Tag: Microsoft Certified: Azure AI Fundamentals

AI, AI-900, Artificial Intelligence (AI), Microsoft Certification, Natural Language Processing (NLP) January 31, 2026

Identify Features and Uses for Translation (AI-900 Exam Prep)

Where This Topic Fits in the Exam

Exam area: Describe features of Natural Language Processing (NLP) workloads on Azure (15–20%)
Sub-area: Identify features of common NLP workload scenarios
Skill focus: Recognizing when translation is the appropriate NLP workload, and understanding Azure services that support it

Translation is a core NLP workload on the AI-900 exam and often appears in short, scenario-based questions.

What Is Translation in NLP?

Translation is the process of converting text (or speech) from one language into another while preserving the original meaning.

Modern AI-powered translation systems use machine learning and deep learning models to understand context, grammar, and semantics rather than performing word-for-word substitutions.

Key Features of Translation Workloads

Translation solutions typically provide the following features:

Text-to-text translation between languages
Support for dozens of languages and dialects
Context-aware translation (not literal word replacement)
Detection of source language
Batch or real-time translation
Integration with applications, websites, and chatbots
Optional customization for domain-specific terminology

Common Uses of Translation

Translation workloads are used whenever language differences create a communication barrier.

Typical scenarios include:

Translating websites or product documentation
Supporting multilingual customer service
Translating chat messages in real time
Localizing applications for global users
Translating social media posts or reviews
Enabling communication across international teams

Azure Services for Translation

In Azure, translation capabilities are provided by:

Azure AI Translator

Azure AI Translator is part of Azure AI Services and offers:

Text translation between supported languages
Language detection
Transliteration (converting text between scripts)
Dictionary lookup and examples
Real-time and batch translation via APIs

This service uses prebuilt models, so no training is required.

Translation vs Other NLP Workloads

It is important to distinguish translation from similar NLP tasks:

NLP Task	Purpose
Translation	Convert text from one language to another
Language detection	Identify which language text is written in
Speech recognition	Convert spoken audio into text
Speech synthesis	Convert text into spoken audio
Sentiment analysis	Identify emotional tone of text

Translation and Speech

Translation workloads may involve:

Text-to-text translation (most common on AI-900)
Speech translation, which combines:
1. Speech recognition
2. Translation
3. Speech synthesis

On the exam, focus primarily on text translation scenarios, unless speech is explicitly mentioned.

Responsible AI Considerations

Translation systems should be designed with responsible AI principles in mind:

Fairness: Avoid biased or culturally inappropriate translations
Reliability: Handle idioms and context accurately
Transparency: Clearly indicate when content is machine-translated
Privacy: Protect sensitive or personal information in translated text

Exam Clues to Watch For

On AI-900, translation workloads are commonly signaled by phrases such as:

“Convert content from one language to another”
“Support multilingual users”
“Translate customer messages”
“Localize an application”

When these appear, translation is the correct NLP workload.

Key Takeaways for AI-900

Translation is an NLP workload that converts text between languages
Azure AI Translator is the primary Azure service for translation
No model training is required
Translation is different from sentiment analysis, entity recognition, and speech workloads
Exam questions are typically scenario-based and concise

Go to the Practice Exam Questions for this topic.

Go to the AI-900 Exam Prep Hub main page.

AI, AI-900, Artificial Intelligence (AI), Machine Learning (ML), Microsoft Certification January 31, 2026

Practice Questions: Identify Features and Labels in a Dataset for Machine Learning (AI-900 Exam Prep)

Practice Exam Questions

Question 1

You are training a model to predict house prices. The dataset includes columns for square footage, number of bedrooms, location, and sale price.
Which column is the label?

A. Square footage
B. Number of bedrooms
C. Location
D. Sale price

Correct Answer: D

Explanation:
The label is the value the model is trained to predict. In this scenario, the goal is to predict the sale price.

Question 2

Which statement best describes a feature in a machine learning dataset?

A. The final prediction made by the model
B. An input value used to make predictions
C. A rule written by a developer
D. The accuracy of the model

Correct Answer: B

Explanation:
Features are the input variables that provide information the model uses to make predictions.

Question 3

A dataset contains customer age, subscription length, monthly charges, and whether the customer canceled the service.
What is the label?

A. Customer age
B. Subscription length
C. Monthly charges
D. Whether the customer canceled

Correct Answer: D

Explanation:
The label represents the outcome being predicted—in this case, whether the customer canceled the service.

Question 4

Which type of machine learning requires both features and labels?

A. Unsupervised learning
B. Reinforcement learning
C. Supervised learning
D. Clustering

Correct Answer: C

Explanation:
Supervised learning uses labeled data so the model can learn the relationship between features and known outcomes.

Question 5

A dataset is used to group customers based on purchasing behavior, but it does not contain any target outcome.
What does this dataset contain?

