Free AI-102 Practice Test Questions and Answers (2026)

View Mode
Q: 1

HOTSPOT You build a chatbot by using Azure OpenAI Studio. You need to ensure that the responses are more deterministic and less creative. Which two parameters should you configure? To answer, select the appropriate parameters in the answer area. NOTE: Each correct answer is worth one point. Microsoft AI-102 question

Your Answer
35 comments in the community discussion
5
Temperature and topp. These two directly affect randomness and creativity, so lowering both makes responses a lot more deterministic. Saw a similar question in practice sets, so pretty sure that’s it, but open if anyone disagrees.
5
Temperature and topp Had something like this in a mock and it's definitely those two. If you set temperature and topp lower, you get more predictable, less creative chatbot responses. Other settings like frequency penalty are for repetition, not randomness or creativity. Let me know if anyone sees it different.
Q: 2
You have a product support manual. You need to build a product support chatbot based on the manual. The solution must minimize development effort and costs. What should you use?
Options
26 comments in the community discussion
2
B Had something like this in a mock and B was the right pick.
2
Makes sense to go with B here. It's basically built for quick Q&A from docs, barely any effort to set up compared to fine-tuning or building an OpenAI+Search stack. Not totally ruling out C if you wanted more control, but for minimum dev and cost, B.
Q: 3

DRAG DROP You have an Azure subscription that contains an Azure Al Search instance named AlSearch1. AlSearch1 contains an index that includes a vector. You need to perform the following actions: • Deploy a new agent by using the Azure Al Agent Service. • Connect the AlSearch1 index to the new agent. • Validate the integration of the index and the agent. Which four actions should you perform in sequence? To answer, move the appropriate actions from the list of actions to the answer area and arrange them in the correct order. Microsoft AI-102 question

Drag & Drop
36 comments in the community discussion
5
Create Azure AI Client and get connection ID, then configure the search tool, create agent and enable search, finally ask the agent questions. I think this is right since the code interpreter steps are a trap here. Integration always starts with the client connection setup. Feel free to disagree if you see it different
5
Create Azure AI Client and get connection ID → Configure Azure AI Search tool → Create agent and enable search tool → Ask the agent questions. Had something like this in a mock-connection setup always comes first, then config, then agent, finally validation. Makes sense logically but let me know if you see it different
Q: 4

DRAG DROP You need to develop an automated call handling system that can respond to callers in their own language. The system will support only French and English. Which Azure Cognitive Services service should you use to meet each requirement? To answer, drag the appropriate services to the correct requirements. Each service may be used once, more than once, or not at all. You may need to drag the split bat between panes or scroll to view content. NOTE: Each correct selection is worth one point. Microsoft AI-102 question

Drag & Drop
30 comments in the community discussion
7
Seen similar on practice tests. Detect the incoming language: Speech to Text, Respond: Text to Speech.
5
Detect language: Speech to Text, respond: Text to Speech. Don’t think Text Analytics fits since it works on text not audio. Translator is a common trap here but the question says reply in the caller’s own language, not translate. Seen similar wording in other practice sets and this seems right.
Q: 5

DRAG DROP You have an Azure subscription that contains an Azure Al Search resource named AS1. You implement a custom skill in AS1 that performs language and sentiment analysis of documents. You ate evaluating the use of AS1 as part of an enrichment pipeline. In which order will AS1 index the documents? To answer, move all indexing stages from the list of stages to the answer area and arrange them in the correct order. Microsoft AI-102 question

Drag & Drop
29 comments in the community discussion
8
document cracking, field mappings, skillset execution, output field mappings, push to index That's the right order for how Azure AI Search pipelines process documents. Cracking comes first (gets the text), then fields get mapped, then any skillsets run (like sentiment analysis), outputs are mapped, and finally it all
6
Had something like this in a mock, order should be document cracking, field mappings, skillset execution, output field mappings, push to index.
Q: 6

HOTSPOT You are building an app that will process incoming email and direct email messages to either French or English language support teams. Which Azure Cognitive Services API should you use? To answer, select the appropriate options in the answer area. NOTE: Each correct selection is worth one point. Microsoft AI-102 question

Your Answer
26 comments in the community discussion
6
Makes sense to use /text/analytics/v3.1/languages for this.
6
Yeah, I'd use the /text/analytics/v3.1/languages endpoint for this. It's all about language detection, not translation, so Text Analytics is the right API here. Translator is only needed if you actually want to convert between languages. Pretty sure that's what the exam expects but open to any other views!
Q: 7

