure Machine Learning, users can insert Python Script or Execute R Script modules within the visual designer to perform advanced operations or custom data transformations.

C# and Scala are not supported directly in Azure Machine Learning Designer. C# is more common in application development, and Scala is primarily used in big data frameworks like Apache Spark.

Hence, the correct answers are C. Python and D. R.

Let’s analyze each application in the context of Microsoft Azure AI Fundamentals (AI-900) and computer vision model types.

App1 – Uses a set of images of tumors to identify whether the tumors are benign or malignant and suggest a treatment → Image classificationThis application is performing image classification, where each image (of a tumor) is assigned to a single predefined category — benign or malignant. Image classification models learn patterns from labeled training images and predict the correct class for new ones. In this case, the model identifies the type of tumor, a classic binary classification scenario.

App2 – Uses images from cameras to track individual livestock as they move around a farm → Object detectionThis scenario describes object detection, which not only identifies what objects (in this case, animals) are in an image but also locates them by drawing bounding boxes. Tracking movement requires detecting the position of each animal frame by frame. Object detection models are well-suited for use cases involving counting, tracking, or monitoring objects in a visual scene.

App3 – Identifies brands in photographs of billboards → Optical character recognition (OCR)This app involves reading and interpreting text (brand names, slogans, or logos) from images of billboards. Optical Character Recognition (OCR), part of Azure AI Vision, extracts textual information from images or scanned documents. Once extracted, that text can be analyzed to identify brand names or keywords.

Summary:

App1 → Image classification

App2 → Object detection

App3 → Optical character recognition (OCR)

Azure Custom Vision is a cognitive service that lets you build, deploy, and improve your own image classifiers. An image classifier is an AI service that applies labels (which represent classes) to images, according to their visual characteristics. Unlike the Computer Vision service, Custom Vision allows you to specify the labels to apply.

Note: The Custom Vision service uses a machine learning algorithm to apply labels to images. You, the developer, must submit groups of images that feature and lack the characteristics in question. You label the images yourself at the time of submission. Then the algorithm trains to this data and calculates its own accuracy by testing itself on those same images. Once the algorithm is trained, you can test, retrain, and eventually use it to classify new images according to the needs of your app. You can also export the model itself for offline use.

Description

Correct Service

Enables the use of natural language to query a knowledge base.

QnA Maker

Enables the real-time transcription of speech-to-text.

Speech

This question tests understanding of Azure Cognitive Services and their use cases as outlined in the Microsoft Azure AI Fundamentals (AI-900) study guide.

“Enables the use of natural language to query a knowledge base.” → QnA MakerAccording to Microsoft Learn’s AI-900 module “Identify features of Natural Language Processing (NLP) workloads and services,” QnA Maker is a cloud-based service that allows developers to build a question-and-answer layer over structured or unstructured content. It enables users to ask questions in natural language, and the service retrieves the most relevant answer from a knowledge base (such as FAQs, manuals, or documents).QnA Maker uses Natural Language Processing (NLP) techniques to interpret user intent and return an appropriate response. It is often integrated into chatbots built with Azure Bot Service to make them capable of conversational question-answering. In the newer Azure Cognitive Services lineup, QnA Maker capabilities are merged into Azure Cognitive Service for Language (Question Answering).

“Enables the real-time transcription of speech-to-text.” → SpeechThe Azure Speech service (part of Azure Cognitive Services) provides the ability to convert spoken language into written text in real time. This feature, called Speech-to-Text, uses deep neural network models to recognize and transcribe human speech with high accuracy.Microsoft’s AI-900 documentation specifies that Speech service capabilities also include text-to-speech, speech translation, and speaker recognition. Real-time transcription is widely used in applications such as voice assistants, captioning systems, call analytics, and accessibility tools.

Other listed services such as Azure Storage and Language Understanding (LUIS) serve different purposes:

Azure Storage handles data storage, not AI workloads.

LUIS identifies user intent from natural language but does not query knowledge bases directly.

According to the Microsoft Azure AI Fundamentals (AI-900) study materials and Microsoft’s Responsible AI guidelines, customers must obtain approval based on their intended usage before accessing and deploying Azure OpenAI Service. This requirement ensures that Microsoft upholds its commitment to Responsible AI principles, which include fairness, reliability and safety, privacy and security, inclusiveness, transparency, and accountability.

The Azure OpenAI Service provides access to powerful language models such as GPT series and Codex, which can generate, summarize, and understand natural language and code. Because of the potential for misuse—such as generating harmful content, misinformation, or unethical automation—Microsoft enforces a use case review and approval process before granting customers access to the service. This process involves submitting an application describing the intended purpose, deployment method, and compliance measures. Only after Microsoft validates that the proposed use aligns with responsible AI practices will access be approved.

This aligns with Microsoft’s documented commitment that “customers are required to submit an application that describes their intended use of the Azure OpenAI Service,” ensuring that all deployments follow ethical and legal standards. This approval step helps maintain transparency and prevent harmful or non-compliant use cases such as deepfake generation, biased automation, or malicious chatbot deployment.

Other options listed in the question are incorrect:

Commit to a minimum level of expenditure – Microsoft does not require financial commitments for ethical approval.

Pay an upfront fee – Payment is handled through normal Azure billing, not a special fee.

Provide credit card details – Not a responsible AI requirement; this is standard for any Azure subscription.

Therefore, the correct and verified answer per Microsoft’s Responsible AI framework and Azure AI-900 study

According to the Microsoft Azure AI Fundamentals (AI-900) official study materials and the Microsoft Learn module “Identify features of common machine learning types,” the evaluation phase occurs after training and testing a machine learning model. Evaluation metrics are used to measure how well the model performs when applied to data it has not seen before (the validation data).

The machine learning workflow includes the following key steps:

Data Preparation – Importing, cleaning, and transforming data.

Splitting the Data – Dividing it into training and validation (or test) sets.

Model Training – Using the training data to teach the model patterns or relationships.

Model Evaluation – Assessing the trained model using the validation data and evaluation metrics such as accuracy, precision, recall, F1 score, and root mean square error (RMSE).

As stated in the AI-900 content, evaluation metrics are crucial after testing, as they help determine if the model is accurate enough or if it requires retraining with different parameters or algorithms.

A. After you clean the data → incorrect, as metrics cannot be reviewed before training.

B. Before you train a model → incorrect, since the model has not yet learned patterns.

C. Before you choose the type of model → incorrect, as metrics depend on the model’s output.

Therefore, the verified answer is D. After you test a model on the validation data, which is when you review evaluation metrics to determine model performance and readiness for deployment.