A. Labels only
B. Features only
C. Training results
D. Predictions

Correct Answer: B

Explanation:
Unsupervised learning datasets contain features but do not include labels.

Question 6

In an email spam detection dataset, which item would most likely be a feature?

A. Spam or not spam
B. Model accuracy score
C. Number of words in the email
D. Final prediction

Correct Answer: C

Explanation:
The number of words is an input characteristic used by the model to make predictions, making it a feature.

Question 7

Which statement about labels is TRUE?

A. Labels are optional in supervised learning
B. Labels are the inputs used by the model
C. Labels represent the value the model predicts
D. Labels are created after predictions are made

Correct Answer: C

Explanation:
Labels are the known outcomes the model is trained to predict in supervised learning scenarios.

Question 8

You are preparing data in Azure Machine Learning to predict product demand.
Which columns should be selected as features?

A. Only the column you want to predict
B. All columns except the target outcome
C. Only numerical columns
D. Only categorical columns

Correct Answer: B

Explanation:
Features are the input columns used to predict the target outcome, which is the label.

Question 9

A dataset includes the following columns: temperature, humidity, wind speed, and weather condition.
If the goal is to predict the weather condition, what are temperature, humidity, and wind speed?

A. Labels
B. Predictions
C. Features
D. Outputs

Correct Answer: C

Explanation:
These values are inputs used to predict the weather condition, making them features.

Question 10

Which scenario best represents a labeled dataset?

A. Customer data grouped by similarity
B. Sensor readings without outcomes
C. Product reviews with sentiment categories
D. Website logs without classifications

Correct Answer: C

Explanation:
Product reviews with sentiment categories include known outcomes, which are labels, making the dataset labeled.

Exam Pattern Tip

On AI-900:

Features = inputs
Labels = outputs
If labels exist → supervised learning
If no labels → unsupervised learning

If you can identify those quickly, you’ll eliminate most wrong answers immediately.

Go to the AI-900 Exam Prep Hub main page.

AI, AI-900, Artificial Intelligence (AI), Computer Vision, Microsoft Certification January 31, 2026

Describe Capabilities of the Azure AI Vision Service (AI-900 Exam Prep)

Overview

Azure AI Vision is Microsoft’s prebuilt computer vision service that enables applications to analyze images and videos without requiring machine learning expertise or custom model training. It provides REST APIs and SDKs that allow developers to easily extract visual insights such as objects, text, faces, and image descriptions.

For the AI-900 exam, you are expected to understand what Azure AI Vision can do, which problems it solves, and how it differs from custom vision solutions—not how to build or tune models.

What Is Azure AI Vision?

Azure AI Vision is part of Azure AI Services and offers ready-to-use computer vision capabilities, including:

Image analysis
Optical Character Recognition (OCR)
Facial detection and analysis
Object detection
Image tagging and categorization

These capabilities are powered by Microsoft-trained deep learning models and are accessed via APIs.

Core Capabilities of Azure AI Vision

1. Image Analysis

Azure AI Vision can analyze images to extract high-level insights, such as:

Objects present in an image (for example, car, building, person)
Scene descriptions in natural language
Image tags and categories
Visual features such as color distribution

Example use cases:

Auto-generating image captions
Content moderation
Organizing image libraries

👉 Exam tip: Image analysis describes what is in an image, not where every object is located with precision.

2. Object Detection

Object detection identifies specific objects in an image and returns:

Object names
Bounding box coordinates
Confidence scores

Example use cases:

Detecting vehicles in traffic images
Identifying products on store shelves

👉 Exam tip: Object detection includes location + object type, unlike image classification which only labels the image as a whole.

3. Optical Character Recognition (OCR)

OCR extracts printed and handwritten text from images and documents.

Azure AI Vision OCR supports:

Multiple languages
Structured and unstructured text
Images, screenshots, and scanned documents

Example use cases:

Digitizing receipts
Reading license plates
Extracting text from scanned forms

👉 Exam tip: OCR is about reading text, not understanding its meaning.

4. Facial Detection and Facial Analysis

Azure AI Vision can detect human faces in images and analyze non-identifying facial attributes, such as:

Face location (bounding boxes)
Facial landmarks
Estimated age range
Facial expressions
Accessories (glasses, masks)

⚠️ It does NOT identify individuals.

Example use cases:

Blurring faces for privacy
Counting people in images
Analyzing expressions in photos

👉 Exam tip:

Facial detection = where faces are
Facial analysis = attributes of faces
Facial recognition = identity (not required for AI-900)

5. Image Tagging and Categorization

Azure AI Vision automatically assigns tags and categories to images, such as:

“outdoor”
“food”
“animal”

These tags help with searchability and organization.

Example use cases:

Image indexing
Content filtering
Metadata enrichment

👉 Exam tip: Tagging helps describe images at a high level, not detect precise objects.

