You have already completed the Test before. Hence you can not start it again.
Test is loading...
You must sign in or sign up to start the Test.
You have to finish following quiz, to start this Test:
Your results are here!! for" AB-410 Practice Test 4 "
0 of 60 questions answered correctly
Your time:
Time has elapsed
Your Final Score is : 0
You have attempted : 0
Number of Correct Questions : 0 and scored 0
Number of Incorrect Questions : 0 and Negative marks 0
Average score
Your score
AB-410
You have attempted: 0
Number of Correct Questions: 0 and scored 0
Number of Incorrect Questions: 0 and Negative marks 0
You can review your answers by clicking on “View Answers” option. Important Note : Open Reference Documentation Links in New Tab (Right Click and Open in New Tab).
Answer Review
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Answer
Review
Unattempted
Correct
Incorrect
Unattempted
Every question in this attempt was answered correctly.
Question 1 of 60
1. Question
Scenario: A developer needs to summarize a 50-page technical document. Which LLM parameter should be increased to ensure the summary is not cut off mid-sentence?
Correct
Max Tokens determines the length of the generated output; for long summaries, this limit must be set high enough.
Incorrect
Max Tokens determines the length of the generated output; for long summaries, this limit must be set high enough.
Unattempted
Max Tokens determines the length of the generated output; for long summaries, this limit must be set high enough.
Question 2 of 60
2. Question
In a RAG architecture, what is the role of the “Orchestrator“ (e.g., Semantic Kernel)?
Correct
The Orchestrator is the “glue“ that manages the sequence of operations required to generate a grounded response.
Incorrect
The Orchestrator is the “glue“ that manages the sequence of operations required to generate a grounded response.
Unattempted
The Orchestrator is the “glue“ that manages the sequence of operations required to generate a grounded response.
Question 3 of 60
3. Question
In the context of Azure OpenAI, what is the purpose of “Tokenization“?
Correct
Tokenization is the process of converting raw text into numerical tokens, which are the basic units of data for transformer-based models.
Incorrect
Tokenization is the process of converting raw text into numerical tokens, which are the basic units of data for transformer-based models.
Unattempted
Tokenization is the process of converting raw text into numerical tokens, which are the basic units of data for transformer-based models.
Question 4 of 60
4. Question
In the context of LLMs, what does the term “Grounding“ refer to?
Correct
Grounding ensures that the AIÂ’s responses are “anchored“ to your specific business data rather than general training knowledge.
Incorrect
Grounding ensures that the AIÂ’s responses are “anchored“ to your specific business data rather than general training knowledge.
Unattempted
Grounding ensures that the AIÂ’s responses are “anchored“ to your specific business data rather than general training knowledge.
Question 5 of 60
5. Question
Which AI Builder model should be used to extract the “Total Balance“ from a photo of a bank statement that has a unique layout?
Correct
The Unstructured Document model is designed to find data points in documents where the layout is not fixed or consistent.
Incorrect
The Unstructured Document model is designed to find data points in documents where the layout is not fixed or consistent.
Unattempted
The Unstructured Document model is designed to find data points in documents where the layout is not fixed or consistent.
Question 6 of 60
6. Question
Scenario: You need to extract “Part Numbers“ from a video stream of a moving assembly line. Which service is required?
Correct
Azure AI Vision can extract text from images and video frames in real-time, even in motion.
Incorrect
Azure AI Vision can extract text from images and video frames in real-time, even in motion.
Unattempted
Azure AI Vision can extract text from images and video frames in real-time, even in motion.
Question 7 of 60
7. Question
Scenario: You need to redact social security numbers from a large dataset of customer emails before using them to ground an AI. Which service should you use?
Correct
PII (Personally Identifiable Information) detection is a feature of Azure AI Language that identifies and masks sensitive data.
Incorrect
PII (Personally Identifiable Information) detection is a feature of Azure AI Language that identifies and masks sensitive data.
Unattempted
PII (Personally Identifiable Information) detection is a feature of Azure AI Language that identifies and masks sensitive data.
