Free Microsoft Azure AI-102 Actual Exam Questions
Dumps Box (DumpsBox) offers up-to-date practice exam questions for AI-102 certification exam which are developed and validated by Microsoft subject domain experts certified in Microsoft Azure AI-102 . These practice questions are update regularly as we keep an eye on any recent changes in AI-102 syllabus, and when there is update our team quickly adjusts the questions. This commitment to providing the best quality exam prep material to certification aspirants is what makes DumpsBox.com the best certification exam prep website. On top of that, our strong, yet strictly moderated, community based feedback keeps the content clean and current. Each question has helpful community discussion that provides it extra perspective and introduces helpful resources for better exam preparation. This also saves students from other outdated practice questions or illicit exam dumps that can have adverse affects on career. Browse through our Microsoft Azure AI-102 exam questions and pass your exam on first try.
HOTSPOT You create a knowledge store for Azure Cognitive Search by using the following JSON. 
Use the drop-down menus to select the answer choice that completes each statement based on the information presented in the graphic. NOTE Each correct selection is worth one point. 
I think A is more like metadata since it includes a title and URL, which help identify the content, while B looks like tags that categorize it further. C definitely contains the main searchable text.
B feels like metadata because it’s concise and categorical. A seems more descriptive but less like core content, so it might be additional metadata or summaries, not the main content which is clearly C.
You need to build a product support chatbot based on the manual. The solution must minimize
development effort and costs.
What should you use?
What about B? It’s designed for Q&A with minimal setup, so less dev work.
I’m thinking option C could add unnecessary complexity here. Grounding with Azure AI Search sounds powerful but setting up and maintaining that index might increase cost and effort. Since the goal is minimal development, B feels more straightforward—it’s designed specifically for custom Q&A without needing extra fine-tuning or data prep. Also, if the manual is mostly text, the custom question answering service should cover it well. Could anyone share if there are performance trade-offs between B and D when dealing with mainly unstructured text?
HOTSPOT You are building an app that will answer customer calls about the status of an order. The app will query a database for the order details and provide the customers with a spoken response. You need to identify which Azure Al service APIs to use. The solution must minimize development effort. Which object should you use for each requirement? To answer, select the appropriate options in the answer area. NOTE: Each correct selection is worth one point. 
I’d go with Speech to Text for converting the customer’s spoken query into text, since that’s the natural first step before checking order status. Then, once you pull up the order info, use Text to Speech to give the response back as audio. The Language Understanding service seems unnecessary here because you’re not really interpreting complex intents, just matching a direct query to data. So B for converting speech input and D for generating the spoken output makes the most sense to me. It keeps things straightforward without extra processing steps.
B for speech-to-text to get live transcription, and LUIS for understanding intent.
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. 
First, regenerate the compromised query key to stop unauthorized use. Then, update the app to use the new read-only query key to limit access. Lastly, enable Azure AD authentication for stronger, managed security without downtime.
I’d start by regenerating the compromised query key to immediately block unauthorized use (that cuts off the leak fast). Next, update the app to use the new read-only query key so users only get the permissions they need. Finally, enable Azure AD authentication to move away from query keys and improve security long-term without downtime. This order keeps the app running smoothly while tightening up access control step by step.
HOTSPOT You have an app that uses the AI Language custom question answering service. You need to ad alternatives for the word testing by using the Authoring API. How should you complete the JSON payload? To answer, select the appropriate options in the answer area. NOTE: Each correct selection is worth one point. 
D works because it directly lists alternatives as an array, no extra nesting.
D fits best since it uses a straightforward array of alternatives under "testing."
HOTSPOT You have a library that contains thousands of images. You need to tag the images as photographs, drawings, or clipart. Which service endpoint and response property should you use? To answer, select the appropriate options in the answer area. NOTE: Each correct selection is worth one point. 
I think the key here is picking the service endpoint that actually supports detailed image classification. The Analyze endpoint works because it’s designed to give categories like photograph, drawing, or clipart. The “categories” property makes the most sense to check since it groups images in those types better than tags, which can be vague or just describe objects. So, I’d go with Analyze and categories, especially since tags might not clearly differentiate clipart and drawings on their own.
The Analyze endpoint is definitely the way to go here, but I think the response property to check is "categories" since it groups images into those broader types better than just "tags."
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.
Which expression should you use to trigger a manual review of the extracted information by a
member of the Consultant-Bookkeeper group?
Maybe B is better since it checks confidence across all fields, not just MerchantName.
B imo, it covers any field under 0.7, safer for catching issues than just MerchantName.
In RG1, you create an App Service plan named AP1.
Which two Azure resources are automatically created in RG1 when you provision the QnA Maker
service? Each correct answer presents part of the solution.
NOTE: Each correct selection is worth one point.
B/D? The QnA Maker definitely needs Azure SQL Database to store the knowledge base data, and Azure Cognitive Search is used for indexing and querying. I don’t think Language Understanding or App Service is automatically created because you already have an App Service plan there. Storage might be involved but not guaranteed as auto-created here. So B and D seem to fit best for the core backend components that get set up automatically.
B/D? The database is necessary to store the QnA data, and Cognitive Search handles querying. Storage (C) and App Service (E) don’t seem to be auto-provisioned here since you already set up the service plan.