Azure AI Vision vs Custom Vision

Feature	Azure AI Vision	Azure Custom Vision
Prebuilt models	✅ Yes	❌ No
Requires training	❌ No	✅ Yes
Quick setup	✅ Yes	❌ No
Specialized scenarios	❌ Limited	✅ Strong
AI-900 focus	✅ Yes	⚠️ Limited

👉 Exam takeaway:
If the question mentions no training, quick setup, or prebuilt models, Azure AI Vision is usually the right answer.

Responsible AI Considerations

Because Azure AI Vision can analyze images of people, Microsoft emphasizes:

Privacy and security of image data
Transparency in how visual data is processed
Fairness and bias mitigation
Appropriate use of facial analysis

👉 Exam tip: Facial capabilities often pair with Responsible AI principles in exam questions.

Common AI-900 Exam Scenarios

You should recognize Azure AI Vision when the scenario involves:

Analyzing images without training a model
Extracting text from images
Detecting faces but not identifying people
Automatically tagging or describing images

Key Exam Takeaways

Azure AI Vision is a prebuilt computer vision service
No machine learning expertise required
Supports image analysis, OCR, object detection, and facial analysis
Focuses on insight extraction, not identity
Frequently tested in scenario-based questions

Go to the Practice Exam Questions for this topic.

Go to the AI-900 Exam Prep Hub main page.

AI, AI-900, Artificial Intelligence (AI), Microsoft Certification, Natural Language Processing (NLP) January 31, 2026

Practice Questions: Describe Capabilities of the Azure AI Language Service (AI-900 Exam Prep)

Practice Exam Questions

Question 1

Which Azure service should you use to analyze customer reviews and determine whether the feedback is positive or negative?

A. Azure Translator
B. Azure AI Vision
C. Azure AI Language
D. Azure Speech

Correct Answer: C

Explanation:
Sentiment analysis is a text analytics capability, which is provided by the Azure AI Language service. Translator is for language conversion, Vision is for images, and Speech is for audio.

Question 2

You want to extract people, organizations, and locations from text documents. Which Azure AI Language capability should you use?

A. Key phrase extraction
B. Named entity recognition
C. Text classification
D. Language detection

Correct Answer: B

Explanation:
Named Entity Recognition (NER) identifies and categorizes entities such as people, organizations, and locations within text.

Question 3

A company wants to automatically identify the main topics discussed in customer feedback emails. Which Azure AI Language feature should be used?

A. Sentiment analysis
B. Entity recognition
C. Key phrase extraction
D. Language detection

Correct Answer: C

Explanation:
Key phrase extraction identifies the most important concepts or talking points in text, making it ideal for summarizing feedback.

Question 4

Which scenario is best suited for the Azure AI Language service?

A. Converting spoken audio into text
B. Translating documents from English to French
C. Analyzing written text for sentiment and meaning
D. Detecting objects in images

Correct Answer: C

Explanation:
The Azure AI Language service specializes in understanding and analyzing text, including sentiment, entities, and key phrases.

Question 5

Which capability of Azure AI Language allows an application to answer natural language questions based on provided documents or FAQs?

A. Sentiment analysis
B. Question answering
C. Key phrase extraction
D. Language detection

Correct Answer: B

Explanation:
Question answering enables applications and chatbots to respond to user questions using structured knowledge sources.

Question 6

A multilingual application needs to determine the language of user-submitted text before processing it further. Which Azure AI Language feature should be used?

A. Translation
B. Language detection
C. Text classification
D. Entity recognition

Correct Answer: B

Explanation:
Language detection identifies the language of input text and is often used before other NLP operations.

Question 7

Which Azure service combines sentiment analysis, entity recognition, and key phrase extraction into a single offering?

A. Azure Translator
B. Azure Speech
C. Azure AI Vision
D. Azure AI Language

Correct Answer: D

Explanation:
The Azure AI Language service provides multiple NLP capabilities under one unified service.

Question 8

You want to categorize incoming support tickets into “Billing,” “Technical,” or “General.” Which Azure AI Language capability should you use?

A. Sentiment analysis
B. Key phrase extraction
C. Text classification
D. Language detection

Correct Answer: C

Explanation:
Text classification assigns predefined categories or labels to text content.

Question 9

Which statement best describes the Azure AI Language service?

A. It focuses only on speech-to-text conversion
B. It translates text between languages
C. It extracts meaning and insights from text
D. It analyzes images and video streams

Correct Answer: C

Explanation:
Azure AI Language is designed to analyze text and extract meaning, including sentiment, entities, and key concepts.

Question 10

Which task would NOT typically be handled by the Azure AI Language service?

A. Identifying sentiment in customer reviews
B. Extracting organizations from news articles
C. Translating text from Spanish to English
D. Detecting key phrases in feedback forms

Correct Answer: C

Explanation:
Text translation is handled by Azure Translator, not the Azure AI Language service. The other tasks are core Language service capabilities.