Question 8 of 60
8. Question
Which Microsoft service provides the “Model Catalog,“ allowing developers to choose between GPT-4, Llama, and Mistral models?
Correct
Azure AI Studio‘s Model Catalog is the central hub for discovering and deploying both frontier and open-source models.
Incorrect
Azure AI Studio‘s Model Catalog is the central hub for discovering and deploying both frontier and open-source models.
Unattempted
Azure AI Studio‘s Model Catalog is the central hub for discovering and deploying both frontier and open-source models.
Question 9 of 60
9. Question
In Copilot Studio, what is a “System Topic“?
Correct
System topics are built-in templates that manage the foundational conversational flow of every bot.
Incorrect
System topics are built-in templates that manage the foundational conversational flow of every bot.
Unattempted
System topics are built-in templates that manage the foundational conversational flow of every bot.
Question 10 of 60
10. Question
Scenario: You are building an agent that needs to look up a user‘s loyalty points in a legacy system via a REST API. Which node in Copilot Studio is used?
Correct
Power Automate flows act as the bridge between Copilot Studio and external systems via connectors or custom APIs.
Incorrect
Power Automate flows act as the bridge between Copilot Studio and external systems via connectors or custom APIs.
Unattempted
Power Automate flows act as the bridge between Copilot Studio and external systems via connectors or custom APIs.
Question 11 of 60
11. Question
What is “Zero-shot“ prompting?
Correct
Zero-shot prompting relies on the modelÂ’s internal general knowledge to fulfill a request without being given specific task-based examples.
Incorrect
Zero-shot prompting relies on the modelÂ’s internal general knowledge to fulfill a request without being given specific task-based examples.
Unattempted
Zero-shot prompting relies on the modelÂ’s internal general knowledge to fulfill a request without being given specific task-based examples.
Question 12 of 60
12. Question
Which AI Builder model would you use to extract information from a document that does not have a fixed layout, such as a legal letter or an email?
Correct
The Unstructured Document model is designed for documents where the text and data points appear in different locations across different files.
Incorrect
The Unstructured Document model is designed for documents where the text and data points appear in different locations across different files.
Unattempted
The Unstructured Document model is designed for documents where the text and data points appear in different locations across different files.
Question 13 of 60
13. Question
Which Microsoft service provides a secure, “Private Link“ connection to ensure that traffic between your app and Azure OpenAI never travels over the public internet?
Correct
Private Endpoints allow you to access Azure AI services over a private IP address within your Virtual Network (VNet).
Incorrect
Private Endpoints allow you to access Azure AI services over a private IP address within your Virtual Network (VNet).
Unattempted
Private Endpoints allow you to access Azure AI services over a private IP address within your Virtual Network (VNet).
Question 14 of 60
14. Question
In “Responsible AI,“ which principle requires that an AI system should not be a “black box“ and its decisions should be explainable?
Correct
Transparency ensures that the logic and data behind AI decisions are clear and understandable to stakeholders and users.
Incorrect
Transparency ensures that the logic and data behind AI decisions are clear and understandable to stakeholders and users.
Unattempted
Transparency ensures that the logic and data behind AI decisions are clear and understandable to stakeholders and users.
Question 15 of 60
15. Question
Scenario: You are building a Power Automate flow that receives an email in French and needs to summarize it in English. Which two AI Builder actions are required?
Correct
Language detection identifies the source, and the AI Translator converts the text to the target language for summarization.
Incorrect
Language detection identifies the source, and the AI Translator converts the text to the target language for summarization.
Unattempted
Language detection identifies the source, and the AI Translator converts the text to the target language for summarization.
Question 16 of 60
16. Question
Scenario: You need to ensure that an AI agent does not generate responses containing medical advice. Where is the most effective place to define this restriction?
Correct
The System Message defines the behavioral constraints and “safety“ rules that the model must follow.
Incorrect
The System Message defines the behavioral constraints and “safety“ rules that the model must follow.
Unattempted
The System Message defines the behavioral constraints and “safety“ rules that the model must follow.
Question 17 of 60
17. Question
Scenario: You want to create a Copilot agent that can pull real-time weather data from an external API. Which Power Platform component is the bridge?