In Azure Machine Learning designer, the Split Data module is specifically designed to divide a dataset into training and validation (or testing) subsets. The AI-900 study guide and the Microsoft Learn module “Split data for training and evaluation” explain that this module allows users to control how data is partitioned, ensuring that models are trained on one portion of the data and tested on unseen data to assess performance.

By default, the Split Data module uses a 70/30 or 80/20 ratio, meaning 70–80% of the data is used for training and the remaining 20–30% for validation or testing. This ensures the model’s generalizability and prevents overfitting.

The other options serve different purposes:

A. Select Columns in Dataset: Used to choose specific columns or features from a dataset.

B. Add Rows: Combines multiple datasets vertically.

D. Join Data: Combines datasets horizontally based on a common key.

Only Split Data performs the function of dividing data into training and validation subsets.

According to the AI-900 official study guide and Microsoft Learn module “Describe features of conversational AI workloads on Azure”, conversational AI refers to artificial intelligence systems that interact with users through natural language via text or speech. These systems include chatbots, virtual assistants, and interactive voice response (IVR) systems that simulate human conversation.

B. Chatbot that provides users with the ability to find answers on a website by themselvesThis is a classic example of conversational AI. Chatbots use natural language understanding (LUIS) and Azure Bot Service to interpret user input, identify intent, and provide relevant responses automatically. They help users self-serve information without human support, such as retrieving account details or answering FAQs.

C. Telephone voice menus to reduce the load on human resourcesAutomated telephone systems or IVRs use conversational AI to interpret spoken commands and route calls intelligently. This is often implemented using Azure Cognitive Services Speech (for speech-to-text and text-to-speech) combined with Azure Bot Service for managing dialogue flow.

According to the Microsoft Azure AI Fundamentals (AI-900) official study guide and Microsoft Learn modules on machine learning concepts, ensuring that the accuracy of a predictive model can be proven requires data partitioning—specifically splitting the available data into training and testing datasets. This is a foundational concept in supervised machine learning.

When you split the data, typically about 70–80% of the dataset is used for training the model, while the remaining 20–30% is used for testing (or validation). The reason behind this approach is to ensure that the model’s performance metrics—such as accuracy, precision, recall, and F1-score—are evaluated on data the model has never seen before. This prevents overfitting and allows you to demonstrate that the model generalizes well to new, unseen data.

In the AI-900 Microsoft Learn content under “Describe the machine learning process”, it is explained that after cleaning and transforming the data, the next essential step is data splitting to “evaluate model performance objectively.” By keeping training and testing data separate, you can prove the reliability and accuracy of the model’s predictions, which is particularly crucial in sensitive domains like clinical or healthcare analytics, where decision transparency and validation are vital.

Option A (Train the model by using the clinical data) is incorrect because you should not train and evaluate on the same data—it would lead to biased results.

Option C (Train the model using automated ML) is incorrect because automated ML is a method for training and tuning, but it doesn’t inherently prove accuracy.

Option D (Validate the model by using the clinical data) is also incorrect if you use the same dataset for validation and training—it would not prove true accuracy.

Therefore, per Microsoft’s official AI-900 study content, the verified correct answer is B. Split the clinical data into two datasets.

The correct answer is Azure AI Custom Vision.

According to the Microsoft Azure AI Fundamentals (AI-900) official study guide and the Microsoft Learn module “Explore computer vision in Azure”, the Azure AI Custom Vision service is specifically designed to allow users to train, test, and deploy custom image classification and object detection models using their own images.

The Azure AI Custom Vision service extends the capabilities of Azure’s general-purpose Computer Vision API by enabling organizations to upload their own labeled datasets, define custom tags (labels), and train models that are optimized for their specific use cases. This makes it ideal for object detection scenarios—such as detecting equipment in a manufacturing line, identifying products on store shelves, or recognizing medical images—where general-purpose models may not suffice.

By contrast, the Azure AI Computer Vision service provides pre-built models for tasks such as image description, tagging, face detection, and OCR (optical character recognition). It does not allow users to train their own models with custom data. Similarly, Azure AI Document Intelligence is used for extracting structured information from documents (forms, receipts, invoices), and Azure Video Analyzer for Media focuses on analyzing video content for insights and metadata extraction.

The AI-900 study guide emphasizes that the Custom Vision service supports two key model types:

Image Classification – categorizing entire images based on predefined tags.

Object Detection – identifying and locating multiple objects within an image by drawing bounding boxes.

Therefore, when the question specifies “train an object detection model by using your own images,” the correct Azure service is Azure AI Custom Vision, as it provides the necessary tools for training, evaluating, and deploying custom computer vision models tailored to a user’s dataset.

Hence, the verified correct answer is: Azure AI Custom Vision.

In unsupervised machine learning, the algorithm learns patterns or structure within data without pre-labeled outputs or target values. The primary goal is to discover hidden relationships or group similar data points automatically. The Microsoft Azure AI Fundamentals (AI-900) study materials identify clustering as the key example of unsupervised learning.

In clustering, algorithms such as K-means, hierarchical clustering, or DBSCAN group data based on feature similarity. For example, a business may cluster customers by purchase behavior to discover natural customer segments without prior category labels. The model finds inherent patterns within the data rather than being told what to predict.

By contrast, classification and regression are supervised learning techniques. In supervised learning, the algorithm is trained using labeled data where correct outputs are already known. Therefore, the correct answer is B. Clustering, as it best represents unsupervised learning in Azure AI-900 principles.

According to the Microsoft Azure AI Fundamentals (AI-900) official study materials and Microsoft Learn module “Identify features of common AI workloads,” the Azure Form Recognizer service is part of Azure Cognitive Services for Document Intelligence. It enables organizations to extract, analyze, and identify information from structured and unstructured documents, including sensitive or confidential data such as names, addresses, financial figures, and identification numbers.

Form Recognizer uses optical character recognition (OCR) combined with machine learning to automatically extract key-value pairs, tables, and text fields from documents like invoices, receipts, contracts, and forms. It can be customized to identify and classify documents that contain specific sensitive data, allowing businesses to automate compliance and data governance tasks.

By contrast:

A. Custom Vision is used for image classification and object detection — it analyzes visual data, not document content.

B. Conversational Language Understanding (formerly LUIS) identifies intent and entities in text conversations, not document structure or sensitive data.

Form Recognizer is explicitly mentioned in the AI-900 course as the tool for document analysis and extraction. It can even integrate with Azure Cognitive Search or Azure Purview for further data management and compliance workflows.

Therefore, the verified and correct answer, aligned with Microsoft’s official training content, is C. Form Recognizer, as it is the Azure Cognitive Service capable of identifying and processing documents containing sensitive information.

