Courseiva

Microsoft Certified: Azure AI Apps and Agents Developer Associate (AI-103) (AI-103) (AI-103) — Questions 301375

510 questions total · 7pages · All types, answers revealed

Page 4

Page 5 of 7

Page 6
301
MCQhard

You need to ensure that an LLM-based agent in your application follows a specific output schema (JSON) for downstream processing. What is the most reliable way to enforce this?

A.Use a few-shot prompt with JSON examples
B.Use a custom parser in Prompt Flow
C.Configure the response_format property to JSON
D.Set the Stop Sequences parameter
AnswerC

This feature forces structured output.

Why this answer

Structured output (Response Format) parameters ensure the model adheres to a specific schema.

302
MCQhard

You are developing a solution that uses custom text classification in Azure AI Language. You choose the multi-label classification architecture. What is the key characteristic of multi-label text classification?

A.A single document can be assigned zero, one, or multiple categories simultaneously.
B.Each document must be assigned to exactly one mutually exclusive category.
C.The output is restricted to binary positive or negative values.
D.Categories are automatically generated by the service without user training data.
AnswerA

Multi-label classification allows documents to be tagged with multiple independent categories.

Why this answer

In multi-label classification, a single document can be assigned zero, one, or multiple distinct category labels simultaneously.

303
MCQhard

Your application integrates Azure AI Vision and processes images stored in an Azure Blob Storage container. To adhere to security best practices, you want to avoid exposing public blob URLs. How should you provide the image to the Image Analysis API?

A.Use FTP to transfer images to the vision service endpoint.
B.Download the image bytes in your application code and send them in the binary request body (application/octet-stream).
C.Permanently change the blob container access level to Public Read.
D.Mount the Blob Storage container directly to the Azure AI Vision service via NFS.
AnswerB

Sending binary bytes directly avoids the need for public URLs or SAS tokens.

Why this answer

Instead of public URLs, you can pass the image binary bytes directly in the HTTP request body (using application/octet-stream), or generate a short-lived SAS token URL.

304
MCQmedium

You are training a Custom Vision multiclass classification model for retail products. After training, you notice that the precision is high, but the recall is low for a specific product category. What does low recall indicate?

A.The training dataset had zero images for that category.
B.The model is incorrectly predicting that product when it is not present (high false positives).
C.The model is missing many actual instances of that product category (high false negatives).
D.The probability threshold is set too high for all classes uniformly.
AnswerC

Recall measures the model's ability to find all actual positive instances; low recall means many true items were missed.

Why this answer

Low recall indicates that the model has a high false negative rate—it misses actual instances of the product category.

305
Multi-Selecteasy

Which TWO file formats are natively supported as inputs for analysis in Azure AI Document Intelligence? (Choose TWO)

Select 2 answers
A.text/html
B.application/pdf
C.application/zip
D.image/jpeg
E.application/vnd.ms-excel
AnswersB, D

PDF files are natively supported by all Document Intelligence models.

Why this answer

Document Intelligence supports standard document formats like PDF, JPEG, PNG, TIFF, and BMP.

306
MCQmedium

You are configuring abstractive summarization in Azure AI Language. You want to control the length of the generated summary output. Which parameter should you configure in your request body?

A.maxSummaryLength
B.documentLengthLimit
C.sentenceCount
D.confidenceThreshold
AnswerA

maxSummaryLength specifies the maximum length of the output summary in abstractive summarization.

Why this answer

Abstractive summarization requests accept a 'max和社会' or max-length parameter depending on the API version to control summary length, specifically 'maxSummaryLength'.

307
Multi-Selectmedium

When designing a Responsible AI evaluation workflow in Azure AI Foundry, which TWO metrics are specifically used to assess content safety and potential harms? (Choose two.)

Select 2 answers
A.Hate speech and unfairness detection
B.Word count frequency
C.BLEU machine translation score
D.Sexual content detection
E.Model parameter count
AnswersA, D

Hate speech and unfairness metrics assess content safety and bias.

Why this answer

Violence, sexual, hate, and self-harm detection along with protected material detection are core safety metrics in Azure AI Foundry.

308
Multi-Selectmedium

You are preparing an ARM template to deploy an Azure AI multi-service account. Which TWO properties are mandatory within the ARM template resource definition? Each correct answer presents part of the solution.

Select 2 answers
A.VirtualMachineSize
B.sku
C.OperatingSystemType
D.KernelVersion
E.location
AnswersB, E

The sku property defines the pricing tier and is required for Azure AI resource provisioning.

Why this answer

Name, kind, sku, and location are essential ARM template properties for cognitive services.

309
MCQmedium

You are configuring a code interpreter tool for an AI agent in Azure AI Agent Service. The agent needs to analyze a CSV dataset uploaded by a user. How does the agent execute this task securely?

A.The agent compiles the CSV data into C# binaries and runs them on the host Azure App Service instance.
B.The agent sends the raw CSV data to an external public third-party execution API.
C.The agent generates Python code to analyze the CSV and executes it inside a secure, isolated sandbox environment managed by the service.
D.The agent translates the CSV data into SQL queries and executes them directly against the primary production database.
AnswerC

The code interpreter operates within an isolated containerized sandbox, ensuring secure execution of generated Python scripts.

Why this answer

The code interpreter runs in a secure, sandboxed execution environment where it dynamically writes and executes Python code to process data and generate files.