Correct
Custom Connectors allow Power Platform tools to communicate with any service that has a RESTful API.
Incorrect
Custom Connectors allow Power Platform tools to communicate with any service that has a RESTful API.
Unattempted
Custom Connectors allow Power Platform tools to communicate with any service that has a RESTful API.
Question 18 of 60
18. Question
Scenario: A developer wants to create an AI agent that can summarize a user‘s recent SharePoint files. Which Copilot Studio feature provides this “out-of-the-box“ connectivity?
Correct
Copilot Studio allows you to add SharePoint or OneDrive as a Knowledge Source, enabling the agent to search and summarize documents automatically.
Incorrect
Copilot Studio allows you to add SharePoint or OneDrive as a Knowledge Source, enabling the agent to search and summarize documents automatically.
Unattempted
Copilot Studio allows you to add SharePoint or OneDrive as a Knowledge Source, enabling the agent to search and summarize documents automatically.
Question 19 of 60
19. Question
In a “Vector Search“ index, how is the similarity between a query and a document primarily calculated?
Correct
Vector search relies on the distance between high-dimensional numerical representations of text to find semantic matches.
Incorrect
Vector search relies on the distance between high-dimensional numerical representations of text to find semantic matches.
Unattempted
Vector search relies on the distance between high-dimensional numerical representations of text to find semantic matches.
Question 20 of 60
20. Question
Which “Responsible AI“ principle requires that AI systems be built to perform reliably and safely even under unexpected conditions?
Correct
This principle ensures the system is resilient, robust, and does not cause harm during failures or edge cases.
Incorrect
This principle ensures the system is resilient, robust, and does not cause harm during failures or edge cases.
Unattempted
This principle ensures the system is resilient, robust, and does not cause harm during failures or edge cases.
Question 21 of 60
21. Question
Scenario: An intelligent application needs to search through thousands of scanned PDF legal documents. Which Azure AI Search feature allows the system to “read“ the text inside the images?
Correct
OCR is a cognitive skill in the AI Search enrichment pipeline that extracts text from images and PDFs for indexing.
Incorrect
OCR is a cognitive skill in the AI Search enrichment pipeline that extracts text from images and PDFs for indexing.
Unattempted
OCR is a cognitive skill in the AI Search enrichment pipeline that extracts text from images and PDFs for indexing.
Question 22 of 60
22. Question
Which tool is used to evaluate the “Groundedness“ of an AIÂ’s response by comparing it to the retrieved context?
Correct
Azure AI Studio provides automated evaluation tools to calculate metrics like groundedness, coherence, and relevance.
Incorrect
Azure AI Studio provides automated evaluation tools to calculate metrics like groundedness, coherence, and relevance.
Unattempted
Azure AI Studio provides automated evaluation tools to calculate metrics like groundedness, coherence, and relevance.
Question 23 of 60
23. Question
In Copilot Studio, what is the purpose of the “On Message Received“ system trigger?
Correct
This trigger allows for global logic, such as logging or pre-processing text, every time a user sends a message.
Incorrect
This trigger allows for global logic, such as logging or pre-processing text, every time a user sends a message.
Unattempted
This trigger allows for global logic, such as logging or pre-processing text, every time a user sends a message.
Question 24 of 60
24. Question
What is a “Hallucination“ in a Large Language Model?
Correct
Hallucinations occur when a model generates plausible-sounding but false information not supported by its training or context.
Incorrect
Hallucinations occur when a model generates plausible-sounding but false information not supported by its training or context.
Unattempted
Hallucinations occur when a model generates plausible-sounding but false information not supported by its training or context.
Question 25 of 60
25. Question
What is the primary advantage of using a “Phi-3“ model over a “GPT-4“ model for a mobile-based intelligent application?
Correct
Phi-3 models are optimized for efficiency and can be deployed in environments where low latency and low power consumption are critical.
Incorrect
Phi-3 models are optimized for efficiency and can be deployed in environments where low latency and low power consumption are critical.
Unattempted
Phi-3 models are optimized for efficiency and can be deployed in environments where low latency and low power consumption are critical.