HOTSPOT You have an Azure subscription that contains an Azure Al Content Safety resource named CS1. You need to use the SDK to call CS1 to identify requests that contain harmful content. How should you complete the code? lo answer, select the appropriate options m the answer area. NOTE: Each correct selection is worth one point 
I’m thinking A for the credential part because typically, Azure SDKs use an AzureKeyCredential object to handle the key securely, not just passing the key directly in client options. So even if B for the endpoint looks right, C might be the better call to actually provide the key properly. It matches what I’ve seen in other Azure AI SDKs where credentials are separated out from general client options.
I think option B for the endpoint is a safe bet since you need to specify the resource’s URL before making calls. For the key, D looks right because the SDK often requires the key as part of the client’s credentials or options. Options A and C don’t seem to fit typical SDK patterns where you set endpoint separately from credentials. Plus, the code snippet style points toward using B and D to correctly configure the client before calling the content safety method.
You deploy the GPT-4 model to the resource.
You need to ensure that you can upload files that will be used as grounding data for the model.
Which two types of resources should you create? Each correct answer presents part of the solution.
NOTE: Each correct selection is worth one point.
D and E, since Blob Storage stores files and AI Search indexes them for GPT-4 to access.
It’s D and C. Blob Storage holds raw files, but Document Intelligence is needed to extract and structure info so GPT-4 can actually use the data effectively. AI Search alone won’t prep the content properly.
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. 
B for search keywords in Insights, Player handles faces and en-US captions.
I think for searching keywords, the Cognitive Insights widget definitely needs the parameter that enables search, so B fits there. For displaying names and faces, that’s handled by the Player widget, so it should have the right option for people insights. Regarding captions, since the question specifically says English (United States), it makes sense to use "en-US" for both widgets to ensure consistency, even if sometimes "en" might work. So I’d put “en-US” where captions are required on both widgets to be safe.
HOTSPOT You have an Azure subscription that has the following configurations: • Subscription ID: 8d3591aa-96b8-4737-ad09-OOf9b1ed35ad • Tenant ID: 3edfe572-cbS4-3ced-ae12-c5c177f39a12 You plan to create a resource that will perform sentiment analysis and optical character recognition (OCR). You need to use an HTTP request to create the resource in the subscription. The solution must use a single key and endpoint. How should you complete the request? To answer, select the appropriate options in the answer area. NOTE: Each correct selection is worth one point. 
I’m pretty sure the key part is that they want a single resource for both OCR and sentiment analysis. That’s what the new “Cognitive Services” resource is for, so the request should create that one. The URL needs to point to the subscription and specify the resource type as “CognitiveServices” and not the old individual services. Also, the endpoint should be the one assigned to that Cognitive Services resource, not the regional or service-specific endpoints, since that’s how you get unified access with one key. This aligns well with options mentioning the new resource type and unified endpoint
I’m thinking the request should hit the /subscriptions path to specify which subscription we’re working with. The unified Cognitive Services resource lets you call both OCR and sentiment analysis through one key, so it won’t need separate endpoints for each. Also, the HTTP method needs to be PUT since you’re creating the resource. If you use POST, it might be for actions or queries but not resource creation. So, a PUT to the subscription-level resource path seems right here.
You need to deploy an Azure resource to the East US Azure region. The resource will be used to
perform sentiment analysis.
How should you call the method?
I think the service name is the main clue here—sentiment analysis belongs to TextAnalytics, not ContentModerator. Does the method specify if the region parameter needs to be lowercase or exactly "East US"?
B/D? I agree TextAnalytics is the right service for sentiment analysis, so A and C are out. Between B and D, the SKU "S0" is the typical tier for TextAnalytics, so that fits better than "Standard." Also, the region name usually doesn’t have spaces and is lowercase, so "eastus" is correct. So B seems more accurate than D based on naming conventions alone.
HOTSPOT You are building a chatbot. You need to use the Content Moderator API to identify aggressive and sexually explicit language. Which three settings should you configure? To answer, select the appropriate settings in the answer area. NOTE: Each correct selection is worth one point. 
I’d go with profanity, adult content, and offensive language like others said. Profanity seems to cover aggressive language, while adult content clearly targets sexual explicitness. Offensive language probably catches anything that’s borderline or doesn’t fit neatly into the other two. Also, the “hate” or “violence” settings seem more specific to threats or hate speech rather than general aggression, so they don’t feel like the best fit here. This way, you cover both the aggressive and explicit aspects without overlapping too much or missing important flags.
Profanity, adult content, and offensive language seem like the right picks here.
HOTSPOT You are building a chatbot by using the Microsoft Bot Framework Composer. You have the dialog design shown in the following exhibit. 
For each of the following statements, select Yes if the statement is true. Otherwise, select No. NOTE: Each correct selection is worth one point. 
I’m going with No on statements that say End dialog clears the entire stack because it just ends the current dialog and returns control to the parent. Cancel all dialogs definitely clears everything, so any statement implying partial clearing has to be No. Also, since the question doesn’t specify Composer version, I’m assuming standard behavior where Cancel all dialogs resets everything, including interruptions. That should help weed out which answers to confirm as Yes or No.
Yes for statements saying Cancel all dialogs clears everything, No for those about End dialog ending all.