310
Multi-Selecthard

You are implementing indirect prompt injection defenses in a RAG application where documents are ingested from public web pages. Which THREE mitigation strategies should you deploy? (Choose THREE)

Select 3 answers
A.Use strict structural delimiters (like XML tags) to isolate retrieved context from system instructions
B.Trust all incoming web page text unconditionally without inspection
C.Disable all content moderation filters to maximize processing speed
D.Implement output validation filters to check model completions for unauthorized actions or exfiltration attempts
E.Enable Prompt Shields for indirect prompt injection in Azure AI Content Safety
AnswersA, D, E

XML tags help the model distinguish untrusted retrieved data from trusted system instructions.

Why this answer

Prompt shields in Content Safety, strict prompt delimiters separating data from instructions, and content validation filters protect against indirect prompt injection.

311
MCQeasy

You are designing a generative AI solution in Azure AI Foundry. You need to ensure that user data sent to Azure OpenAI is not used to train Microsoft's foundational models. What is Microsoft's default data privacy policy for Azure OpenAI?

A.Customer data is stored in public Azure storage accounts
B.Customer data is not used to train foundation models by default
C.Customer data is automatically published to public hugging face repositories
D.Customer data is used for model training unless you opt out by submitting a support ticket
AnswerB

Azure OpenAI enterprise privacy guarantees that prompt and completion data remain strictly confidential to your tenant.

Why this answer

Microsoft guarantees that customer data and prompts sent to Azure OpenAI are not used to train or improve foundation models.

312
MCQhard

You are processing health-related clinical notes using the Azure AI Language service. You need to extract medical concepts and link them to standardized medical vocabularies such as ICD-10-CM and SNOMED-CT. Which specific feature and endpoint task should you use?

A.Named Entity Recognition (NER) with standard prebuilt-entity models
B.Key phrase extraction with medical domain parameters
C.Custom Named Entity Recognition with a trained ICD-10 dictionary
D.Healthcare text analytics (Healthcare-specific NER and relation extraction)
AnswerD

Healthcare text analytics extracts health-related entities and links them to specialized ontologies like UMLS, SNOMED-CT, and ICD-10.

Why this answer

Healthcare text analytics is a specialized capability of the Azure AI Language service designed to extract and link medical entities to standard terminologies.

313
Multi-Selectmedium

You are designing an agentic workflow in Prompt Flow. Which TWO components are required to define a tool-based agent? (Choose two)

Select 2 answers
A.A LLM to act as the agent brain
B.Tool definition (OpenAPI spec or Python function)
C.A SQL Database
D.A Power Automate cloud flow
E.An Azure Container Registry
AnswersA, B

The model performs the reasoning to decide which tool to call.

Why this answer

Agents require a tool definition (to perform actions) and a base model (to reason).

314
Multi-Selecthard

You are designing a high-availability architecture for an Azure AI Speech service solution. Which THREE practices should you incorporate into your design? (Choose three.)

Select 3 answers
A.Hardcode the primary region IP address in all client applications without using a routing DNS layer.
B.Configure Azure Front Door or Traffic Manager with health probes to route traffic away from unhealthy regions.
C.Set up Azure Monitor alert rules to notify operations teams of sudden spikes in server error rates or latency.
D.Deploy Speech resources in multiple Azure regions for active-passive or active-active failover.
E.Rely exclusively on a single-region deployment without backups to minimize operational complexity.
AnswersB, C, D

Traffic routing services with health probes enable automated failover.

Why this answer

High availability involves multi-region deployments, automated health checks with Azure Front Door, and proper monitoring/alerts.

315
MCQhard

You are designing a multi-tenant SaaS application on Azure. Each tenant requires isolation of their Azure OpenAI token usage and cost tracking. What is the recommended provisioning strategy to achieve strict cost and quota isolation per tenant?

A.Use a single Azure OpenAI resource for all tenants and track usage using custom HTTP headers.
B.Deploy all tenants in the F0 free pricing tier.
C.Provision a separate Azure OpenAI resource for each tenant.
D.Rely on Azure Cost Management resource group tags alone without resource separation.
AnswerC

Dedicated resources per tenant ensure clean billing separation, isolated rate limits, and distinct quota management.

Why this answer

Provisioning a dedicated Azure OpenAI resource per tenant provides absolute isolation for quotas, usage metrics, and cost tracking.

316
MCQmedium

Your organization has multiple Azure AI services deployed across different resource groups. Your security officer requests a centralized dashboard to monitor security recommendations, vulnerability assessments, and compliance posture for all AI resources. Which Azure service should you open?

A.Azure Advisor
B.Microsoft Entra ID protection portal
C.Azure Policy compliance blade
D.Microsoft Defender for Cloud
AnswerD

Defender for Cloud provides secure score recommendations, vulnerability tracking, and compliance assessments for Azure AI services.

Why this answer

Microsoft Defender for Cloud provides centralized visibility and security posture recommendations across all Azure resources, including AI services.

317
MCQeasy

You are building an app to analyze customer feedback. You want to extract named entities such as people, locations, and organizations. Which Azure AI Language feature should you use?

A.Language detection
B.Sentiment analysis
C.Named Entity Recognition (NER)
D.Key phrase extraction
AnswerC

NER identifies and categorizes entities like people, locations, and organizations.

Why this answer

Named Entity Recognition (NER) extracts known entities from text and categorizes them into predefined classes such as people, places, and organizations.

318
Multi-Selectmedium

When configuring a Custom Vision project, which TWO project types are available? (Choose two.)

Select 2 answers
A.Speech Synthesis
B.Form Extraction
C.Classification
D.Text Recognition (OCR)
E.Object Detection
AnswersC, E

Classification is a supported project type for tagging entire images.

Why this answer

Custom Vision supports two main project types: Classification (Multiclass/Multilabel) and Object Detection.