Question 26 of 60
26. Question
In the context of LLMs, what is “Model Distillation“?
Correct
Distillation creates more efficient Small Language Models (SLMs) that retain much of the reasoning capability of larger models but at a lower cost.
Incorrect
Distillation creates more efficient Small Language Models (SLMs) that retain much of the reasoning capability of larger models but at a lower cost.
Unattempted
Distillation creates more efficient Small Language Models (SLMs) that retain much of the reasoning capability of larger models but at a lower cost.
Question 27 of 60
27. Question
Which Microsoft service allows you to define “Data Loss Prevention“ (DLP) policies to control which connectors can be used with AI-driven Power Automate flows?
Correct
The Power Platform Admin Center is the hub for managing DLP policies to ensure sensitive data isn‘t leaked through unauthorized connectors.
Incorrect
The Power Platform Admin Center is the hub for managing DLP policies to ensure sensitive data isn‘t leaked through unauthorized connectors.
Unattempted
The Power Platform Admin Center is the hub for managing DLP policies to ensure sensitive data isn‘t leaked through unauthorized connectors.
Question 28 of 60
28. Question
Which Responsible AI principle is primarily addressed by ensuring that an AI system is accessible to users who rely on screen readers?
Correct
Inclusiveness ensures that technology is usable by everyone, regardless of their physical or cognitive abilities.
Incorrect
Inclusiveness ensures that technology is usable by everyone, regardless of their physical or cognitive abilities.
Unattempted
Inclusiveness ensures that technology is usable by everyone, regardless of their physical or cognitive abilities.
Question 29 of 60
29. Question
Which Microsoft tool is used to manage the Application Lifecycle Management (ALM) of Copilot agents, moving them from “Dev“ to “Test“ to “Production“?
Correct
Solutions are the containers used in the Power Platform to bundle and move apps, flows, and bots across different environments.
Incorrect
Solutions are the containers used in the Power Platform to bundle and move apps, flows, and bots across different environments.
Unattempted
Solutions are the containers used in the Power Platform to bundle and move apps, flows, and bots across different environments.
Question 30 of 60
30. Question
Scenario: A customer service bot needs to distinguish between a “Refund“ request and a “Exchange“ request. Which Copilot Studio feature handles this?
Correct
Trigger phrases and intent recognition map the user‘s natural language to the specific logic for either refunds or exchanges.
Incorrect
Trigger phrases and intent recognition map the user‘s natural language to the specific logic for either refunds or exchanges.
Unattempted
Trigger phrases and intent recognition map the user‘s natural language to the specific logic for either refunds or exchanges.
Question 31 of 60
31. Question
A developer is building a multi-agent system where one agent retrieves data and another summarizes it. Which component decides which agent to call first based on the user‘s goal?
Correct
The Planner is the architectural component that decomposes a complex goal into a sequence of steps or agent calls.
Incorrect
The Planner is the architectural component that decomposes a complex goal into a sequence of steps or agent calls.
Unattempted
The Planner is the architectural component that decomposes a complex goal into a sequence of steps or agent calls.
Question 32 of 60
32. Question
Scenario: You are using AI Builder to extract data from invoices. The invoice layout is consistent across all vendors. Which model type should you use?
Correct
Structured Document models (form processing) are optimized for documents with consistent layouts like standard invoices or applications.
Incorrect
Structured Document models (form processing) are optimized for documents with consistent layouts like standard invoices or applications.
Unattempted
Structured Document models (form processing) are optimized for documents with consistent layouts like standard invoices or applications.
Question 33 of 60
33. Question
In the context of RAG, what is “Chunking“?
Correct
Chunking is necessary because LLMs have context window limits; smaller segments ensure the most relevant pieces of information are retrieved.
Incorrect
Chunking is necessary because LLMs have context window limits; smaller segments ensure the most relevant pieces of information are retrieved.
Unattempted
Chunking is necessary because LLMs have context window limits; smaller segments ensure the most relevant pieces of information are retrieved.
Question 34 of 60
34. Question
Scenario: An application needs to verify if a user‘s uploaded profile picture contains a human face before allowing the upload. Which Azure AI Vision feature is most appropriate?