The correct service to identify similar faces in a set of images is Azure AI Vision, which includes the Face API capability. According to the Microsoft Learn module “Analyze images with Azure AI Vision”, this service provides prebuilt models for face detection, facial recognition, and similarity matching.

The Face API can detect individual faces in images and extract unique facial features to create a face embedding (a numerical representation of the face). It then compares these embeddings across multiple images to determine whether faces are similar or belong to the same person. This functionality is commonly used in identity verification, photo management systems, and security solutions.

The other options are incorrect:

B. Azure AI Custom Vision is used for custom image classification or object detection but does not provide face similarity or recognition features.

C. Azure AI Language processes text-based data (sentiment, entities, key phrases) — not visual content.

D. Azure OpenAI Service focuses on text generation, summarization, and conversation, not facial analysis.

Therefore, the Microsoft-verified service for identifying similar faces across images is A. Azure AI Vision.

The correct answer is A. Configure the Content filters settings.

When using the Azure OpenAI Service for text or image generation, Microsoft provides built-in content filtering to help detect and block potentially harmful or unsafe outputs. These filters are part of Microsoft’s Responsible AI framework and are designed to prevent the generation of offensive, violent, sexual, or otherwise restricted content.

To ensure the highest level of protection, you can configure content filter settings within the Azure OpenAI deployment. This allows you to define stricter policies based on your organization’s safety requirements. For image generation models such as DALL·E, enabling or strengthening these filters ensures that inappropriate or unsafe images are not generated or returned.

B (Customize an LLM): Customization affects behavior but not safety filtering.

C (Configure the system prompt): Adjusts response style but doesn’t guarantee content safety.

D (Change the model): Different models have similar filter systems; protection level depends on filter configuration, not the model itself.

According to the Microsoft Azure AI Fundamentals (AI-900) official study materials, a webchat bot is defined as a conversational AI application that interacts with users through a web-based chat interface. It simulates human conversation using text (and sometimes voice) to answer questions, assist with transactions, or provide information automatically. Microsoft Learn’s “Describe features of common AI workloads” module highlights conversational AI as a major AI workload, where bots and virtual agents are used to provide automated, intelligent responses in real time through web, mobile, or messaging platforms.

In this scenario, the chatbot on the festival website provides immediate answers about scheduled events and ticket purchases. This aligns exactly with how a webchat bot operates — interacting with users through a website, handling repetitive inquiries, and providing consistent information without human intervention. This type of solution is commonly built using Azure Bot Service integrated with Azure Cognitive Services for Language, which allows the bot to understand user intent and respond naturally.

Let’s examine the other options to reinforce why D is correct:

A describes a text analytics or sentiment analysis scenario, not a conversational bot, because it classifies text sentiment but doesn’t “chat” with a user.

B is an example of machine translation using the Translator service, not a chatbot.

C is an email classification or natural language processing task, not a webchat interaction.

The AI-900 exam objectives clearly distinguish conversational AI from other cognitive services such as translation or sentiment analysis. Conversational AI focuses on dialogue and interaction through natural language conversation channels like websites or messaging apps.

Therefore, the verified and officially aligned answer is D. From a website interface, answer common questions about scheduled events and ticket purchases for a music festival.

According to the Microsoft Azure AI Fundamentals (AI-900) official study guide and the Microsoft Learn module “Describe core concepts of machine learning on Azure”, when building a predictive machine learning model, features are the input variables used by the algorithm to predict the target label. The target label is the output or value the model is trained to predict.

In this dataset, the target variable is clearly the Income Range, since the goal is to predict a customer’s income bracket. Therefore, Income Range (D) is the label, not a feature. Features must be other attributes that help the model make this prediction.

The fields Education Level (A) and Age (C) are the most relevant features because both can logically and statistically influence income level.

Education Level is a categorical variable that often correlates strongly with income. Individuals with higher education levels tend to earn more on average, making this an important predictor.

Age is a numerical variable that typically affects income level due to factors such as experience and career progression.

By contrast:

First Name (E) and Last Name (B) are irrelevant as features because they are identifiers, not meaningful predictors of income. Including them could lead to bias or model overfitting without contributing to accurate predictions.

Hence, according to AI-900 principles, the features used to train a model predicting income range would be Education Level and Age.

Object detection is similar to tagging, but the API returns the bounding box coordinates (in pixels) for each object found. For example, if an image contains a dog, cat and person, the Detect operation will list those objects together with their coordinates in the image. You can use this functionality to process the relationships between the objects in an image. It also lets you determine whether there are multiple instances of the same tag in an image.

The Detect API applies tags based on the objects or living things identified in the image. There is currently no formal relationship between the tagging taxonomy and the object detection taxonomy. At a conceptual level, the Detect API only finds objects and living things, while the Tag API can also include contextual terms like " indoor " , which can ' t be localized with bounding boxes.

In Microsoft Azure AI Language Service, both Named Entity Recognition (NER) and Key Phrase Extraction are core features for text analytics. They serve distinct purposes in analyzing and structuring unstructured text data.

Named Entity Recognition (NER):NER is used to identify and categorize specific entities within text, such as people, organizations, locations, dates, times, and quantities. According to Microsoft Learn’s “Analyze text with Azure AI Language” module, NER scans text to extract these entities along with their types. Therefore, the statement “Named entity recognition can be used to retrieve dates and times in a text string” is True (Yes).

Key Phrase Extraction:This feature identifies the most important phrases or main topics in a block of text. It is useful for summarization or highlighting central ideas without classifying them into specific categories. Therefore, the statement “Key phrase extraction can be used to retrieve important phrases in a text string” is also True (Yes).

City Name Retrieval:While key phrase extraction highlights major phrases, it does not extract specific entities like cities or dates. Extracting such details requires Named Entity Recognition, which is designed to find named entities such as city names, people, or organizations. Hence, the statement “Key phrase extraction can be used to retrieve all the city names in a text string” is False (No).

According to the Microsoft Azure AI Fundamentals (AI-900) official study guide and Microsoft Learn module “Identify features of regression machine learning”, regression is a type of supervised machine learning used when the target variable (the value you want to predict) is a continuous numeric value.

In this scenario, the task is to predict how many hours of overtime a delivery person will work based on the number of orders received. Both the input (number of orders) and the output (hours of overtime) are numeric variables. Since the goal is to estimate a quantitative value rather than categorize or group data, this is a classic example of a regression problem.

Regression models analyze the relationship between variables to make numerical predictions. For example, the model might learn that each additional 20 orders increases overtime by about two hours. Common algorithms used for regression include linear regression, decision tree regression, and boosted regression models. These models produce outputs such as “expected overtime = 5.6 hours,” which are continuous numeric results.