319
Multi-Selecthard

Which THREE of the following are best practices for Responsible AI when using Generative AI solutions?

Select 3 answers
A.Removing all system instructions to maximize creativity
B.Implementing content safety filters
C.Collecting all user PII for model training
D.Maintaining human-in-the-loop oversight
E.Providing transparency to users that they are interacting with an AI
AnswersB, D, E

Filtering mitigates harmful content.

Why this answer

Transparency, safety filtering, and human oversight are essential pillars.

320
MCQhard

Your team is implementing 'Human-in-the-loop' (HITL) for an agent that performs account updates. Which mechanism should you use to pause the agent and wait for approval?

A.Set a long timeout in the API call.
B.Increase the model's 'stop_sequences'.
C.Use an infinite loop in Python.
D.Use a Prompt Flow 'Pause' node or custom state management logic.
AnswerD

Custom flow design can handle waiting for external triggers, such as human approval.

Why this answer

In agentic workflows, an interruption or 'callback' mechanism allows the system to pause execution until human input is received.

321
MCQhard

You are performing Retrieval-Augmented Generation. You need to ensure the retrieved data includes metadata for filtering. Where should this metadata be defined?

A.Environment variables
B.Prompt Flow input schema
C.System message content
D.Azure AI Search Index Fields
AnswerD

Metadata must be indexed as fields.

Why this answer

Metadata is defined in the Azure AI Search index schema.

322
MCQeasy

You are deploying a model in Azure AI Foundry. You need to ensure that the model endpoint is protected by a virtual network. Which configuration should you use?

A.Enable public internet access with IP firewall rules
B.Enable Managed Identity on the endpoint
C.Deploy a dedicated Azure Application Gateway
D.Configure a Private Endpoint for the AI resource
AnswerD

Private endpoints map the resource to a private IP in your VNet.

Why this answer

Private endpoints provide secure access to Azure AI resources within a virtual network.

323
MCQmedium

You are using Azure AI Language custom text classification. You have uploaded your training dataset and trained a model. You notice that your evaluation metrics show low precision. What does low precision indicate about your model?

A.Many of the items predicted as a specific class actually belonged to a different class.
B.The model failed to find many actual positive instances in the dataset.
C.The latency of the REST API calls is too high.
D.The training dataset contained fewer than 5 documents per class.
AnswerA

Low precision indicates high false positives, meaning items were incorrectly labeled as the target class.

Why this answer

Precision measures the ratio of true positive predictions to total predicted positives. Low precision means the model produces many false positives.

324
Multi-Selecthard

You are developing an application that uses the Azure AI Language REST API for text analysis. Which TWO requirements must be met when constructing HTTP requests to the service? (Choose two.)

Select 2 answers
A.Format all request payloads as XML documents.
B.Include the master account password in the request query string.
C.Set the 'Content-Type' header to 'application/json'.
D.Authenticate exclusively using HTTP Basic Authentication with username and password.
E.Include the 'Ocp-Apim-Subscription-Key' header containing your API subscription key.
AnswersC, E

JSON payloads require Content-Type: application/json.

Why this answer

Requests to Azure AI Language require the Ocp-Apim-Subscription-Key header (or token authorization) and the Content-Type header set to application/json.

325
Multi-Selecthard

You are designing a deployment strategy for an Azure AI Custom Vision project in an enterprise pipeline. Which THREE practices should you follow to ensure secure and repeatable deployments? Each correct answer presents part of the solution.

Select 3 answers
A.Store API keys and connection strings securely in Azure Key Vault rather than hardcoding them.
B.Manually create and configure all production resources directly in the Azure portal without documentation.
C.Use Infrastructure as Code (ARM or Bicep) to provision Azure resources consistently.
D.Implement automated integration and validation tests in CI/CD pipelines before production release.
E.Store production API keys as plaintext comments in Bicep template files.
AnswersA, C, D

Key Vault protects sensitive credentials from exposure in code repositories.

Why this answer

Using IaC, securing secrets in Key Vault, and automated validation tests are key practices.

326
Multi-Selecthard

You are troubleshooting an Azure OpenAI deployment in Azure AI Foundry where agents are failing to complete tasks due to excessive timeouts. Which THREE diagnostic steps should you take? (Choose three.)

Select 3 answers
A.Check if the tokens-per-minute (TPM) quota is throttling requests, causing queued delays.
B.Double the model temperature to 2.0 to accelerate reasoning speed.
C.Delete all vector indices to force local memory processing.
D.Inspect Application Insights telemetry to analyze execution duration per agent step and tool call latency.
E.Increase the client-side HTTP timeout threshold on the API request client.
AnswersA, D, E

TPM throttling causes requests to queue or time out before processing.

Why this answer

Troubleshooting agent timeouts involves examining token limits, network latency, and model complexity.

327
MCQhard

You are auditing your Azure AI services for cost optimization. You discover an Azure AI Search service that has been over-provisioned with too many search units and replica nodes. How can you right-size the service without losing your existing indexes?

A.Switch the pricing tier from Standard to Free.
B.Export all index definitions, delete the service, and re-import the indexes into a new Basic tier service.
C.Modify the scale settings in the Azure portal to reduce the number of replicas and partitions.
D.Delete the search service and recreate it with fewer replicas and partitions.
AnswerC

Scaling down via the scale blade adjusts resource allocation while preserving index data.

Why this answer

You can scale down search units and replicas directly through the Azure portal or management API without losing indexes, provided the data fits within the reduced storage limits.