Correct
Face Detection identifies the presence of human faces and can also provide attributes like head pose or emotions.
Incorrect
Face Detection identifies the presence of human faces and can also provide attributes like head pose or emotions.
Unattempted
Face Detection identifies the presence of human faces and can also provide attributes like head pose or emotions.
Question 35 of 60
35. Question
Which Responsible AI principle is primarily addressed by providing a “Report a Problem“ or “Give Feedback“ button on an AI‘s response?
Correct
Accountability involves creating mechanisms for human feedback and oversight to ensure the system is operating correctly and safely.
Incorrect
Accountability involves creating mechanisms for human feedback and oversight to ensure the system is operating correctly and safely.
Unattempted
Accountability involves creating mechanisms for human feedback and oversight to ensure the system is operating correctly and safely.
Question 36 of 60
36. Question
When configuring an Azure OpenAI deployment, what does the “Max Tokens“ parameter specifically limit?
Correct
Max Tokens limits the response length and helps manage costs by preventing the model from generating excessively long outputs.
Incorrect
Max Tokens limits the response length and helps manage costs by preventing the model from generating excessively long outputs.
Unattempted
Max Tokens limits the response length and helps manage costs by preventing the model from generating excessively long outputs.
Question 37 of 60
37. Question
What does the “Top P“ parameter do in an LLM configuration?
Correct
Top P (Nucleus Sampling) is an alternative to Temperature that controls the diversity of the output by narrowing the word choice pool.
Incorrect
Top P (Nucleus Sampling) is an alternative to Temperature that controls the diversity of the output by narrowing the word choice pool.
Unattempted
Top P (Nucleus Sampling) is an alternative to Temperature that controls the diversity of the output by narrowing the word choice pool.
Question 38 of 60
38. Question
Which component is used to securely store and manage the API keys for Azure AI Services when deploying an application to Azure App Service?
Correct
Azure Key Vault is the industry-standard service for securely storing secrets, keys, and certificates.
Incorrect
Azure Key Vault is the industry-standard service for securely storing secrets, keys, and certificates.
Unattempted
Azure Key Vault is the industry-standard service for securely storing secrets, keys, and certificates.
Question 39 of 60
39. Question
Scenario: You are building an agent that must perform a calculation and then update a record in a legacy SQL database. Which tool is best for this action?
Correct
Power Automate acts as the logic engine that allows an agent to perform multi-step actions across different connectors.
Incorrect
Power Automate acts as the logic engine that allows an agent to perform multi-step actions across different connectors.
Unattempted
Power Automate acts as the logic engine that allows an agent to perform multi-step actions across different connectors.
Question 40 of 60
40. Question
In Microsoft Copilot Studio, what is the purpose of the “On Unknown Intent“ system topic?
Correct
This topic handles situations where the orchestrator cannot confidently map the user‘s message to an existing user-defined topic.
Incorrect
This topic handles situations where the orchestrator cannot confidently map the user‘s message to an existing user-defined topic.
Unattempted
This topic handles situations where the orchestrator cannot confidently map the user‘s message to an existing user-defined topic.
Question 41 of 60
41. Question
What is the “Context Window“ of a Large Language Model?
Correct
The context window defines the memory limit for a single conversation or prompt session.
Incorrect
The context window defines the memory limit for a single conversation or prompt session.
Unattempted
The context window defines the memory limit for a single conversation or prompt session.
Question 42 of 60
42. Question
What is the primary purpose of “Semantic Ranking“ in an Azure AI Search index?
Correct
Semantic ranking uses language models to understand the intent behind a query, ensuring the most relevant results appear first.
Incorrect
Semantic ranking uses language models to understand the intent behind a query, ensuring the most relevant results appear first.
Unattempted
Semantic ranking uses language models to understand the intent behind a query, ensuring the most relevant results appear first.
Question 43 of 60
43. Question
What is the primary role of “Microsoft Entra ID“ in an intelligent application architecture?
Correct
Entra ID (formerly Azure AD) handles authentication and authorization, ensuring only permitted users/apps can access the AI.