Final Exam Tips for This Topic

If the question involves understanding text, think Azure AI Language
If the question involves translation, think Azure Translator
If the question involves speech, think Azure Speech
AI-900 questions focus on what the service does, not how to code it

Go to the AI-900 Exam Prep Hub main page.

AI, AI-900, Artificial Intelligence (AI), Microsoft Certification, Natural Language Processing (NLP) January 31, 2026

Describe Capabilities of the Azure AI Language Service (AI-900 Exam Prep)

Where This Fits in the Exam

Exam: AI-900 – Microsoft Azure AI Fundamentals
Domain: Describe features of Natural Language Processing (NLP) workloads on Azure (15–20%)
Sub-area: Identify Azure tools and services for NLP workloads

At this level, the exam focuses on what the service does, when to use it, and how it differs from other Azure AI services—not on implementation or coding.

What Is the Azure AI Language Service?

The Azure AI Language service is a cloud-based NLP service that enables applications to understand, analyze, and extract meaning from text.

It brings together several NLP capabilities under a single unified service, making it easier to build text-based AI solutions such as:

Customer feedback analysis
Chatbots
Document processing
Knowledge mining

For AI-900, think of it as “the main Azure service for understanding text.”

Key Capabilities of the Azure AI Language Service

1. Text Analytics

Text Analytics allows applications to analyze raw text and extract insights.

Main features include:

Sentiment analysis
Key phrase extraction
Named entity recognition
Language detection

These features are widely tested on the exam.

2. Sentiment Analysis

What it does:
Determines whether text expresses a positive, negative, neutral, or mixed sentiment.

Example use cases:

Analyzing customer reviews
Measuring brand perception on social media
Evaluating survey responses

Exam tip:
Sentiment analysis answers “How does the text feel?”

3. Key Phrase Extraction

What it does:
Identifies the main talking points in a block of text.

Example:

“The hotel had great service but poor Wi-Fi.”

Key phrases might include:

great service
poor Wi-Fi

Common exam scenario:
Summarizing long documents or feedback automatically.

4. Named Entity Recognition (NER)

What it does:
Detects and categorizes entities mentioned in text.

Common entity types:

People
Organizations
Locations
Dates
Products

Example:

“Satya Nadella is the CEO of Microsoft.”

Entities detected:

Person: Satya Nadella
Organization: Microsoft

5. Language Detection

What it does:
Identifies the language a piece of text is written in.

Why it matters:

Enables multilingual applications
Often used before translation or sentiment analysis

Exam tip:
Azure AI Language can detect language without being told what it is.

6. Question Answering

What it does:
Allows applications to answer natural language questions based on provided content.

Key points:

Replaces the older QnA Maker
Uses FAQs, documents, or URLs as knowledge sources
Commonly used in chatbots and helpdesk systems

Example:

User: “What is your return policy?”
Bot responds using stored knowledge.

7. Text Classification

What it does:
Assigns predefined categories or labels to text.

Examples:

Classifying emails as billing, technical support, or general inquiry
Tagging support tickets automatically

Important distinction:
This is about categorizing content, not detecting sentiment.

8. Custom Language Models

What it does:
Allows organizations to train custom NLP models using their own data.

Used for:

Domain-specific terminology
Industry-specific language (legal, healthcare, finance)

AI-900 focus:
Know that customization is possible, not how to train models.

Azure AI Language Service vs Other Azure AI Services

This distinction is frequently tested.

Service	Primary Purpose
Azure AI Language	Understand and analyze text
Azure Translator	Translate text between languages
Azure Speech	Speech-to-text and text-to-speech
Azure Vision	Analyze images and video

Exam shortcut:
If the scenario is about meaning, sentiment, or structure of text, the answer is usually Azure AI Language service.

Common Exam Scenarios to Watch For

You’ll often see questions like:

“Which Azure service should you use to analyze customer reviews?”
“Which service extracts people and locations from documents?”
“Which NLP service powers chatbots with question answering?”

If it involves text understanding, not translation or speech → Azure AI Language.

Key Takeaways for AI-900

Azure AI Language service is the primary NLP service for text analysis
It supports:
- Sentiment analysis
- Key phrase extraction
- Entity recognition
- Language detection
- Question answering
- Text classification
It is different from Translator and Speech
AI-900 focuses on capabilities and use cases, not APIs or code

Go to the Practice Exam Questions for this topic.

Go to the AI-900 Exam Prep Hub main page.

AI-900, Artificial Intelligence (AI), Microsoft Certification, Natural Language Processing (NLP) January 31, 2026

Practice Questions: Describe Capabilities of the Azure AI Speech Service (AI-900 Exam Prep)

Practice Exam Questions

Question 1

A company wants to automatically convert recorded customer support calls into written transcripts for analysis.
Which Azure service should they use?