328
Multi-Selecthard

You are designing an advanced agentic solution in Azure AI Foundry that requires tool integration and state management. Which TWO practices should you implement to ensure robust agent reliability and security? (Choose two)

Select 2 answers
A.Implement human-in-the-loop confirmation steps before the agent executes high-impact or destructive transactional tool calls.
B.Validate and sanitize all parameters generated by the model before passing them into downstream enterprise tool APIs.
C.Disable all logging and telemetry to minimize latency during agent execution loops.
D.Grant the agent unrestricted administrative permissions to all corporate databases to maximize autonomous problem-solving.
E.Rely solely on the model's internal memory to maintain multi-session conversation state without external persistence.
AnswersA, B

Human-in-the-loop validation ensures critical actions (like financial transactions or data deletion) are approved by a user.

Why this answer

Validating tool inputs and implementing human-in-the-loop validation for sensitive actions are critical best practices for agent reliability and security.

329
MCQeasy

You need to detect the language of a batch of text documents and want to minimize the number of API calls. What is the maximum number of documents you can include in a single request payload to the language detection endpoint?

A.1 document
B.1,000 documents
C.10 documents
D.100,000 documents
AnswerC

Synchronous batch requests support up to 10 documents per call.

Why this answer

Azure AI Language batch APIs typically support up to 10 documents per synchronous request payload.

330
MCQmedium

Your organization uses Azure Monitor diagnostic settings to route logs and metrics for all Azure AI services to a centralized Log Analytics workspace. You need to write a Kusto Query Language (KQL) query to find all incoming API requests that resulted in an HTTP 500 server error over the last 24 hours. Which table should you query?

A.AppTraces
B.AzureDiagnostics
C.AzureActivity
D.SecurityEvent
AnswerB

AzureDiagnostics captures service-specific telemetry and log categories for Cognitive Services.

Why this answer

Azure AI Services diagnostic logs are sent to the AzureDiagnostics table when using legacy resource-specific or workspace-based diagnostic settings.

331
MCQmedium

You are deploying a custom text classification model using Azure AI Language. You need to implement cost management controls and track exact expenditures by tagging the Azure AI resource with the internal project cost center 'CC-409'. Where should you configure these tags in the Azure Portal?

A.Within the Networking blade under Private endpoint connections
B.Under the Resource management > Tags section of the Azure AI Services resource blade
C.In the Access control (IAM) settings as a custom role assignment
D.In the Diagnostic settings by creating a new metric alert rule
AnswerB

Resource tags are managed directly in the Tags blade of the Azure resource.

Why this answer

Tags applied at the Azure Resource Group or individual Azure AI resource level in the Azure Portal propagate to Azure Cost Management for granular tracking and reporting. Access control, network configuration, and diagnostic settings do not handle resource-level billing tags.

332
MCQeasy

Your organization requires that all data sent to Azure OpenAI Service remains encrypted using a key that your security team fully controls and rotates. Which feature should you configure during resource creation?

A.Azure Information Protection client-side encryption
B.Customer-managed keys (CMK) with Azure Key Vault
C.Service-managed keys (Default encryption)
D.Transparent Data Encryption with a SQL Managed Instance
AnswerB

CMK allows you to use your own Key Vault keys to encrypt resource data.

Why this answer

Customer-managed keys (CMK) via Azure Key Vault allow organizations to manage encryption keys used to secure Azure AI service data.

333
MCQhard

You notice your Azure AI agent is producing offensive content. You need to configure a guardrail that intercepts the model output before it is sent to the user. Which service should you integrate?

A.Azure AI Search Filters
B.Prompt Template Validation
C.Azure Monitor
D.Azure AI Content Safety
AnswerD

Content Safety provides real-time filtering.

Why this answer

Azure AI Content Safety is the standard guardrail for filtering model output.

334
MCQeasy

Which role is responsible for ensuring the ethical deployment of an AI agent within an organization?

A.The Data Scientist only.
B.The end user.
C.The Azure AI Developer.
D.The Azure Billing Admin.
AnswerC

The developer is responsible for implementing the necessary safety and ethical configurations.

Why this answer

Responsible AI practices are a cross-functional responsibility, but the developer must implement the technical guardrails provided by the organization.

335
MCQmedium

You are designing a customer feedback processing system. You need to extract key phrases from customer reviews and store them in an Azure Cosmos DB database. When you inspect the JSON response from the key phrase API, what data structure is returned for the key phrases?

A.A single concatenated string with semicolons separating each phrase.
B.An array of strings representing the key phrases found in the document.
C.A nested dictionary mapping phrase weights to sentiment scores.
D.An array of objects containing 'text', 'category', and 'confidenceScore' properties.
AnswerB

Key phrases are returned as an array of strings.

Why this answer

The key phrase extraction response returns a list of string values representing the extracted key phrases for each document.

336
MCQeasy

You are reviewing security settings for an Azure AI Services account. You need to verify whether local authentication using API keys is enabled or disabled. Where in the Azure portal can you find this setting?

A.Azure Cost Management -> Billing scope
B.Resource management -> Resource normalization / Networking / Security settings
C.Microsoft Entra ID -> App registrations
D.Azure Monitor -> Diagnostic settings
AnswerB

Azure AI accounts include settings to disable local authentication (API key access) in favor of Microsoft Entra ID exclusively.

Why this answer

The 'Resource management' -> 'Networking' or 'Access configuration' blade in the Azure portal manages local authentication settings (local auth disable toggle).

337
MCQmedium

You are analyzing product reviews using Azure AI Language. You receive sentiment analysis output where a sentence has a positive confidence score of 0.85, a neutral score of 0.10, and a negative score of 0.05. What is the sentiment label assigned to this sentence?