To contrast with the other options:

Classification is used for predicting categories or labels, such as “overtime required” vs. “no overtime,” or “high-risk” vs. “low-risk.” It deals with discrete outputs rather than continuous numbers.

Clustering is an unsupervised learning approach used to group similar data points based on shared characteristics, such as grouping delivery staff by performance patterns or customer types.

As emphasized in Microsoft’s Responsible AI and Machine Learning Fundamentals learning paths, regression models are ideal for numeric forecasting problems such as predicting sales, revenue, demand, or working hours.

Therefore, the correct answer is: Regression.

Generative AI focuses on creating new content rather than just analyzing existing data. As per Microsoft’s AI-900 curriculum and Azure OpenAI documentation, typical use cases include generating text, images, code, or other creative outputs based on input prompts.

A. Generating draft responses for customer service agents — ✅ Correct. GPT-based models can automatically generate draft replies to customer queries, enabling agents to refine responses and increase efficiency.

B. Creating original artwork from textual descriptions — ✅ Correct. DALL-E, available through Azure OpenAI, can produce unique images based on natural language prompts.

Options C and D are incorrect because they involve predictive or classification models, not generative ones:

C. Predicting sales revenue → Regression (machine learning).

D. Classifying email messages → Classification (machine learning).

Correct answers: A and B.

According to the Microsoft Azure OpenAI Service documentation and AI-900 official study materials, the DALL-E model is specifically designed to generate and edit images from natural language prompts. When a user provides a descriptive text input such as “a futuristic city skyline at sunset”, DALL-E interprets the textual prompt and produces an image that visually represents the content described. This functionality is known as text-to-image generation and is one of the creative AI capabilities supported by Azure OpenAI.

DALL-E belongs to the family of generative models that can create new visual content, expand existing images, or apply transformations to images based on textual instructions. Within Azure OpenAI, the DALL-E API enables developers to integrate image creation directly into applications—useful for design assistance, marketing content generation, or visualization tools. The model learns from vast datasets of text–image pairs and is optimized to ensure alignment, diversity, and accuracy in the produced visuals.

By contrast, the other options serve different purposes:

A. GPT-4 is a large language model for text-based generation, reasoning, and conversation, not for creating images.

C. GPT-3 is an earlier text generation model, primarily used for language tasks like summarization, classification, and question answering.

D. Whisper is an automatic speech recognition (ASR) model used to convert spoken language into written text; it has no image-generation capability.

Therefore, when the requirement is to generate images based on user prompts, the only Azure OpenAI model that fulfills this purpose is DALL-E. This aligns directly with the AI-900 learning objective covering Azure OpenAI generative capabilities for text, code, and image creation.

According to the Microsoft Azure AI Fundamentals (AI-900) official study guide and the Microsoft Learn module “Identify features of common machine learning types”, there are three main types of machine learning: supervised learning, unsupervised learning, and reinforcement learning. Within supervised learning, two common approaches are regression and classification, while clustering is a primary example of unsupervised learning.

“You train a regression model by using unlabeled data.” – No.Regression models are trained with labeled data, meaning the input data includes both features (independent variables) and target labels (dependent variables) representing continuous numerical values. Examples include predicting house prices or sales forecasts. Unlabeled data (data without target output values) cannot be used to train regression models; such data is used in unsupervised learning tasks like clustering.

“The classification technique is used to predict sequential numerical data over time.” – No.Classification is used for categorical predictions, where outputs belong to discrete classes, such as spam/not spam or disease present/absent. Predicting sequential numerical data over time refers to time series forecasting, which is typically a regression or forecasting problem, not classification. The AI-900 syllabus clearly separates classification (categorical prediction) from regression (continuous value prediction) and time series (temporal pattern analysis).

“Grouping items by their common characteristics is an example of clustering.” – Yes.This statement is correct. Clustering is an unsupervised learning technique used to group similar data points based on their features. The AI-900 study materials describe clustering as the process of “discovering natural groupings in data without predefined labels.” Common examples include customer segmentation or document grouping.

Therefore, based on Microsoft’s AI-900 training objectives and definitions:

Regression → supervised learning using labeled continuous data (No)

Classification → categorical prediction, not sequential numeric forecasting (No)

Clustering → grouping by similarity (Yes)

Full Detailed Explanation (250–300 words):

The correct answer is A. AI systems should treat people fairly.

This statement aligns with one of Microsoft’s six Responsible AI principles, which are:

Fairness – AI systems should treat all people fairly and avoid bias.

Reliability and Safety

Privacy and Security

Inclusiveness

Transparency

Accountability

The principle of Fairness ensures that AI models do not discriminate based on factors such as race, gender, age, or socioeconomic background. For example, a loan approval or hiring model must provide equal opportunity to all qualified applicants regardless of demographic differences.

B (Not revealing design details) contradicts Transparency, which promotes openness about AI functionality.

C (Black-box models) goes against Microsoft’s push for Explainable AI.

D (Protect developers’ interests) is not part of Microsoft’s Responsible AI framework.

Therefore, the verified correct answer is A. AI systems should treat people fairly.

In Azure, the Language service unifies several natural language capabilities, including LUIS, QnA Maker, and Text Analytics, into one comprehensive service. To determine a user’s intent in a chatbot, you use the Conversational Language Understanding (CLU) feature of the Language service, which is the evolution of LUIS.

CLU helps chatbots and applications comprehend natural language input by identifying the intent (the purpose of the user’s statement) and extracting entities (important details). For example, when a user types “Book a meeting for tomorrow,” the model recognizes the intent (BookMeeting) and the entity (tomorrow).

The other options do not determine intent:

Translator (A) is used for language translation.

Azure Cognitive Search (B) retrieves documents based on search queries.

Speech (C) converts audio to text but doesn’t analyze meaning.

Thus, to determine a user’s intent in a chatbot scenario, the correct service is D. Language.

According to the Microsoft Azure AI Fundamentals (AI-900) Official Study Guide and the Microsoft Learn module “Explore computer vision in Microsoft Azure,” Optical Character Recognition (OCR) is a computer vision capability that detects and extracts printed or handwritten text from images or scanned documents and converts it into machine-readable digital text.

In this scenario, a historian wants to digitize newspaper articles — which means converting physical or scanned images of printed text into digital text for easier searching, archiving, and analysis. This is exactly the function of OCR. By using OCR, the historian can take photos or scans of old newspapers and extract the words into editable digital documents, preserving valuable historical information.

OCR is a key feature of the Azure Computer Vision service, which provides capabilities such as:

Extracting text from images or PDFs.

Reading both printed and handwritten text in multiple languages.

Converting physical documents into searchable digital files.