A. Azure AI Language
B. Azure AI Vision
C. Azure AI Speech
D. Azure Translator

✅ Correct Answer: C

Explanation:
Azure AI Speech provides Speech to Text, which converts spoken audio into written text. Azure AI Language analyzes existing text but does not process audio.

Question 2

An application needs to read written instructions aloud to users using natural-sounding voices.
Which Azure AI Speech capability is required?

A. Speech to Text
B. Text to Speech
C. Speaker Recognition
D. Speech Translation

✅ Correct Answer: B

Explanation:
Text to Speech converts written text into spoken audio. This is commonly used for accessibility and voice assistants.

Question 3

A global company wants users to speak in Spanish and hear an English audio response in real time.
Which Azure AI Speech feature supports this scenario?

A. Text Analytics
B. Azure Translator
C. Speech Translation
D. Speaker Identification

✅ Correct Answer: C

Explanation:
Speech Translation enables real-time translation of spoken language and can output translated speech or text.

Question 4

Which scenario is best suited for Azure AI Speech instead of Azure AI Language?

A. Extracting key phrases from emails
B. Detecting sentiment in product reviews
C. Transcribing audio from meetings
D. Identifying entities in documents

✅ Correct Answer: C

Explanation:
Azure AI Speech handles audio-based workloads such as transcribing meetings. Azure AI Language is used for written text analysis.

Question 5

A banking app needs to verify a user’s identity based on their voice.
Which Azure AI Speech capability should be used?

A. Speech to Text
B. Speaker Recognition
C. Text to Speech
D. Language Detection

✅ Correct Answer: B

Explanation:
Speaker Recognition is used to verify or identify individuals based on voice characteristics.

Question 6

Which Azure AI Speech capability converts spoken language into written text in real time?

A. Speech Translation
B. Text to Speech
C. Speech to Text
D. Speaker Identification

✅ Correct Answer: C

Explanation:
Speech to Text converts audio input into text and supports real-time transcription.

Question 7

A developer wants to generate lifelike, human-sounding voices for a virtual assistant.
Which feature of Azure AI Speech makes this possible?

A. Optical character recognition
B. Neural voices
C. Language modeling
D. Sentiment analysis

✅ Correct Answer: B

Explanation:
Azure AI Speech uses neural voices to produce natural-sounding speech output.

Question 8

Which input type is primarily required when using the Azure AI Speech service?

A. Images
B. Video streams
C. Audio data
D. Structured tables

✅ Correct Answer: C

Explanation:
Azure AI Speech is designed to process audio input, such as spoken language or sound recordings.

Question 9

Which scenario would require combining multiple Azure AI Speech capabilities?

A. Detecting faces in images
B. Translating written documents
C. Speaking in one language and hearing a translated spoken response
D. Analyzing sentiment in customer feedback

✅ Correct Answer: C

Explanation:
This scenario combines Speech to Text, Translation, and Text to Speech to deliver a speech-to-speech experience.

Question 10

Which statement best describes Azure AI Speech?

A. It analyzes written documents for meaning
B. It processes images and videos
C. It enables spoken language understanding and generation
D. It is used only for chatbots

✅ Correct Answer: C

Explanation:
Azure AI Speech focuses on spoken language, including recognition, synthesis, translation, and speaker identification.

Final Exam Tips 🧠

If the question mentions audio, voice, or speech, think Azure AI Speech
Know the difference between:
- Speech to Text
- Text to Speech
- Speech Translation
- Speaker Recognition
AI-900 questions are conceptual and scenario-based, not technical

Go to the AI-900 Exam Prep Hub main page.

AI, AI-900, Artificial Intelligence (AI), Microsoft Certification, Natural Language Processing (NLP) January 31, 2026

Describe Capabilities of the Azure AI Speech Service (AI-900 Exam Prep)

Where This Fits in the Exam

Exam: AI-900 – Microsoft Azure AI Fundamentals
Domain: Describe features of Natural Language Processing (NLP) workloads on Azure (15–20%)
Sub-area: Identify Azure tools and services for NLP workloads

For AI-900, Microsoft expects you to understand what the Azure AI Speech service does, when to use it, and how it differs from other AI services — not how to code it.

What Is the Azure AI Speech Service?

The Azure AI Speech service is a cloud-based service that enables applications to process spoken language. It allows systems to:

Convert speech into text
Convert text into natural-sounding speech
Translate spoken language
Recognize speakers and voices

It is part of Azure AI Services and focuses on audio and voice-based NLP workloads.

Core Capabilities of Azure AI Speech

1. Speech to Text

Speech to Text converts spoken audio into written text.