A.neutral
B.positive
C.mixed
D.negative
AnswerB

The highest confidence score is positive (0.85), so the label is positive.

Why this answer

The sentiment label corresponds to the highest confidence score among positive, neutral, and negative. Here, 0.85 is positive.

338
MCQmedium

You are using Azure AI Foundry to deploy an agent. You want to ensure that the agent only uses approved data sources. Which feature should you configure?

A.Prompt engineering for 'Do not use external info'.
B.Grounding data configuration in the Agent settings.
C.Model temperature control.
D.Content Safety policy filters.
AnswerB

Restricting data sources to a specific index ensures grounding compliance.

Why this answer

Grounding with specific data sources restricts the agent to provided data, preventing the use of external knowledge.

339
MCQhard

You are analyzing performance metrics of a Custom Vision object detection iteration. You observe a Mean Average Precision (mAP) of 88%, but when testing with real-world user photos, the model fails to detect objects that are partially occluded. What should you do to improve occlusion handling?

A.Switch from General domain to Food domain.
B.Lower the confidence threshold to 0.01.
C.Add training images containing partially occluded objects and ensure they are correctly tagged.
D.Increase image compression artifacts.
AnswerC

Training with occluded examples teaches the model to recognize partial features of the object.

Why this answer

To improve detection of occluded objects, you must include training images that feature partially blocked or occluded instances of the target objects.

340
MCQeasy

You are developing an application that needs to automatically extract key-value pairs and tabular data from standardized invoices without any prior labeled training data. Which Azure AI Document Intelligence model should you use?

A.Azure AI Document Intelligence General Document model
B.Azure AI Document Intelligence Custom Extraction model
C.Azure AI Document Intelligence Prebuilt Invoice model
D.Azure AI Vision OCR feature
AnswerC

The Prebuilt Invoice model extracts standard invoice fields automatically without requiring custom training.

Why this answer

The Prebuilt Read model extracts text, while the Prebuilt Invoice model is specifically pretrained to extract key-value pairs, line items, and other invoice-specific fields without requiring any training data.

341
Multi-Selectmedium

Which TWO of the following are valid methods to improve the quality of retrieved data in a RAG system?

Select 2 answers
A.Reducing the number of vector dimensions
B.Implementing a hybrid search strategy
C.Optimizing document chunking strategies
D.Changing the chat history session length
E.Increasing the number of tokens in the prompt
AnswersB, C

Hybrid search uses both keyword and vector search.

Why this answer

Better chunking strategies and hybrid search improve retrieval quality.

342
MCQhard

You are configuring diagnostic settings for an Azure AI Search service to audit administrative actions and query logs. Which destination should you choose if you want to stream these logs into a SIEM tool like Microsoft Sentinel for real-time security analysis?

A.Send to an Azure DevOps pipeline
B.Export to a local CSV file via Azure CLI
C.Send to an Event Hub
D.Archive to a storage account
AnswerC

Streaming to an Event Hub is the standard ingestion mechanism for third-party or native SIEM solutions like Microsoft Sentinel.

Why this answer

Streaming diagnostic logs to an Event Hub allows external SIEM tools like Microsoft Sentinel to ingest and analyze the log data in real time.

343
MCQhard

You are building a custom named entity recognition model in Language Studio. You want to extract product serial numbers that follow a specific pattern (two uppercase letters followed by four digits, a hyphen, and three digits). Which entity type definition method should you use?

A.List entity type
B.Learned entity type
C.Regex entity type
D.Prebuilt entity type
AnswerC

Regex entity components allow you to define strict pattern matching for structured codes like serial numbers.

Why this answer

When defining custom entities with rigid patterns, you can use a 'Regex' (regular expression) component to extract items matching exact rules.

344
MCQmedium

Your application processes receipt images using Azure AI Vision Read API. You need to extract text specifically from tables within the receipts. Which service is better suited for advanced table structure and key-value pair extraction compared to base Azure AI Vision Read?

A.Azure AI Face Service
B.Azure AI Document Intelligence
C.Custom Vision Service
D.Azure AI Content Safety
AnswerB

Document Intelligence specializes in extracting tables, key-value pairs, and structured fields from forms and receipts.

Why this answer

Azure AI Document Intelligence (formerly Form Recognizer) is specifically designed for structured data, tables, and key-value extraction from documents and receipts.

345
MCQmedium

You are analyzing product reviews and need to identify specific product features mentioned by users along with whether users expressed positive or negative opinions about those specific features. Which advanced feature of sentiment analysis should you enable?

A.Extractive summarization
B.Opinion mining
C.Key phrase extraction
D.Named Entity Recognition
AnswerB

Opinion mining extracts aspect-level targets and sentiments.

Why this answer

Opinion mining is an advanced feature of sentiment analysis that extracts aspects (targets) and their associated opinions (assessments).

346
MCQeasy

You need to determine the primary dominant colors and whether an image is predominantly black-and-white using Azure AI Vision. Which visual feature parameter provides this information?

A.features=denseCaptions
B.features=color
C.features=people
D.features=tags
AnswerB

The color feature analyzes color properties including dominant colors and black-and-white status.

Why this answer

The color feature returns dominant foreground color, background color, accent color, and whether the image is black and white.

347
Multi-Selecthard

Which THREE actions occur when you 'deploy' an agent from Azure AI Foundry?

Select 3 answers
A.The model is automatically retrained on new data.
B.The agent's configuration is persisted as a versioned deployment.
C.All previous chat history is deleted from the server.
D.An HTTP endpoint is generated for application integration.
E.Capacity and quota are allocated for the deployment.
AnswersB, D, E

Versioning is a core part of deployment.