Let’s examine the incorrect options:

Facial analysis: Detects facial features, age, gender, and emotions — unrelated to text extraction.

Image classification: Identifies what an image contains (e.g., “dog,” “car,” or “building”) but doesn’t extract text.

Object detection: Identifies and locates objects within an image using bounding boxes, not suitable for text recognition.

Therefore, to digitize newspaper articles and convert printed words into editable digital text, the correct technology to use is Optical Character Recognition (OCR), provided by the Azure Computer Vision API.

✅ Final Answer: optical character recognition (OCR)

According to the Microsoft Azure AI Fundamentals (AI-900) Official Study Guide and the Microsoft Learn module “Explore fundamental principles of machine learning,” a regression model is used when the goal is to predict a continuous numerical value based on historical data.

In this question, the task is to predict the sale price of auctioned items, which is a numeric output that can take on a wide range of values (for example, $50.25, $199.99, etc.). This makes it a regression problem because the output is continuous rather than categorical.

Regression models analyze the relationship between input features (such as item type, condition, age, bidding history, or demand) and a numerical target variable (the sale price). Common regression algorithms include linear regression, decision tree regression, and neural network regression. In Azure Machine Learning, these models are trained using labeled datasets containing known outcomes to learn patterns and make future predictions.

Let’s review the incorrect options:

Classification: Used to predict discrete categories or labels, such as “sold” vs. “unsold” or “low,” “medium,” “high.” It cannot output continuous numeric predictions.

Clustering: An unsupervised technique used to group similar data points based on shared characteristics, not to predict specific numeric outcomes.

Therefore, because predicting a sale price involves forecasting a continuous numerical value, the correct model type is Regression.

This aligns with Microsoft’s AI-900 teaching that regression is used for tasks such as:

Predicting house prices

Forecasting sales revenue

Estimating car values or auction prices

According to the Microsoft Azure AI Fundamentals (AI-900) curriculum and Azure OpenAI documentation, the appropriate service for detecting and managing harmful, unsafe, or inappropriate content in text, images, or other generative AI outputs is Azure AI Content Safety.

Azure AI Content Safety is designed to automatically detect potentially harmful material such as hate speech, violence, self-harm, sexual content, or profanity. It ensures that generative AI applications like chatbots, image generators, and content creation tools comply with Microsoft’s Responsible AI principles — specifically Reliability & Safety and Accountability.

This service integrates directly with the Azure OpenAI Service, meaning that when developers build AI solutions using models like GPT-4 or DALL·E, they can use Content Safety to filter and moderate both input prompts and model outputs. This protects users from unsafe or offensive content generation.

Let’s analyze why the other options are incorrect:

A. Face – The Face service detects and analyzes human faces in images or videos. It is unrelated to moderating harmful textual or generative content.

B. Video Analysis – This service analyzes video streams to detect objects, actions, or events but not inappropriate or harmful text or imagery from AI models.

C. Language – The Azure AI Language service focuses on text understanding tasks like sentiment analysis, entity recognition, and translation, not content safety filtering.

Therefore, per Microsoft Learn’s official AI-900 guidance, when identifying or filtering harmful content in a generative AI solution built with Azure OpenAI, the correct and verified service to use is Azure AI Content Safety.

In the context of Microsoft Azure AI Fundamentals (AI-900) and general machine learning principles, regression refers to a type of supervised learning used to predict continuous numerical values based on historical data. The goal of regression is to model the relationship between input variables (features) and a continuous output variable (target).

In this scenario, the task is to predict how many vehicles will travel across a bridge on a given day. The number of vehicles is a numerical value that can vary continuously depending on factors such as time of day, weather, weekday/weekend, or traffic trends. Because the output is numeric and not categorical, this problem type clearly fits into regression analysis.

Microsoft’s official learning content for AI-900, under “Identify features of regression and classification machine learning models,” specifies that regression models are used to predict values such as sales forecasts, demand estimation, temperature prediction, or traffic volume—all of which share the same underlying objective: predicting a quantity.

To clarify other options:

Classification is used when predicting categories or discrete classes, such as determining whether an email is spam or not spam, or if an image contains a cat or a dog.

Clustering is an unsupervised learning technique used to group similar data points without predefined labels (for example, grouping customers by purchasing behavior).

Since predicting the number of vehicles results in a continuous numerical output, it aligns precisely with the regression workload type described in the Microsoft AI-900 study materials.

This question tests understanding of Microsoft’s six guiding principles for Responsible AI, which are: fairness, reliability and safety, privacy and security, inclusiveness, transparency, and accountability. These principles, as described in the Microsoft Azure AI Fundamentals (AI-900) study guide and Microsoft Learn Responsible AI module, help ensure that AI systems are developed and used ethically and responsibly.

Transparency – Yes:Transparency means users should understand how and why an AI system makes certain decisions. Providing an explanation of the outcome of a credit loan application clearly supports transparency because it helps customers know the reasoning behind approval or rejection. According to Microsoft Learn, transparency ensures that “AI systems are understandable by users and stakeholders,” especially in sensitive applications such as finance and credit scoring. Thus, the first statement is Yes.

Reliability and Safety – Yes:The reliability and safety principle ensures AI systems perform consistently, safely, and as intended, even in complex or high-risk environments. A triage bot that prioritizes insurance claims based on injury type aligns with this principle—it must be accurate, dependable, and safe to ensure claims are processed correctly and not influenced by errors or faulty algorithms. Microsoft teaches that AI should be “reliable under expected and unexpected conditions” to prevent harm or misjudgment. Therefore, this statement is Yes.

Inclusiveness – No:Inclusiveness focuses on ensuring AI systems empower and benefit all users, especially those with different abilities or backgrounds. Offering an AI solution at different prices across sales territories is a business decision, not an ethical or inclusiveness principle issue. It does not relate to accessibility or equal participation of diverse users. Therefore, this final statement is No.

For evaluating a classification model, the appropriate metric from the options provided is the True Positive Rate (TPR), also known as Sensitivity or Recall. According to the Microsoft Azure AI Fundamentals (AI-900) official study guide and Microsoft Learn module “Evaluate model performance”, classification models are evaluated using metrics that measure how accurately the model predicts categorical outcomes such as “yes/no,” “spam/not spam,” or “approved/denied.”

The True Positive Rate measures the proportion of correctly identified positive cases out of all actual positive cases. Mathematically, it is expressed as:

True Positive Rate (Recall)=True PositivesTrue Positives + False Negatives\text{True Positive Rate (Recall)} = \frac{\text{True Positives}}{\text{True Positives + False Negatives}}True Positive Rate (Recall)=True Positives + False NegativesTrue Positives​

This metric is important when missing positive predictions carries a high cost, such as in medical diagnosis or fraud detection. Microsoft Learn highlights classification evaluation metrics such as accuracy, precision, recall, F1 score, and AUC (Area Under the Curve) as suitable for classification models.