Key features:

Real-time transcription
Batch transcription of audio files
Support for multiple languages
Automatic punctuation and formatting

Common use cases:

Transcribing meetings or calls
Voice-controlled applications
Call center analytics
Accessibility tools (captions and subtitles)

📌 AI-900 exam tip:
If the question mentions converting spoken words into text, the answer is Azure AI Speech (Speech to Text).

2. Text to Speech

Text to Speech converts written text into natural-sounding spoken audio.

Key features:

Neural voices that sound human-like
Multiple languages and accents
Adjustable pitch, speed, and tone
Support for voice styles (e.g., cheerful, calm)

Common use cases:

Voice assistants
Read-aloud applications
Accessibility for visually impaired users
Automated announcements

📌 AI-900 exam tip:
If the scenario describes reading text out loud, think Text to Speech.

3. Speech Translation

Speech Translation converts spoken language into another language, either as text or synthesized speech.

Key features:

Real-time speech translation
Multi-language support
Can output translated speech or text

Common use cases:

Multilingual meetings
Travel and tourism apps
International customer support

📌 AI-900 exam tip:
Speech translation handles spoken language, while Azure Translator handles written text.

4. Speaker Recognition

Speaker Recognition identifies or verifies who is speaking based on their voice.

Capabilities include:

Speaker verification (confirming identity)
Speaker identification (determining who is speaking)

Common use cases:

Secure voice authentication
Call center speaker tracking
Personalized voice experiences

📌 AI-900 note:
You only need to understand what it does, not how voice models are trained.

5. Speech-to-Speech Scenarios

By combining Speech to Text, Translation, and Text to Speech, Azure AI Speech supports end-to-end voice experiences, such as:

Speaking in one language and hearing a response in another
Voice-based chatbots
Smart devices and assistants

How Azure AI Speech Differs from Other Azure AI Services

Service	Primary Purpose
Azure AI Speech	Spoken language (audio)
Azure AI Language	Written text analysis
Azure Translator	Text translation
Azure AI Vision	Images and video

📌 Exam pattern to watch for:
Microsoft often tests whether you can choose the right service based on the input type (audio vs text vs image).

Typical AI-900 Scenarios Involving Azure AI Speech

You should choose Azure AI Speech when a scenario involves:

Audio recordings
Live speech
Voice input or output
Real-time transcription
Spoken translation

Key Takeaways for the AI-900 Exam

Azure AI Speech focuses on spoken language, not written text
Core capabilities:
- Speech to Text
- Text to Speech
- Speech Translation
- Speaker Recognition
Exam questions are scenario-based, not technical
If the question mentions audio, voice, or speech, Azure AI Speech is usually the answer

Go to the Practice Exam Questions for this topic.

Go to the AI-900 Exam Prep Hub main page.

AI, AI-900, Artificial Intelligence (AI), Generative AI, Microsoft Certification January 31, 2026

Practice Questions: Identify Features of Generative AI Models (AI-900 Exam Prep)

Practice Questions

Question 1

Which scenario is the best example of a generative AI workload?

A. Predicting tomorrow’s temperature based on historical data
B. Classifying emails as spam or not spam
C. Generating a product description from a short prompt
D. Detecting anomalies in server performance metrics

✅ Correct Answer: C

Explanation:
Generative AI models are designed to create new content, such as text, images, or code. Generating a product description is a content creation task, which is a core feature of generative AI.

Question 2

What is a key characteristic that distinguishes generative AI models from traditional machine learning models?

A. They require labeled training data
B. They produce deterministic outputs
C. They generate new data similar to training data
D. They can only be used for classification tasks

✅ Correct Answer: C

Explanation:
Generative AI models learn patterns from data and generate new outputs that resemble the data they were trained on, rather than only predicting labels or numeric values.

Question 3

What role does a prompt play when working with a generative AI model?

A. It retrains the model with new data
B. It defines how the model should generate a response
C. It validates the accuracy of the model
D. It encrypts the generated output

✅ Correct Answer: B

Explanation:
A prompt provides instructions or context that guide the model’s output. It does not retrain the model or affect its underlying parameters.

Question 4

Why can the same prompt sometimes produce different responses from a generative AI model?

A. The model uses rule-based logic
B. The model is deterministic
C. The model generates probabilistic outputs
D. The training data changes after each request

✅ Correct Answer: C

Explanation:
Generative AI models use probabilistic methods, meaning they select likely next outputs rather than fixed responses, which can result in variation.

Question 5

Which feature enables a generative AI model to produce human-like text responses?

A. Feature engineering
B. Context awareness and large-scale pretraining
C. Manual rule definition
D. Binary classification

✅ Correct Answer: B

Explanation:
Generative AI models are pretrained on massive datasets and use context to generate fluent, coherent, human-like responses.

Question 6

Which statement best describes the training approach used by most generative AI models?