DRAG DROP You are building an app that will scan confidential documents and use the Language service to analyze the contents. You provision an Azure Cognitive Services resource. You need to ensure that the app can make requests to the Language service endpoint. The solution must ensure that confidential documents remain on-premises. Which three actions should you perform in sequence? To answer, move the appropriate actions from the list of actions to the answer area and arrange them in the correct order. Microsoft AI-102 question

Drag & Drop
45 comments in the community discussion
6
Looks good to me! You set up on-prem K8s with internet, pull the image from Microsoft Container Registry, then run the container using API key and endpoint. That's how you keep docs on-prem but still use Azure for billing. Pretty sure that's what MS expects here, but open to other views.
6
Provision on-premises Kubernetes (with internet), pull from MCR, then run container with API key + endpoint.
Q: 8
You are building an app that will use the Azure Video Indexer service. You plan to train a language model to recognize industry-specific terms. You need to upload a file that contains the industry-specific terms. Which file format should you use?
Options
32 comments in the community discussion
1
Its C, saw this on a similar exam recently, only TXT files work for adding vocab in Video Indexer.
1
Why would you pick XML or XLS here? For Video Indexer, they just want a plain .txt file with each term per line. Unless the question mentions extra data like definitions, C fits best in most cases.
Q: 9
You need to measure the public perception of your brand on social media messages. Which Azure Cognitive Services service should you use?
Options
41 comments in the community discussion
4
A. seen similar in practice tests and official docs suggest Text Analytics for sentiment tasks.
1
B is what I went with since Content Moderator scans messages for violations, which kinda relates to perception if nasty stuff damages the brand. Not 100% sure though, maybe I'm mixing up moderation with sentiment.
Q: 10

HOTSPOT For each of the following statements, select Yes if the statement is tine. Otherwise, select No. NOTE: Each correct selection is worth one point. Microsoft AI-102 question

Your Answer
45 comments in the community discussion
6
Not quite, statement 2 trips people up-batch isn't immediate at all. It's Yes, No, Yes for this set.
5
Nah, statement 2 is the classic batch trap-batch isn't real-time. Pretty sure it's Yes for 1 and 3, No for 2.
Q: 11

DRAG DROP You have a web app that uses Azure AI search. When reviewing activity, you see greater than expected search query volumes. You suspect that the query key is compromised. You need to prevent unauthorized access to the search endpoint and ensure that users only have read only access to the documents collection. The solution must minimize app downtime. Which three action should you perform in sequence? To answer, more the appropriate actions from the list of actions to the answer area and arrange them in the correct order. Microsoft AI-102 question

Drag & Drop
4 comments in the community discussion
4
Yep, add new query key, update app to use it, delete the old one. No admin key needed here.
2
Add a new query key → change app to use it → then delete the compromised key. That's how you rotate in Azure Search with no downtime and keep user access to read only. Pretty sure this matches what Microsoft recommends.
Q: 12

DRAG DROP You are developing a webpage that will use the Video Indexer service to display videos of internal company meetings. You embed the Player widget and the Cognitive Insights widget into the page. You need to configure the widgets to meet the following requirements: Ensure that users can search for keywords. Display the names and faces of people in the video. Show captions in the video in English (United States). How should you complete the URL for each widget? To answer, drag the appropriate values to the correct targets. Each value may be used once, more than once, or not at all. You may need to drag the split bar between panes or scroll to view content. NOTE: Each correct selection is worth one point. Microsoft AI-102 question

Drag & Drop
5 comments in the community discussion
3
widgets: people, keywords controls: search showcaptions: true captions: en-US. Casey Q. used "people, search" in widgets but that won't let users search inside the video, only see detected people. For keyword searching, pretty sure you need "keywords" here. The trap is mixing up the controls and widgets settings.
3
widgets: people, keywordscontrols: searchshowcaptions: truecaptions: en-US
Q: 13

DRAG DROP You have an app that manages feedback. You need to ensure that the app can detect negative comments by using the Sentiment Analysis API in Azure Cognitive Service for Language. The solution must ensure that the managed feedback remains on your company's internal network. Which three actions should you perform in sequence? To answer, move the appropriate actions from the list of actions to the answer area and arrange them in the correct order. NOTE: More than one order of answer choices is correct You will receive credit for any of the correct orders you select. Microsoft AI-102 question