Why this answer

Deployment makes the model/agent available as an endpoint, often configuring the model version and resource allocation.

348
Multi-Selecteasy

You are building an application to process customer reviews. Which TWO metrics are returned in the standard JSON response of a sentiment analysis document evaluation? (Choose two.)

Select 2 answers
A.Overall document sentiment label
B.Estimated reading time in minutes
C.Detected language translation confidence score
D.Total word count of the document
E.Sentiment confidence scores (positive, neutral, negative)
AnswersA, E

The overall sentiment label (positive, negative, neutral, mixed) is returned.

Why this answer

Sentiment analysis document responses return the overall sentiment label and confidence scores (positive, neutral, negative).

349
MCQmedium

Your company requires strict data privacy. Where should you host your Document Intelligence resources to ensure data remains in a specific region?

A.Regional Azure resource
B.Any available region
C.Global service endpoint
D.Public multi-region endpoint
AnswerA

Deploying the resource in a specific region binds data processing to that location.

Why this answer

Choosing the appropriate Azure region during resource creation ensures data residency compliance.

350
MCQhard

You have an Azure AI application that is hitting rate limits (429 errors). You need to determine which specific client application is causing the issue. How do you identify this?

A.Check the 'Activity Log'.
B.Increase the resource quota in the 'Quotas' blade.
C.Enable diagnostic logging to Log Analytics and run a KQL query.
D.Restart the AI service.
AnswerC

Diagnostic logs provide the granular detail needed to analyze request sources and error patterns.

Why this answer

Diagnostic logs (sent to Log Analytics) can be configured to include client-side metadata or IP information for request tracing.

351
MCQhard

You are building an agent in Azure AI Foundry using a prompt flow. You need to ensure that the agent only accesses specific internal documents and avoids hallucinating outside the provided context. Which grounding technique should you implement?

A.Implement grounding with data using Azure AI Search
B.Use a system message to forbid outside information
C.Enable prompt caching
D.Increase the temperature parameter to 1.0
AnswerA

Grounding with data connects the agent to specific indices.

Why this answer

In Azure AI Foundry, grounding with data allows you to restrict the model to specific sources, minimizing hallucinations.

352
MCQeasy

You are using Azure AI Document Intelligence to extract information from identity documents, such as passports and driver licenses. Which prebuilt model should you select?

A.prebuilt-tax.us.w2
B.prebuilt-idDocument
C.prebuilt-W2
D.prebuilt-businessCard
AnswerB

This model extracts first name, last name, date of birth, document number, and expiry date from identity documents.

Why this answer

The prebuilt ID document model (prebuilt-idDocument) is designed to extract fields from passports and driver licenses.

353
Multi-Selecthard

You are configuring network isolation and security for an Azure AI Search service that connects to an Azure SQL Database backend. Which THREE features can you utilize to secure this data flow? (Choose three.)

Select 3 answers
A.IP firewall rules restricting Azure SQL Database access to the Azure AI Search service outbound IPs.
B.Azure Private Endpoint for Azure AI Search and Azure SQL Database
C.Managed identities and secure connection strings with Entra authentication where supported.
D.Disabling all transport layer security (TLS) requirements to speed up query execution.
E.Storing plain-text database admin credentials in the search index configuration file.
AnswersA, B, C

IP firewalls restrict database access to authorized service endpoints.

Why this answer

Securing data flows between PaaS services involves private endpoints, IP firewalls, and managed identity connections.

354
MCQhard

You are managing an Azure AI Search service with multiple replicas and partitions. During peak query hours, latency increases significantly, and CPU utilization on the search service hits 95%. You need to scale the service to improve query throughput without altering the index data schema. Which scaling action should you take?

A.Increase the number of search partitions.
B.Disable semantic ranking to reduce compute overhead.
C.Increase the number of search replicas in the Azure portal or via CLI.
D.Migrate the service to the Free pricing tier.
AnswerC

Replicas handle search query load; adding replicas increases query throughput and read capacity.

Why this answer

Scaling out by adding search replicas increases query throughput and high availability.

355
MCQhard

You are building an agentic workflow in Azure AI Foundry. You need the agent to perform a multi-step task: search a database, summarize results, and then send an email. Which component is best suited to manage this orchestration?

A.Azure AI Agent Service.
B.Azure OpenAI fine-tuning service.
C.Azure AI Search indexer.
D.Azure AI Content Safety API.
AnswerA

This service provides the execution engine for agentic workflows.

Why this answer

The Agent Service provides the orchestration framework for agents to manage state, tools, and multi-step reasoning.

356
MCQhard

You are processing customer feedback using sentiment analysis in the Azure AI Language service. You need to extract finer-grained information about specific product attributes mentioned in the reviews (e.g., finding out that the screen is bright but the battery life is poor). Which feature should you enable in your API request?

A.Opinion mining
B.PII detection
C.Language detection
D.Custom text classification
AnswerA

Opinion mining performs aspect-based sentiment analysis, linking sentiments directly to target features or aspects in the text.

Why this answer

Opinion mining is an extension of sentiment analysis that provides more granular information about the opinions related to product or service attributes (aspect-based sentiment analysis).

357
Multi-Selecthard

You are optimizing an enterprise RAG application in Azure AI Foundry for high accuracy and minimal latency. Which THREE advanced retrieval techniques should you consider implementing? (Choose THREE)

Select 3 answers
A.Semantic ranking to re-rank top search results using deep learning models
B.Using random sentence selection without indexing
C.Hybrid search combining full-text BM25 and vector embeddings
D.Disabling vector search entirely to rely solely on random string matching
E.Query expansion and re-writing to enrich user prompts with synonyms and related terms
AnswersA, C, E

Semantic ranking elevates the most relevant passages to the top of the result set.