The other options—Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and Coefficient of Determination (R²)—are regression metrics used to evaluate models that predict numeric values rather than categories. For example, they apply to predicting house prices or temperatures, not yes/no decisions.

Therefore, the correct classification evaluation metric among the choices is A. True Positive Rate.

Box1: Sentiment analysis

Sentiment Analysis is the process of determining whether a piece of writing is positive, negative or neutral.

Box 2: Broad entity extraction

Broad entity extraction: Identify important concepts in text, including key

Key phrase extraction/ Broad entity extraction: Identify important concepts in text, including key phrases and named entities such as people, places, and organizations.

Box 3: Entity Recognition

Named Entity Recognition: Identify and categorize entities in your text as people, places, organizations, date/time, quantities, percentages, currencies, and more. Well-known entities are also recognized and linked to more information on the web.

According to the Microsoft Azure AI Fundamentals (AI-900) official study guide and Microsoft Learn documentation for Azure AI Custom Vision, this service is a specialized part of the Azure AI Vision family that enables developers to train custom image classification and object detection models. It allows organizations to build tailored computer vision models that recognize images or specific objects relevant to their business needs.

Detect objects in an image → YesThe Azure AI Custom Vision service supports both image classification (assigning an image to one or more categories) and object detection (identifying and locating objects within an image using bounding boxes). This means it can indeed detect and differentiate multiple objects in a single image, making this statement true.

Requires your own data to train the model → YesThe Custom Vision service is designed to be customizable. Unlike prebuilt Azure AI Vision models that work out of the box, Custom Vision requires you to upload and label your own dataset for training. The model then learns from your examples to perform specialized image recognition tasks relevant to your domain. Thus, this statement is also true.

Analyze video files → NoWhile Custom Vision can analyze images, it does not directly process or analyze video files. Video analysis is handled by a different service—Azure Video Indexer—which can extract insights such as spoken words, scenes, and faces from videos.

In summary:

✅ Yes – Detect objects in images

✅ Yes – Requires your own data

❌ No – Does not analyze video files.

Yes, Yes, No.

According to the Microsoft Azure AI Fundamentals (AI-900) study materials, conversational AI enables applications, websites, and digital assistants to interact with users via natural language. A chatbot is a key conversational AI workload and can be integrated into multiple channels such as web pages, Microsoft Teams, Facebook Messenger, and Cortana using Azure Bot Service and Bot Framework.

“A restaurant can use a chatbot to answer queries through Cortana” — Yes.Azure Bot Service supports multi-channel deployment, which includes Cortana integration. This means the same bot can respond to voice or text input via Cortana, making it a valid use case for a restaurant to provide menu details, reservations, or order tracking through voice-based AI assistants.

“A restaurant can use a chatbot to answer inquiries about business hours from a webpage” — Yes.This is a standard scenario for chatbots embedded on a company website. As per Microsoft Learn’s Describe features of conversational AI module, a chatbot can be added to a website to handle FAQs such as business hours, location, or menu details, thereby improving response time and reducing repetitive human workload.

“A restaurant can use a chatbot to automate responses to customer reviews on an external website” — No.Azure bots and other conversational AI tools cannot automatically interact with or post on external third-party platforms where the business does not control the data or API integration. Automated posting or replying to reviews on external review sites (e.g., Yelp or Google Reviews) would violate both ethical and technical boundaries of responsible AI usage outlined by Microsoft.

According to the Microsoft Azure AI Fundamentals (AI-900) official study guide and the Microsoft Learn module “Identify features of Natural Language Processing (NLP) workloads on Azure”, Azure Cognitive Services provides several text analytics and language understanding workloads that perform different types of language processing tasks. Each workload extracts specific information or performs distinct analysis operations on text data.

Entity Recognition → Extracts persons, locations, and organizations from the textEntity recognition is a feature of Azure Cognitive Service for Language (formerly Text Analytics). It identifies and categorizes named entities in unstructured text, such as people, organizations, locations, dates, and more. The study guide defines this workload as: “Entity recognition locates and classifies named entities in text into predefined categories.” This allows applications to extract structured information from raw text data—for example, identifying “Microsoft” as an organization and “Seattle” as a location.

Sentiment Analysis → Evaluates text along a positive–negative scaleSentiment analysis determines the emotional tone or opinion expressed in a piece of text. It classifies text as positive, negative, neutral, or mixed, which is widely used for social media monitoring, customer feedback, and product reviews. Microsoft’s official documentation describes it as: “Sentiment analysis evaluates text and returns a sentiment score indicating whether the sentiment is positive, negative, neutral, or mixed.”

Translation → Returns text translated to the specified target languageThe Translator service, part of Azure Cognitive Services, automatically translates text from one language to another. It supports multiple languages and provides near real-time translation. The AI-900 content specifies that “translation workloads are used to automatically translate text between languages using machine translation models.”

In summary:

Entity Recognition → Extracts entities like names and locations.

Sentiment Analysis → Determines emotional tone.

Translation → Converts text between languages.

✅ Final Answers:

Extracts persons, locations, and organizations → Entity recognition

Evaluates text along a positive–negative scale → Sentiment analysis

Returns text translated to the specified target language → Translation

According to the Microsoft Azure AI Fundamentals (AI-900) learning path and Microsoft Learn documentation on GitHub Copilot, GitHub Copilot is powered by OpenAI Codex, a specialized language model derived from the GPT-3 family but fine-tuned specifically on programming languages and code data.

OpenAI Codex was designed to translate natural language prompts into executable code in multiple programming languages, including JavaScript, Python, C#, TypeScript, and Go. It can understand comments, function names, and code structure to generate relevant code suggestions in real time.

When a developer writes client-side JavaScript, GitHub Copilot uses Codex to analyze the context of the file and generate intelligent suggestions, such as completing functions, writing boilerplate code, or suggesting improvements. Codex can also explain what specific code does and provide inline documentation, which enhances developer productivity.

Option A (GPT-4): While some newer versions of GitHub Copilot (Copilot X) may integrate GPT-4 for conversational explanations, the core code completion engine remains based on Codex, as per the AI-900-level content.

Option C (DALL-E): Used for image generation, not for programming tasks.

Option D (GPT-3): Codex was fine-tuned from GPT-3 but has been further trained specifically for code generation tasks.

Therefore, the verified and official answer from Microsoft’s AI-900 curriculum is B. Codex — the OpenAI model used by GitHub Copilot to make suggestions for client-side JavaScript and other programming languages.

Model Explain Ability.