Drag & Drop
11 comments in the community discussion
6
These drag-and-drops are always wordy. Provision Language service, deploy container on-prem, then run and query it.
6
Yeah, this flow makes sense: provision Language service in Azure, then deploy Docker container on-prem, finally run and query it locally. Keeps everything inside your own network and satisfies the data residency piece. I think that's what Microsoft expects here but always possible there's a weird edge case.
Q: 14
Note: This question is part of a series of questions that present the same scenario. Each question in the series contains a unique solution that might meet the stated goals. Some question sets might have more than one correct solution, while others might not have a correct solution. After you answer a question in this section, you will NOT be able to return to it. As a result, these questions will not appear in the review screen. You build a language model by using a Language Understanding service. The language model is used to search for information on a contact list by using an intent named FindContact. A conversational expert provides you with the following list of phrases to use for training. Find contacts in London. Who do I know in Seattle? Search for contacts in Ukraine. You need to implement the phrase list in Language Understanding. Solution: You create a new pattern in the FindContact intent. Does this meet the goal?
Options
9 comments in the community discussion
2
Option B I remember a similar scenario from labs. Patterns aren't for adding example phrases, they're more for templated formats. These should go as utterances instead. Pretty sure about this one.
1
Doesn't look like A is correct, since patterns are mainly for fixed templates. B.
Q: 15
You plan to perform predictive maintenance. You collect IoT sensor data from 100 industrial machines for a year. Each machine has 50 different sensors that generate data at one-minute intervals. In total, you have 5,000 time series datasets. You need to identify unusual values in each time series to help predict machinery failures. Which Azure Cognitive Services service should you use?
Options
11 comments in the community discussion
2
A imo. I've seen similar questions in other sets and Anomaly Detector is the main Azure service for time series outlier detection, especially when you want to surface unusual sensor readings for predictive maintenance. B (Cognitive Search) is great for searching content but doesn't do anomaly ML itself. Anyone disagree
1
B tbh, Cognitive Search seems logical for searching sensor logs if the focus is quick lookups, not just automatic outlier detection. Trap for those thinking only pure anomaly detection matters here.
Q: 16

HOTSPOT You plan to deploy an Azure OpenAI resource by using an Azure Resource Manager (ARM) template. You need to ensure that the resource can respond to 600 requests per minute. How should you complete the template? To answer, select the appropriate options in the answer area. NOTE: Each correct selection is worth one point.

Your Answer
6 comments in the community discussion
5
Isn’t the capacity property in ARM templates what actually controls throughput for Azure OpenAI, not maxRequestsPerMinute? The question wants 600 RPM consistently, so setting capacity to 100 makes sense if you follow the token/request ratio. Am I missing a detail here, or is there a trap ab
1
Shouldn’t the first box be maxRequestsPerMinute? The ARM template docs mention that property for rate limiting, not capacity.
Q: 17
You have a Language Understanding resource named lu1. You build and deploy an Azure bot named bot1 that uses lu1. You need to ensure that bot1 adheres to the Microsoft responsible AI principle of inclusiveness. How should you extend bot1?
Options
3 comments in the community discussion
1
Option D Adds speech support so more users can interact, which is what inclusiveness is about.
I get the logic for B, but active learning mostly helps the bot adapt to new phrases, not necessarily make it accessible for more users. D fits better because adding Direct Line Speech gives people with different needs (like visual or motor impairments) another way to interact with the bot. Pretty sure D is right, but
Q: 18
You need to develop an extract solution for the receipt images. The solution must meet the document processing requirements and the technical requirements. You upload the receipt images to the From Recognizer API for analysis, and the API returns the following JSON. Microsoft AI-102 question Which expression should you use to trigger a manual review of the extracted information by a member of the Consultant-Bookkeeper group?
Options
10 comments in the community discussion
8
D. not totally sure but looks right since it's checking if confidence is under 0.7 for MerchantName. Can someone confirm?
3
D. that's correct for this scenario.
Q: 19

DRAG DROP You train an Azure Custom Vision object detection model to identify a company's products by using the Retail domain. You plan to deploy the model as part of a mobile app for Android phones. You need to prepare the model for deployment. Which three actions should you perform in sequence? To answer, move the appropriate actions from the list of actions to the answer area and arrange them in the correct order. Microsoft AI-102 question

Drag & Drop
9 comments in the community discussion
5
Change domain, retrain, export is the right order here.
5
Change domain, retrain, export is the order. Saw a similar flow in labs and the official guide matches this too.
Q: 20