Incorrect
Entra ID (formerly Azure AD) handles authentication and authorization, ensuring only permitted users/apps can access the AI.
Unattempted
Entra ID (formerly Azure AD) handles authentication and authorization, ensuring only permitted users/apps can access the AI.
Question 44 of 60
44. Question
Which Microsoft Dataverse feature ensures that an AI agent cannot read payroll data unless the logged-in user has the correct security clearance?
Correct
Field-level security allows administrators to restrict access to specific sensitive columns, even if the user has access to the rest of the table.
Incorrect
Field-level security allows administrators to restrict access to specific sensitive columns, even if the user has access to the rest of the table.
Unattempted
Field-level security allows administrators to restrict access to specific sensitive columns, even if the user has access to the rest of the table.
Question 45 of 60
45. Question
Scenario: You are building an app to scan warehouse shelves. You need to identify if a specific shelf is “Empty“ or “Full.“ Which AI Builder model type should you train?
Correct
Image Classification is used to categorize an entire image (or a specific area) into pre-defined labels like “Empty“ or “Full.“
Incorrect
Image Classification is used to categorize an entire image (or a specific area) into pre-defined labels like “Empty“ or “Full.“
Unattempted
Image Classification is used to categorize an entire image (or a specific area) into pre-defined labels like “Empty“ or “Full.“
Question 46 of 60
46. Question
Scenario: You are building a Power Automate flow that extracts data from shipping labels. The labels are often blurry or tilted. Which AI Builder model is most resilient for this?
Correct
The Unstructured Document model uses deep learning to identify text and fields even when the layout varies or the image quality is imperfect.
Incorrect
The Unstructured Document model uses deep learning to identify text and fields even when the layout varies or the image quality is imperfect.
Unattempted
The Unstructured Document model uses deep learning to identify text and fields even when the layout varies or the image quality is imperfect.
Question 47 of 60
47. Question
In Azure AI Search, what is “Hybrid Search“?
Correct
Hybrid search combines traditional keyword matching with modern vector search to provide the most accurate retrieval results.
Incorrect
Hybrid search combines traditional keyword matching with modern vector search to provide the most accurate retrieval results.
Unattempted
Hybrid search combines traditional keyword matching with modern vector search to provide the most accurate retrieval results.
Question 48 of 60
48. Question
What is a “Hallucination“ in Generative AI?
Correct
Hallucinations occur when an LLM generates data that is factually incorrect but appears plausible.
Incorrect
Hallucinations occur when an LLM generates data that is factually incorrect but appears plausible.
Unattempted
Hallucinations occur when an LLM generates data that is factually incorrect but appears plausible.
Question 49 of 60
49. Question
Which “Responsible AI“ principle requires that a developer be able to explain why an AI system made a specific decision?
Correct
Transparency focuses on the “explainability“ of AI systems, ensuring that their logic and decision-making processes can be understood by humans.
Incorrect
Transparency focuses on the “explainability“ of AI systems, ensuring that their logic and decision-making processes can be understood by humans.
Unattempted
Transparency focuses on the “explainability“ of AI systems, ensuring that their logic and decision-making processes can be understood by humans.
Question 50 of 60
50. Question
Scenario: A developer wants to build a bot that can query a real-time stock price database. Which component acts as the bridge between the bot and the database API?
Correct
Custom Connectors allow Power Platform tools to securely communicate with external REST APIs or databases.
Incorrect
Custom Connectors allow Power Platform tools to securely communicate with external REST APIs or databases.
Unattempted
Custom Connectors allow Power Platform tools to securely communicate with external REST APIs or databases.
Question 51 of 60
51. Question
Which Azure AI service feature allows you to “Translate“ a text document while preserving the original complex formatting like tables and headers?
Correct
Document Translation is a specialized feature of the Translator service that maintains the layout and structure of the original file.
Incorrect
Document Translation is a specialized feature of the Translator service that maintains the layout and structure of the original file.
Unattempted
Document Translation is a specialized feature of the Translator service that maintains the layout and structure of the original file.
Question 52 of 60
52. Question
Scenario: You need to redact Social Security Numbers from a dataset before using it to ground an AI agent. Which service should you use?