Most businesses run on trust and being able to open the ML “black box” helps build transparency and trust. In heavily regulated industries like healthcare and banking, it is critical to comply with regulations and best practices. One key aspect of this is understanding the relationship between input variables (features) and model output. Knowing both the magnitude and direction of the impact each feature (feature importance) has on the predicted value helps better understand and explain the model. With model explain ability, we enable you to understand feature importance as part of automated ML runs.

The correct answers are B. a smart device in the home that responds to questions such as, " What will the weather be like today? " and D. a website that uses a knowledge base to interactively respond to users ' questions.

According to the Microsoft Azure AI Fundamentals (AI-900) official study guide and Microsoft Learn module “Identify features of Natural Language Processing (NLP) workloads on Azure”, Natural Language Processing (NLP) is a branch of artificial intelligence that enables computers to understand, interpret, and generate human language in a meaningful way. NLP bridges the gap between human communication and machine understanding, allowing systems to process both spoken and written language.

Option B – A smart device in the home that responds to questions such as “What will the weather be like today?”This is an example of an NLP workload because the device must process spoken language (speech-to-text), interpret the user’s intent (language understanding), and generate a relevant spoken response (text-to-speech). This workflow involves several Azure Cognitive Services, such as Speech Service for recognizing and synthesizing speech, and Language Understanding (LUIS) for interpreting intent. This aligns with conversational AI and NLP tasks in the AI-900 syllabus.

Option D – A website that uses a knowledge base to interactively respond to users’ questions.This is also an NLP workload because the system interprets text input from users and retrieves appropriate answers from a knowledge base. Microsoft’s QnA Maker (now part of the Azure AI Language service) and Azure Bot Service enable such behavior. The model uses NLP to understand the user’s question, find the most relevant response, and generate an appropriate reply — key characteristics of natural language processing.

Incorrect options:

A (assembly line machinery) represents automation or robotics, not NLP.

C (monitoring temperature to activate a fan) is an example of an IoT (Internet of Things) or rule-based system, not related to language processing.

According to the Microsoft Azure AI Fundamentals (AI-900) official study guide and Microsoft Learn modules on machine learning concepts, ensuring that the accuracy of a predictive model can be proven requires data partitioning—specifically splitting the available data into training and testing datasets. This is a foundational concept in supervised machine learning.

When you split the data, typically about 70–80% of the dataset is used for training the model, while the remaining 20–30% is used for testing (or validation). The reason behind this approach is to ensure that the model’s performance metrics—such as accuracy, precision, recall, and F1-score—are evaluated on data the model has never seen before. This prevents overfitting and allows you to demonstrate that the model generalizes well to new, unseen data.

In the AI-900 Microsoft Learn content under “Describe the machine learning process”, it is explained that after cleaning and transforming the data, the next essential step is data splitting to “evaluate model performance objectively.” By keeping training and testing data separate, you can prove the reliability and accuracy of the model’s predictions, which is particularly crucial in sensitive domains like clinical or healthcare analytics, where decision transparency and validation are vital.

Option A (Train the model by using the clinical data) is incorrect because you should not train and evaluate on the same data—it would lead to biased results.

Option C (Train the model using automated ML) is incorrect because automated ML is a method for training and tuning, but it doesn’t inherently prove accuracy.

Option D (Validate the model by using the clinical data) is also incorrect if you use the same dataset for validation and training—it would not prove true accuracy.

Therefore, per Microsoft’s official AI-900 study content, the verified correct answer is B. Split the clinical data into two datasets.

Inclusiveness: At Microsoft, we firmly believe everyone should benefit from intelligent technology, meaning it must incorporate and address a broad range of human needs and experiences. For the 1 billion people with disabilities around the world, AI technologies can be a game-changer.

According to the Microsoft Azure AI Fundamentals (AI-900) study material and Azure OpenAI Service documentation, large language models (LLMs) such as GPT are capable of performing multiple natural language processing (NLP) tasks depending on the intent of the prompt. These tasks generally fall into categories like classification, generation, summarization, and translation, each with a distinct purpose and output type.

Classifying – This process involves analyzing text and assigning it to a predefined category or label based on its content. The scenario “Detect the genre of a work of fiction” clearly fits this category. The model must evaluate the text and determine whether it belongs to genres like mystery, romance, or science fiction. This is a classic text classification problem, as the output is a discrete category derived from textual features.

Summarizing – This process means condensing lengthy text into a shorter version that preserves the key information. In the scenario “Create a list of bullet points based on text input,” the model extracts essential information and reformats it as concise bullet points, which is an abstraction form of summarization. Summarization models help users quickly understand the main ideas from long documents, meeting efficiency and readability goals.

Generating – This refers to the LLM’s ability to produce new, creative content based on input instructions. The task “Create advertising slogans from a product description” represents generation because it requires the model to construct original text that didn’t previously exist. Generation tasks showcase the creativity and contextual fluency of models like GPT in marketing and content creation.

Thus, these mappings align directly with the Azure OpenAI LLM capabilities taught in AI-900, linking each NLP process with its most suitable real-world task.

According to the Microsoft Azure AI Fundamentals (AI-900) official study guide and the Microsoft Learn module “Identify features of common AI workloads”, OCR (Optical Character Recognition) is a Computer Vision technology that detects and extracts printed or handwritten text from images and scanned documents. OCR allows organizations and individuals to convert physical or image-based text into machine-readable, editable, and searchable digital text.

In the context of this question, a historian working with old newspaper articles or archival documents would use OCR to digitize printed content. For instance, the historian can scan or photograph old newspaper pages, and then use an OCR tool—such as Azure Computer Vision’s OCR API—to automatically recognize and extract the textual content from those images. This process enables the historian to store, edit, and analyze the content digitally without manually typing everything.

OCR works by using deep learning algorithms trained on thousands of text samples. The system analyzes patterns, shapes, and spatial relationships of characters to identify text accurately, even from low-quality or aged paper documents. Once extracted, the digital text can be indexed, translated, or processed further using Natural Language Processing (NLP) tools for content analysis.

Now, addressing the other options:

Facial analysis is used to detect emotions, age, or gender from human faces—irrelevant to text digitization.

Image classification identifies entire images by categories (e.g., cat, car, flower).

Object detection identifies and locates multiple objects within an image but doesn’t extract text.

Therefore, per the AI-900 learning objectives under the Computer Vision workload, the correct and verified completion is:

Statements

Yes

No

A bot that responds to queries by internal users is an example of a conversational AI workload.

✅ Yes

An application that displays images relating to an entered search term is an example of a conversational AI workload.

✅ No

A web form used to submit a request to reset a password is an example of a conversational AI workload.