HOTSPOT You have an Azure subscription that contains an Azure Al Content Safety resource named Resource1 and a storage account named storage1. You create a blob container named container1 and upload a sample set of image files to container1. You need to validate whether Resource1 can identify images that contain potential violence. How should you complete the cURL command? To answer, select the appropriate options in the answer area. NOTE: Each correct selection is worth one point. Microsoft AI-102 question

Your Answer
8 comments in the community discussion
8
contentsafety and /image:analyze. Legacy "contentmoderator" would miss the violence detection part here.
2
contentmoderator and /ProcessImage. It trips me up since both seem related to scanning, but pretty sure that's a common tripwire.
Question 1 of 20

What's covered in this practice questions set

4: Implement natural language processing solutions · 9 questions

📖 About this Domain

This domain covers implementing solutions that process text and speech using Azure AI services. You will use Azure AI Language to extract insights from text and Azure AI Speech to integrate voice capabilities. Core tasks involve provisioning resources, performing text analysis, building conversational language understanding models, and implementing speech synthesis and recognition.

🎓 What You Will Learn

  • You will learn to use the Azure AI Language service for key phrase extraction, named entity recognition (NER), and sentiment analysis.
  • You will learn to build, train, and deploy custom text classification and custom named entity recognition models for specific domains.
  • You will learn to create a conversational language understanding model to map user utterances to intents and extract entities.
  • You will learn to integrate the Azure AI Speech service for speech-to-text transcription and text-to-speech synthesis using the SDK.

🛠️ Skills You Will Build

  • You will build the skill to provision Azure AI Language resources and use the client library to perform text analytics.
  • You will build proficiency in designing a conversational language understanding schema, including defining intents, entities, and training utterances.
  • You will build the ability to train, evaluate, and deploy custom NLP models for tasks like custom text classification.
  • You will build the capability to implement real-time speech recognition and synthesis in an application using the Azure AI Speech SDK.

💡 Top Tips to Prepare

  • Gain hands-on experience with the Python or C# SDKs for both Azure AI Language and Azure AI Speech services.
  • Master the core components of conversational language understanding, specifically the distinction between intents, entities, and utterances.
  • Understand the process for provisioning Azure AI services resources, managing keys, and identifying the correct endpoint URI.
  • Know when to use a pre-built model versus when to build a custom model for tasks like text classification or NER.

1: Plan and manage an Azure AI solution · 5 questions

📖 About this Domain

This domain covers the foundational process of selecting the correct Azure AI services based on business requirements. It emphasizes planning for security, monitoring, and the implementation of Responsible AI principles within a solution.

🎓 What You Will Learn

  • You will learn to select the appropriate Azure AI service, including Azure Cognitive Services, Azure Applied AI Services, and Azure Machine Learning.
  • You will learn to plan and configure security for an Azure AI solution by managing authentication, access control with RBAC, and secrets with Azure Key Vault.
  • You will learn to create and manage Azure AI services, including resource provisioning, endpoint management, and configuring diagnostic logging.
  • You will learn to plan and implement a Responsible AI solution by applying Microsoft's principles and using tools like the Responsible AI dashboard and Azure AI Content Safety.

🛠️ Skills You Will Build

  • You will build the skill to architect AI solutions by mapping business requirements to specific Azure AI services and components.
  • You will build skills in securing AI workloads by implementing authentication, role-based access control (RBAC), and managed identities.
  • You will build the skill to provision and manage the lifecycle of Azure AI resources using the Azure portal, CLI, and ARM templates.
  • You will build the skill to integrate Responsible AI practices, including fairness, transparency, and content safety, into your solution design.

💡 Top Tips to Prepare

  • Focus on the specific use cases for each service in the Azure AI portfolio to master service selection questions.
  • Gain hands-on experience with Azure Active Directory, creating service principals, assigning RBAC roles, and managing keys in Azure Key Vault.
  • Practice provisioning Azure AI resources, paying close attention to available pricing tiers, network configurations like private endpoints, and monitoring options.
  • Memorize the six Microsoft Responsible AI principles and understand how tools like Azure Machine Learning's Responsible AI dashboard address them.

3: Implement Azure AI Vision solutions · 4 questions

📖 About this Domain

This domain covers the implementation of computer vision solutions using Azure AI Vision services. You will analyze images and videos to extract information, identify objects, and read text. The focus is on integrating pre-built models and creating custom vision solutions.

🎓 What You Will Learn

  • You will learn to use the Image Analysis service to extract a wide range of visual features from images.
  • You will learn to implement the Read API for Optical Character Recognition (OCR) to extract text from images and documents.
  • You will learn to build, train, and deploy custom image classification and object detection models using the Custom Vision service.
  • You will learn to use the Face service to detect, analyze, and identify human faces in images.