A. They are trained only on small, task-specific datasets
B. They are pretrained on large datasets and adapted for many tasks
C. They require real-time retraining for each request
D. They are trained exclusively using reinforcement learning

✅ Correct Answer: B

Explanation:
Generative AI models are typically large pretrained models that can perform multiple tasks without retraining.

Question 7

Which scenario would most likely require the use of a generative AI model?

A. Predicting customer churn
B. Assigning product categories
C. Writing a summary of a long document
D. Detecting fraudulent transactions

✅ Correct Answer: C

Explanation:
Summarization involves creating new text, which is a hallmark of generative AI workloads.

Question 8

What is a common risk associated with generative AI models that requires responsible AI controls?

A. Overfitting to training data
B. Hallucinations and biased outputs
C. Low model accuracy
D. Inability to scale

✅ Correct Answer: B

Explanation:
Generative AI models can produce confident but incorrect information or biased content, making responsible AI safeguards essential.

Question 9

Which feature allows a generative AI model to continue a conversation in a meaningful way?

A. Feature scaling
B. Context retention
C. Label encoding
D. Data normalization

✅ Correct Answer: B

Explanation:
Context retention enables generative AI models to understand previous inputs and generate coherent multi-turn conversations.

Question 10

Which statement best describes the scope of tasks generative AI models can perform?

A. They are limited to a single predefined task
B. They can perform multiple tasks using the same model
C. They must be retrained for each task
D. They only work with numerical data

✅ Correct Answer: B

Explanation:
Generative AI models are general-purpose, capable of handling a wide variety of tasks such as summarization, translation, content generation, and question answering.

Final Exam Tip 💡

If an AI-900 question mentions:

Creating text, images, or code
Prompts
Conversations
Human-like responses

👉 Think: Generative AI model

Go to the AI-900 Exam Prep Hub main page.

AI, AI-900, Artificial Intelligence (AI), Generative AI, Microsoft Certification January 31, 2026

Identify Features of Generative AI Models (AI-900 Exam Prep)

Introduction

Generative AI models are a class of artificial intelligence systems designed to create new content rather than simply analyze or classify existing data. In the AI-900 exam, Microsoft focuses on conceptual understanding, not implementation details. You are expected to recognize what generative AI models do, how they behave, and what makes them different from traditional machine learning models.

Generative AI underpins many modern Azure AI solutions, including Azure OpenAI Service, and plays a central role in text, image, code, and audio generation workloads.

What Is a Generative AI Model?

A generative AI model learns patterns, structure, and relationships from large datasets and uses that knowledge to generate new, original outputs that resemble the data it was trained on.

Unlike predictive models (which output labels or numeric values), generative models produce:

Text
Images
Code
Audio
Synthetic data

Key Features of Generative AI Models (Exam Focus)

1. Content Generation

Generative AI models can create new content rather than selecting from predefined responses.

Examples:

Writing emails, stories, or summaries
Generating images from text descriptions
Producing computer code
Creating conversational responses

AI-900 cue: If the scenario involves creating something new, it likely involves generative AI.

2. Large Pretrained Models

Generative AI models are typically pretrained on massive datasets containing text, images, or other media.

Key characteristics:

Trained on diverse, large-scale data
Capture language structure, context, and semantics
Can generalize to many tasks without retraining

Examples:

Large language models (LLMs)
Multimodal foundation models

3. Prompt-Based Interaction

Generative AI models are commonly controlled using prompts, which are natural language instructions or inputs.

Prompts can:

Ask questions
Provide instructions
Set constraints or styles
Include examples (few-shot prompting)

Exam tip: Prompts guide how the model responds but do not retrain the model.

4. Probabilistic Output (Non-Deterministic)

Generative AI models produce probabilistic responses, meaning:

The same prompt can produce different outputs
Responses are not fixed or guaranteed
Outputs are generated based on likelihood, not rules

This enables creativity but also requires careful validation.

5. Context Awareness

Generative AI models can use context provided in a conversation or prompt to influence responses.

Examples:

Remembering earlier parts of a conversation
Adjusting tone or topic based on prior input
Generating coherent multi-turn dialogue

This is especially relevant for chat-based AI systems.

6. General-Purpose Capability

Generative AI models are often multi-task by design.

A single model can:

Answer questions
Summarize text
Translate languages
Generate explanations
Write code

This contrasts with traditional ML models, which are typically task-specific.

7. Fine-Tuning and Customization

While generative AI models are pretrained, they can be:

Fine-tuned with domain-specific data
Prompt-engineered for specific use cases
Configured with system instructions

For AI-900, it’s important to know customization is possible, not how to implement it.

8. Human-Like Outputs

Generative AI models are designed to produce outputs that appear:

Natural
Fluent
Contextually relevant
Similar to human-generated content

This is especially true for text and conversational AI.

9. Support for Multimodal Data

Some generative AI models can work across multiple data types, such as:

Text → Image
Image → Text
Text → Code

AI-900 expects recognition of this capability, not technical depth.