✅ No

According to the Microsoft Azure AI Fundamentals (AI-900) official study materials, conversational AI workloads are those that enable interaction between humans and AI systems through natural language conversation, either by text or speech. These workloads are typically implemented using Azure Bot Service, Azure Cognitive Services for Language, and Azure OpenAI Service. The key characteristic of a conversational AI workload is the presence of dialogue—the AI interprets user intent and provides a meaningful, contextual response in a conversation-like manner.

“A bot that responds to queries by internal users is an example of a conversational AI workload.” → YESThis fits the definition perfectly. A chatbot that helps employees (internal users) by answering questions about policies, IT issues, or HR procedures is a typical example of conversational AI. It uses natural language understanding to interpret questions and provide automated responses. Microsoft Learn explicitly identifies chatbots as conversational AI solutions designed for both internal and external interactions.

“An application that displays images relating to an entered search term is an example of a conversational AI workload.” → NOThis is not conversational AI because there is no dialogue or language understanding involved. It is an example of information retrieval or computer vision if it uses image recognition, but not conversation.

“A web form used to submit a request to reset a password is an example of a conversational AI workload.” → NOA password reset form is a simple UI-driven process that doesn’t require AI or conversational logic. It performs a fixed function based on user input but does not understand or respond to natural language.

Therefore, based on the AI-900 study guide, only the first statement is an example of a conversational AI workload, while the second and third statements are not.

Reliability and safety: To build trust, it ' s critical that AI systems operate reliably, safely, and consistently under normal circumstances and in unexpected conditions. These systems should be able to operate as they were originally designed, respond safely to unanticipated conditions, and resist harmful manipulation.

According to the Microsoft Azure AI Fundamentals (AI-900) official study guide and Microsoft Learn module “Describe features of natural language processing (NLP) workloads on Azure,” sentiment analysis is an NLP workload that determines the emotional tone or opinion expressed in a piece of text. This could be positive, negative, or neutral sentiment.

When an app analyzes social media posts to identify their tone, it is performing sentiment analysis, since it aims to understand the emotional context behind user-generated text such as tweets, reviews, or comments. Azure provides this functionality through the Azure Cognitive Services – Text Analytics API, which evaluates text and returns sentiment scores.

Other options are not suitable:

Key phrase extraction identifies main ideas in text but not tone.

Entity recognition identifies names of people, organizations, or locations.

Speech recognition converts spoken words into text, not emotional analysis.

Therefore, analyzing social media tone is an example of sentiment analysis, a key NLP workload in Microsoft’s AI-900 syllabus.

This question examines your understanding of Natural Language Processing (NLP) as described in the Microsoft Azure AI Fundamentals (AI-900) official study guide and Microsoft Learn module “Explore natural language processing.” NLP is a branch of artificial intelligence that enables computers to analyze, understand, and generate human language — both written and spoken. Typical NLP tasks include text analytics, language understanding, sentiment analysis, key phrase extraction, and profanity detection.

Monitoring online service reviews for profanities → YesThis is a classic example of NLP. Detecting profane or inappropriate words in customer reviews requires analyzing text content. Azure Cognitive Services offers Content Moderator and Text Analytics APIs that can detect and filter profanity, sentiment, and offensive language automatically. Microsoft Learn states: “Natural language processing is used to process and analyze text to detect sentiment, key phrases, and inappropriate content.” Hence, this task is correctly classified as NLP.

Identifying brand logos in an image → NoThis task belongs to Computer Vision, not NLP. The Computer Vision API and Custom Vision service in Azure are designed to detect and classify visual elements like logos, objects, or scenes. Since it involves images, not text, it is unrelated to natural language processing.

Monitoring public news sites for negative mentions of a product → YesThis is another valid example of NLP. The process involves analyzing the sentiment of text from online articles to determine whether mentions of a product are positive, neutral, or negative. Azure Text Analytics provides prebuilt sentiment analysis and entity recognition capabilities that help automate such monitoring.

Full Detailed Explanation (250–300 words):

“You can fine-tune some Azure OpenAI models by using your own data.” – YESThis statement is true. Azure OpenAI allows customers to fine-tune certain models like GPT-3, GPT-3.5, and some embedding models with their own data. Fine-tuning customizes a model to perform better on specific tasks or match a company’s domain terminology, tone, or context. According to Microsoft Learn’s AI-900 and Azure OpenAI documentation, fine-tuning is supported for approved use cases while maintaining Microsoft’s Responsible AI oversight and compliance process.

“Pretrained generative AI models are a component of Azure OpenAI.” – YESThis statement is also true. Azure OpenAI provides access to pretrained large language and generative AI models such as GPT-3.5, GPT-4, Codex, and DALL·E. These models are pretrained on vast datasets and made available via APIs, allowing developers to generate text, code, and images without needing to train their own models. This is a core feature of Azure OpenAI’s service offering.

“To build a solution that complies with Microsoft responsible AI principles, you must build and train your own model.” – NOThis statement is false. Compliance with Microsoft Responsible AI principles (Fairness, Reliability & Safety, Privacy & Security, Inclusiveness, Transparency, Accountability) does not require building custom models. Prebuilt Azure AI and OpenAI services already align with Responsible AI standards. Developers simply need to use these services responsibly, applying governance and ethical design practices.

Classification

According to the Microsoft Azure AI Fundamentals (AI-900) official study guide and the Microsoft Learn module “Describe features of common AI workloads”, classification is a supervised machine learning technique used when the goal is to predict which category or class an item belongs to. In supervised learning, the model is trained with labeled data—data that already contains known outcomes. The system learns patterns and relationships between input features and their corresponding labels so it can predict future classifications accurately.

In the scenario provided — “A banking system that predicts whether a loan will be repaid” — the model’s output is a binary decision, meaning there are two possible outcomes:

The loan will be repaid (positive class)

The loan will not be repaid (negative class)

This kind of problem involves predicting a discrete value (a label or category), not a continuous numeric output. Therefore, it perfectly fits the classification type of machine learning.

The AI-900 learning materials describe classification as being used in many real-world examples, including:

Determining whether an email is spam or not spam.

Predicting whether a customer will churn (leave) or stay.

Detecting fraudulent transactions.

Assessing medical test results as positive or negative.

By contrast:

Regression predicts continuous numeric values, such as predicting house prices, temperatures, or sales revenue. It would not apply here because repayment prediction is not a numeric value but a categorical decision.

Clustering is an unsupervised learning method that groups similar data points without predefined categories, such as segmenting customers by purchasing behavior.

Thus, based on Microsoft’s Responsible AI and AI-900 study guide concepts, a banking system that predicts whether a loan will be repaid uses the Classification type of machine learning.