Why this answer

Hybrid search, semantic ranking, and query expansion/rewriting are advanced retrieval optimization techniques in Azure AI Search.

358
MCQmedium

You need to ensure that your agent is compliant with regional data residency requirements. Where should you check this?

A.In the Prompt Flow runtime.
B.In the Content Safety policy.
C.In the model configuration settings.
D.In the Azure AI Foundry resource's region settings.
AnswerD

Regional residency is defined at the resource level.

Why this answer

Data residency is determined by the Azure region where the service instance is deployed.

359
MCQhard

You are building an automated asset management system that processes millions of images annually. To optimize costs and latency, you want to minimize the number of API calls while extracting tags, bounding box objects, and OCR text simultaneously. How should you structure your API request?

A.Submit three separate asynchronous requests for tags, objects, and read.
B.Include multiple comma-separated values in the features parameter, such as features=tags,objects,read.
C.Use the legacy Computer Vision v3.2 batch analysis endpoint.
D.Train a Custom Vision model to combine all three tasks.
AnswerB

Combining features in a single analyze request avoids redundant API calls and processing overhead.

Why this answer

Image Analysis 4.0 allows combining multiple features in a single API call using the 'features' query parameter.

360
MCQhard

You are deploying an Azure AI Speech container on an Azure Kubernetes Service (AKS) cluster. Your security policy dictates that pod-to-pod communication within the cluster must be encrypted in transit using mTLS, and egress traffic to Azure must flow through a strict outbound firewall. Which Kubernetes feature should you implement?

A.Azure Container Registry vulnerability scanning
B.Kubernetes Network Policies combined with a service mesh for mTLS
C.Azure Bastion deployed inside the AKS worker node subnet
D.AKS cluster auto-scaler configuration
AnswerB

Network policies govern IP-level pod isolation, while a service mesh enforces mTLS encryption for pod-to-pod communication.

Why this answer

Kubernetes network policies control traffic flow between pods, while a service mesh (like Istio or Linkerd) provides mTLS encryption in transit.

361
MCQeasy

You need to evaluate customer reviews to determine whether the overall tone is positive, negative, or neutral. Which Azure AI Language feature should you use?

A.Language detection
B.Named Entity Recognition
C.Key phrase extraction
D.Sentiment analysis
AnswerD

Sentiment analysis evaluates the emotional tone of text.

Why this answer

Sentiment analysis provides sentiment labels (positive, negative, neutral) and confidence scores at both the document and sentence levels.

362
MCQeasy

When sharing a Prompt Flow with a teammate, which action allows them to run the flow without having to recreate the connections?

A.Delete the connection and recreate it
B.Grant permissions to the shared Connection resource
C.Hardcode credentials into the flow
D.Export the flow as a text file
AnswerB

Permissions allow others to use the defined credentials.

Why this answer

Sharing the project and ensuring appropriate RBAC or shared connection access allows collaboration.

363
MCQeasy

You are provisioning a new Azure AI multi-service account in the West US region for an enterprise application. Which pricing tier should you select to ensure the account supports all available Azure AI capabilities including Custom Neural Voice and Content Moderator?

A.Free (F0)
B.Isolated (I0)
C.Basic (B0)
D.Standard (S0)
AnswerD

The S0 tier provides full access to all Azure AI multi-service capabilities without feature restrictions.

Why this answer

The Standard (S0) pricing tier is required for full access to standard and advanced Azure AI services. The Free (F0) tier has transaction limits and restricted feature sets.

364
MCQmedium

You are migrating an Azure AI Speech service resource to a different Azure subscription within the same tenant. Which mechanism should you use to preserve the existing custom voice models and configurations?

A.Use the Azure portal resource move feature to transfer the Speech resource to the target subscription.
B.Export custom models as JSON and import them via the Speech Studio in the new subscription.
C.Export the resource ARM template and deploy it to the target subscription.
D.Recreate the custom models from scratch in the new subscription using a new Speech resource.
AnswerA

Moving the resource directly via Azure resource move preserves all associated models and endpoints.

Why this answer

Azure Resource Mover or subscription-level resource move operations allow moving cognitive service resources while retaining custom models and configurations, provided regional constraints are met.

365
MCQeasy

You are provisioning an Azure AI services account and want to view your current API key. Where can you retrieve the API key in the Azure portal?

A.Resource graph explorer
B.Azure Advisor cost optimization pane
C.Azure Active Directory overview page
D.Keys and Endpoint blade of the resource
AnswerD

The Keys and Endpoint blade provides access to Key 1 and Key 2 for authenticating API requests.

Why this answer

The 'Keys and Endpoint' blade of an Azure AI resource displays the generated primary and secondary keys.

366
Multi-Selecthard

When configuring a custom Named Entity Recognition (NER) project in Language Studio using the Azure AI Language service, which THREE steps or configuration options are required during the model lifecycle?

Select 3 answers
A.Define your entity schema and types
B.Fine-tune the underlying base transformer weights via direct GPU gradient backpropagation code
C.Manually convert all input text files into pre-tokenized JSON vectors
D.Train the model using your labeled dataset
E.Deploy the trained model to a deployment name for client application consumption
AnswersA, D, E

You must define the entity types (e.g., ProductNumber, VendorName) before labeling data.

Why this answer

Building a custom NER model involves defining the entity schema, tagging or labeling your documents, training the model, evaluating performance metrics, and finally deploying the trained model.