10. Need for Responsible AI Controls

Generative AI models require safeguards due to risks such as:

Hallucinations (incorrect but confident outputs)
Bias
Harmful or inappropriate content

Microsoft emphasizes:

Content filtering
Responsible AI principles
Human oversight

Generative AI vs Traditional Machine Learning (High-Yield Comparison)

Aspect	Traditional ML	Generative AI
Primary goal	Predict or classify	Create new content
Output type	Labels or numbers	Text, images, code, audio
Task scope	Narrow, specific	Broad, general-purpose
Interaction style	Structured inputs	Natural language prompts
Creativity	None	High

Azure Context (What AI-900 Expects You to Recognize)

Generative AI workloads on Azure are commonly delivered through:

Azure OpenAI Service
Integrated Azure AI tooling
Secure, enterprise-ready AI deployments

You are not expected to know APIs or pricing — only capabilities and use cases.

Common Exam Triggers to Watch For 👀

If a question mentions:

Writing text
Creating images
Generating code
Conversational responses
Prompt-based interaction

➡️ Think: Generative AI model

Summary

For the AI-900 exam, generative AI models are defined by their ability to:

Generate new content
Respond to prompts
Operate probabilistically
Handle multiple tasks
Produce human-like outputs
Require responsible AI safeguards

Understanding these features, not implementation details, is the key to scoring well in this exam section.

Go to the Practice Exam Questions for this topic.

Go to the AI-900 Exam Prep Hub main page.

AI, AI-900, Artificial Intelligence (AI), Generative AI, Machine Learning (ML), Microsoft Certification January 31, 2026

Workload Scenarios to Correct AI Approach mappings (AI-900 Exam Prep)

Here is some additional information to help you solidify your knowledge and prepare for the AI-900 exam.

1. Core AI Approaches

AI Approach	What It’s Best At
Traditional (Rule-Based) AI	Fixed logic, deterministic decisions
Predictive Machine Learning	Predicting values or classifying outcomes
Generative AI	Creating new content from prompts

Another way to relay the same information:

If it follows rules, it’s traditional AI.
If it predicts, it’s ML.
If it creates, it’s generative AI

2. Scenario-to-Approach Mapping

Business Rules & Automation

Scenario	Correct AI Approach	Why
Approve a loan if income > threshold	Traditional AI	Rule-based, no learning required
Route support tickets based on keywords	Traditional AI	Deterministic logic
Enforce compliance policies	Traditional AI	Rules must be followed exactly

Predictive & Analytical Scenarios

Scenario	Correct AI Approach	Why
Predict customer churn	Predictive ML (Classification)	Binary outcome
Forecast product demand	Predictive ML (Regression)	Numeric prediction
Detect credit card fraud	Predictive ML (Classification)	Probability-based decision
Predict house prices	Predictive ML (Regression)	Continuous value
Segment customers	Predictive ML (Clustering)	Discover groups

Natural Language Processing (NLP)

Scenario	Correct AI Approach	Why
Analyze customer sentiment	Predictive ML (NLP)	Classification of sentiment
Extract key phrases from text	Predictive ML (NLP)	Pattern recognition
Recognize named entities	Predictive ML (NLP)	Identify structured info
Translate text	Generative AI / NLP	Generates new text
Summarize documents	Generative AI	Content creation

Computer Vision

Scenario	Correct AI Approach	Why
Identify objects in an image	Predictive ML (Vision)	Classification/detection
Detect faces in images	Predictive ML (Vision)	Pattern recognition
Read printed text from images (OCR)	Predictive ML (Vision)	Extraction task
Generate images from text	Generative AI	Creates new images

Speech Workloads

Scenario	Correct AI Approach	Why
Convert speech to text	Predictive ML (Speech)	Recognition task
Convert text to speech	Generative AI	Generates audio
Identify spoken language	Predictive ML	Classification

Generative AI Scenarios

Scenario	Correct AI Approach	Why
Generate an email from a prompt	Generative AI	New content
Write code from a description	Generative AI	Content generation
Answer questions conversationally	Generative AI	Dynamic responses
Create images from text prompts	Generative AI	Creative output

3. Azure Service Mapping

Scenario Type	Azure Service
Predictive ML	Azure Machine Learning
NLP (Sentiment, Entities)	Azure AI Language
Speech workloads	Azure AI Speech
Vision workloads	Azure AI Vision
Generative AI	Azure OpenAI Service
Rule-based workflows	Logic Apps / Power Automate

4. Common AI-900 Exam Traps

Trap	Correct Thinking
“Translation is classification”	❌ Translation generates text
“Chatbots are always rule-based”	❌ Modern chatbots use generative AI
“OCR generates text”	❌ OCR extracts existing text
“Generative AI replaces ML”	❌ Different goals

Go to the AI-900 Exam Prep Hub main page.