🛠️ Skills You Will Build

  • You will build skills to provision and manage Azure AI Vision resources through the Azure portal and CLI.
  • You will build skills to integrate the Azure AI Vision SDK into applications to perform image analysis and OCR.
  • You will build skills to train, evaluate, and publish custom models using the Custom Vision portal and SDK.
  • You will build skills to process video streams for near-real-time analysis using Azure Video Indexer.

💡 Top Tips to Prepare

  • Gain hands-on experience with the Azure AI Vision SDKs in both Python and C# to understand API request and response structures.
  • Clearly distinguish between the use cases for the Image Analysis service versus a custom-trained Custom Vision model.
  • Master the asynchronous operation pattern of the Read API, including how to submit a job and retrieve the results.
  • Practice the end-to-end workflow for Custom Vision projects, from image tagging and training to model evaluation and endpoint consumption.

5: Implement knowledge mining and document intelligence solutions · 1 questions

📖 About this Domain

This domain covers implementing knowledge mining solutions using Azure AI Search to index and query unstructured data. You will also learn to implement Azure AI Document Intelligence to extract data from forms and documents, automating data entry.

🎓 What You Will Learn

  • You will learn to provision an Azure AI Search service, create a data source, and define a search index.
  • You will learn to create and attach a cognitive skillset to an indexer to enrich data during ingestion.
  • You will learn to use Azure AI Document Intelligence prebuilt and custom models to extract key-value pairs and tables.
  • You will learn to create a client application that queries a search index to find information.

🛠️ Skills You Will Build

  • You will build the skill to configure an Azure AI Search indexer, including field mappings and scheduling.
  • You will build the skill to implement a custom skillset to extend the data enrichment pipeline.
  • You will build the skill to train a custom Azure AI Document Intelligence model using labeled data in Document Intelligence Studio.
  • You will build the skill to write complex search queries using Lucene syntax and filtering expressions.

💡 Top Tips to Prepare

  • Focus on the components of the Azure AI Search indexing pipeline: data source, skillset, index, and indexer.
  • Practice creating and testing cognitive skillsets, as they are central to knowledge mining.
  • Understand the differences between prebuilt, custom template, and custom neural models in Azure AI Document Intelligence.
  • Complete the hands-on labs in the official Microsoft Learn modules for AI-102 to gain practical experience with both services.

2: Implement decision support solutions · 1 questions

📖 About this Domain

This domain covers implementing AI solutions that extract insights from unstructured data to aid organizational decision-making. It focuses on Azure AI services for knowledge mining, content moderation, and intelligent document processing. You will build systems that analyze text, images, and documents to provide actionable intelligence.

🎓 What You Will Learn

  • You will learn to implement Azure AI Content Safety to detect and filter inappropriate user-generated content in text and images.
  • You will learn to create a knowledge mining solution using Azure AI Search by building an index, ingesting data, and applying a cognitive skillset.
  • You will learn to use Azure AI Document Intelligence to extract text, key-value pairs, and tables from documents with prebuilt and custom models.
  • You will learn to integrate various decision support services to create a comprehensive AI solution.

🛠️ Skills You Will Build

  • You will build the skill to configure and manage an Azure AI Search service, including defining indexers, data sources, and skillsets.
  • You will build the ability to train, deploy, and use custom models in Azure AI Document Intelligence for specialized document extraction.
  • You will build proficiency in using the Azure AI Content Safety service to moderate content based on specific harm categories and severity levels.
  • You will build the technical skill to consume these services programmatically using their respective REST APIs and SDKs.

💡 Top Tips to Prepare

  • Gain hands-on experience with the Azure AI Document Intelligence Studio to label and train both custom template and custom neural models.
  • Focus on the components of an Azure AI Search skillset, understanding how to chain built-in skills to enrich an index during ingestion.
  • Understand the specific harm categories and severity levels within Azure AI Content Safety for both text and image analysis.
  • Practice creating and querying Azure AI Search indexes using both the Azure portal and the REST API to understand the full workflow.

Premium Access Includes

  • Quiz Simulator
  • Exam Mode
  • Progress Tracking
  • Question Saving
  • Flash Cards
  • Drag & Drops
  • 3 Months Access
  • PDF Downloads
Get Premium Access
Scroll to Top

FLASH OFFER

Days
Hours
Minutes
Seconds

avail 10% DISCOUNT on YOUR PURCHASE