367
MCQmedium

You are using Prompt Flow to iterate on an agent. What is the benefit of using 'Environment Variables' in your flow?

A.To manage configuration settings like API keys and endpoints across environments.
B.To reduce the model response time.
C.To store large data files.
D.To increase the prompt length.
AnswerA

Environment variables provide a secure and flexible way to handle configuration.

Why this answer

Environment variables allow you to switch between development, test, and production configurations without changing code.

368
MCQeasy

You are building an application that analyzes user feedback. You need to detect instances where users share their phone numbers, email addresses, and credit card numbers so that you can automatically redact them. Which feature should you use?

A.Sentiment analysis
B.Language detection
C.Named Entity Recognition (NER)
D.PII detection
AnswerD

PII detection identifies, classifies, and optionally redacts sensitive personal data.

Why this answer

Personally Identifiable Information (PII) detection identifies and redacts sensitive information such as phone numbers, emails, and financial data.

369
Multi-Selectmedium

You are planning the security architecture for an enterprise Azure AI multi-service account. Management requires that access to the resource be strictly controlled and audited. Which TWO security features or configurations should you implement? Each correct answer presents part of the solution.

Select 2 answers
A.Disable local authentication (API key access) and require Microsoft Entra ID authentication.
B.Store all admin API keys permanently in an unencrypted Azure Blob Storage container.
C.Configure an Azure Private Endpoint to restrict data access to the virtual network.
D.Grant the Contributor role to all end-user client applications.
E.Enable anonymous public access across all subnets.
AnswersA, C

Disabling local auth eliminates static API keys and enforces centralized identity governance.

Why this answer

Disabling local auth enforces Entra ID, and Private Endpoints secure network traffic.

370
MCQmedium

Your organization is deploying an Azure OpenAI resource. Your security policy requires that all network traffic destined for the OpenAI service must traverse a secure hub-and-spoke virtual network topology and pass through an Azure Firewall instance. How should you route outbound traffic from your application subnet to Azure OpenAI?

A.Configure Azure Traffic Manager dns routing.
B.Attach an Azure Bastion subnet to the firewall.
C.Rely on automatic default system routing without route tables.
D.Configure User Defined Routes (UDR) on the application subnet with a next hop pointing to the Azure Firewall.
AnswerD

UDRs route traffic from the subnet through the firewall before reaching external or PaaS endpoints.

Why this answer

User Defined Routes (UDR) allow you to override default Azure routing tables and force traffic through an Azure Firewall or network virtual appliance.

371
MCQhard

You are experiencing high latency in your RAG agent. You confirm the model response time is fine, but the retrieval step is slow. What should you investigate?

A.The system prompt length.
B.The content safety threshold.
C.The model temperature.
D.The index retrieval configuration and semantic ranker parameters.
AnswerD

Search index settings and ranking overhead directly impact retrieval speed.

Why this answer

Vector search latency is often tied to index configuration, partitioning, or query complexity.

372
Multi-Selectmedium

Your organization is implementing a model versioning and rollout strategy for Azure OpenAI deployments. Which TWO practices ensure safe deployment and quick rollback capabilities? Each correct answer presents part of the solution.

Select 2 answers
A.Delete the entire Azure AI resource group prior to every model update.
B.Overwrite existing model weights in place without changing deployment names.
C.Deploy new model versions with distinct deployment names alongside existing versions.
D.Configure traffic splitting weights to canary test new models with a small percentage of users before full rollout.
E.Hardcode model version numbers directly inside immutable compiled client binaries.
AnswersC, D

Distinct deployment names allow running multiple versions simultaneously for testing.

Why this answer

Using unique deployment names and traffic splitting ensures safe rollouts and rollbacks.

373
MCQmedium

You are using Azure AI Vision Read API in your application. You want to ensure that the extracted text blocks return word-level bounding polygons rather than just line-level bounding boxes. What parameter or response structure should you examine?

A.Add ?wordLevel=true to the query string.
B.Examine the 'words' array within each returned 'lines' object in the JSON response payload.
C.Call the legacy OCR endpoint instead.
D.Set the model-version parameter to v2.0.
AnswerB

The JSON response for Read results contains line objects which include nested word objects with individual bounding polygons.

Why this answer

The Read API response structure automatically includes word-level details inside each line object (`lines[*].words[*]`), providing bounding polygons for both lines and individual words.

374
Multi-Selecthard

You are troubleshooting a Custom Vision object detection model that exhibits poor performance in production. Which THREE actions can you take using the Custom Vision portal or APIs to diagnose and improve the model? (Choose three.)

Select 3 answers
A.Delete all project iteration history to reset model weights to random initialization.
B.Examine prediction suggestions to find low-confidence or incorrectly predicted images and add them to training data.
C.Review precision and recall metrics per tag in the Performance tab.
D.Adjust the probability threshold on the performance page to evaluate precision/recall trade-offs.
E.Manually edit model neural network weights via the REST API.
AnswersB, C, D

Adding hard examples from prediction suggestions (active learning) improves model robustness.

Why this answer

Portal diagnostics include reviewing prediction performance metrics (Precision, Recall, mAP), tagging missed detections (Smart Labeling / Suggested Tags), and adjusting probability thresholds.

375
MCQeasy

Which command line tool is primarily used for managing Azure resources, including AI services, from a terminal?

A.Azure PowerShell.
B.Azure CLI.
C.Docker.
D.Kubectl.
AnswerB

The Azure CLI is the primary tool for cross-platform Azure resource management.

Why this answer

The Azure CLI (az) is the standard command-line interface for managing Azure resources.

Page 4

Page 5 of 7

Page 6

All pages