Correct
PII (Personally Identifiable Information) detection identifies and masks sensitive data points within a text string.
Incorrect
PII (Personally Identifiable Information) detection identifies and masks sensitive data points within a text string.
Unattempted
PII (Personally Identifiable Information) detection identifies and masks sensitive data points within a text string.
Question 53 of 60
53. Question
Which Microsoft service provides “Prompt Shields“ to protect against Indirect Prompt Injection (where malicious instructions are hidden in data)?
Correct
Prompt Shields are specifically designed to detect and block malicious content hidden within user inputs or retrieved data.
Incorrect
Prompt Shields are specifically designed to detect and block malicious content hidden within user inputs or retrieved data.
Unattempted
Prompt Shields are specifically designed to detect and block malicious content hidden within user inputs or retrieved data.
Question 54 of 60
54. Question
In Copilot Studio, what is the purpose of the “On Conversation Start“ system trigger?
Correct
This trigger is essential for personalizing the start of a session or verifying authentication before the user speaks.
Incorrect
This trigger is essential for personalizing the start of a session or verifying authentication before the user speaks.
Unattempted
This trigger is essential for personalizing the start of a session or verifying authentication before the user speaks.
Question 55 of 60
55. Question
Scenario: You are using Azure AI Video Indexer. You need to identify the exact moments when a specific company logo appears on screen. Which feature provides this?
Correct
Brand Detection identifies thousands of world-famous logos and brands within the visual frames of a video.
Incorrect
Brand Detection identifies thousands of world-famous logos and brands within the visual frames of a video.
Unattempted
Brand Detection identifies thousands of world-famous logos and brands within the visual frames of a video.
Question 56 of 60
56. Question
Which Azure service provides the “Prompt Shields“ capability to detect and block Indirect Prompt Injection attacks?
Correct
Azure AI Content Safety includes Prompt Shields to protect against malicious inputs designed to subvert model instructions.
Incorrect
Azure AI Content Safety includes Prompt Shields to protect against malicious inputs designed to subvert model instructions.
Unattempted
Azure AI Content Safety includes Prompt Shields to protect against malicious inputs designed to subvert model instructions.
Question 57 of 60
57. Question
Scenario: You want to restrict an AI bot to only answer in “French“ regardless of the user‘s input language. Where is the best place to define this?
Correct
The System Message defines the core behavior, persona, and linguistic constraints for the AI session.
Incorrect
The System Message defines the core behavior, persona, and linguistic constraints for the AI session.
Unattempted
The System Message defines the core behavior, persona, and linguistic constraints for the AI session.
Question 58 of 60
58. Question
Scenario: A retail company wants an AI agent that can check a SQL database for product availability in real-time. How should this be implemented in Copilot Studio?
Correct
Plugins allow agents to perform actions and fetch live data from external systems via Power Automate connectors.
Incorrect
Plugins allow agents to perform actions and fetch live data from external systems via Power Automate connectors.
Unattempted
Plugins allow agents to perform actions and fetch live data from external systems via Power Automate connectors.
Question 59 of 60
59. Question
Scenario: You are designing a Power App that uses AI to scan warehouse inventory. You need to identify if a specific shelf contains a “Box“ or a “Pallet.“ Which AI Builder model is best?
Correct
Image Classification is used to tag or categorize an entire image into pre-defined labels.
Incorrect
Image Classification is used to tag or categorize an entire image into pre-defined labels.
Unattempted
Image Classification is used to tag or categorize an entire image into pre-defined labels.
Question 60 of 60
60. Question
Which Dataverse feature allows an AI agent to see if a specific product is “In Stock“ by looking at a calculated sum of several warehouses?
Correct
Rollup columns automatically calculate aggregates (like sum or count) from related records, providing real-time data for the AI.
Incorrect
Rollup columns automatically calculate aggregates (like sum or count) from related records, providing real-time data for the AI.
Unattempted
Rollup columns automatically calculate aggregates (like sum or count) from related records, providing real-time data for the AI.
X
Use Page numbers below to navigate to other practice tests