AI-103 Practice Exam 2 — 20 Questions

Question 1

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?

Accepted Answer

Configure a Private Endpoint for the AI resource. Configuring a Private Endpoint for the AI resource is correct because Azure Private Link creates a private IP address within your virtual network for the Azure AI Foundry resource, ensuring all traffic to the model endpoint stays on the Microsoft backbone and never traverses the public internet. This is the standard mechanism for VNet-protected PaaS endpoints in Azure.

Answer

Enable public internet access with IP firewall rules

Answer

Enable Managed Identity on the endpoint

Answer

Deploy a dedicated Azure Application Gateway

Question 2

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?

Accepted Answer

Under the Resource management > Tags section of the Azure AI Services resource blade. Tags in Azure are configured under the Resource management > Tags section of the Azure AI Services resource blade. This allows you to assign metadata such as cost center 'CC-409' to the resource, which is then used for cost tracking and management. Tags are key-value pairs that can be applied to resources, resource groups, and subscriptions for organizational and billing purposes.

Answer

Within the Networking blade under Private endpoint connections

Answer

In the Access control (IAM) settings as a custom role assignment

Answer

In the Diagnostic settings by creating a new metric alert rule

Question 3

In Azure AI Foundry, which tool would you use to measure the quality of your RAG application against a golden dataset?

Accepted Answer

Evaluation. The Evaluation tool in Azure AI Foundry is designed to measure the quality of generative AI applications, including RAG systems, against a golden dataset using metrics such as groundedness, relevance, coherence, and fluency. It provides both built-in and custom evaluators, making it the correct choice for quality measurement. Model Catalog, Deployment, and Prompt Flow Tracing serve different purposes.

Answer

Model Catalog

Answer

Deployment

Answer

Prompt Flow Tracing

Question 4

You are developing an application to extract specific fields from invoices using Azure AI Document Intelligence. Which prebuilt model should you use for this task?

Accepted Answer

prebuilt-invoice. The prebuilt-invoice model is purpose-built for invoice documents and returns structured fields such as VendorName, InvoiceDate, InvoiceTotal, DueDate, CustomerName, and line items without any training. Because the task is extracting specific fields from invoices, this prebuilt model directly matches the document type and required output. Custom models or generic layout models would require extra work and would not return invoice-specific fields out of the box.

Answer

custom-template

Answer

prebuilt-layout

Answer

prebuilt-document

Question 5

You are troubleshooting a low confidence score on specific fields extracted by your Azure AI Document Intelligence custom template model. Which TWO actions can help improve the extraction confidence for those fields? (Choose TWO)

Accepted Answer

Ensure all training documents are of high visual quality and free from blur or heavy skew.. Option C is correct because custom template models learn field positions and visual patterns from the labeled samples, so blurry, skewed, or low-quality training documents degrade the model's ability to reliably locate and read those fields, producing low confidence scores; supplying sharp, straight, high-quality scans directly improves feature extraction and OCR accuracy. Option D is correct because low confidence on specific fields usually indicates the model has not seen enough layout variation for those fields, so adding diverse training documents that capture the different layouts, fonts, and positions where those fields appear lets the model generalize and raise its confidence. Option A is wrong because the confidence threshold in the analyzeDocument request only filters which results are returned or flagged; it does not change the model's actual extraction confidence. Option B is wrong because converting to grayscale and downsampling to 72 DPI reduces image fidelity and typically lowers OCR and extraction quality rather than improving confidence. Option E is wrong because a prebuilt business card model is a fixed, general-purpose model that cannot be trained on your specific fields or layouts, so it cannot improve confidence for your custom template fields.

Answer

Increase the confidence threshold parameter in the analyzeDocument request body to 0.99.

Answer

Convert all training PDFs to grayscale and reduce their resolution to 72 DPI to speed up processing.

Answer

Switch the model type from custom template to a prebuilt business card model.

Question 6

You are developing a global news aggregation application using the Azure AI Language service. You need to automatically identify the primary language of incoming user-submitted articles across 120 different languages before routing them to translation pipelines. Which API endpoint should you call?

Accepted Answer

POST /language/:analyze-text?api-version=2023-04-01 with the TextLanguageDetection task. The TextLanguageDetection task in the /language/:analyze-text endpoint is specifically designed to identify the language of a document and return the detected language code and confidence score. It supports the full range of Azure AI Language languages (currently 120+), making it the correct choice for routing multilingual articles.

Answer

POST /language/translate?api-version=3.0

Answer

POST /language/:analyze-text?api-version=2023-04-01 with the CustomEntityRecognition task

Answer

POST /language/analyze-text/jobs?api-version=2023-04-01 with the TextSentimentAnalysis task

Question 7

You are deploying an LLM in Azure AI Foundry and need to monitor for toxicity and jailbreak attempts in real-time. Which feature should you enable?

Accepted Answer

Azure AI Content Safety. Azure AI Content Safety is the service designed to detect harmful content, including toxicity and jailbreak attempts, in real time for LLM applications. It provides classifiers for hate, violence, sexual, and self-harm content, plus Prompt Shields for jailbreak and indirect prompt injection detection. Enabling it allows pre- and post-processing moderation of prompts and completions.

Answer

Azure Policy

Answer

Microsoft Purview Compliance Manager

Answer

Azure Monitor Log Analytics

Question 8

You are configuring an Azure AI Document Intelligence custom extraction pipeline that processes thousands of diverse medical forms daily. You notice that extraction confidence scores for certain hand-written fields are low. What is the most effective remediation step within the Document Intelligence workflow?

Accepted Answer

Add additional labeled training documents that specifically feature diverse handwriting samples for those low-confidence fields.. The most effective remediation for low-confidence extraction of handwritten fields in a custom Document Intelligence model is to provide more labeled training data that represents the variability causing the low confidence. Handwriting is highly variable, so adding diverse handwriting samples for the specific fields helps the model learn the patterns and improve accuracy. This directly addresses the root cause—insufficient training examples for those field variations—within the Document Intelligence workflow.

Answer

Disable the OCR engine parameter in the API call to bypass text recognition steps.

Answer

Decrease the model training iterations configuration setting to prevent overfitting on handwriting.

Answer

Switch from a Custom Neural model to a Custom Template model.

Question 9

Your custom extraction model is failing to identify fields on a new version of a document. You have labeled 50 documents. What is the most likely solution?

Accepted Answer

Add more training documents with the new layout. When a custom extraction model fails on a new document layout, the most likely fix is to add more training documents that represent the new layout so the model can learn the additional variations. Custom models (especially neural) improve with more diverse, labeled examples covering the layouts they must handle; 50 documents of the old layout are insufficient for the new version. Retraining with representative samples is the standard remedy.

Answer

Change the API version to v2.1

Answer

Switch to a template model

Answer

Increase the confidence threshold

Question 10

Your company's security policy requires that all Azure AI Services accounts disable public network access completely. However, internal applications hosted inside a specific Azure Virtual Network must still communicate with the AI services. How should you configure network access?

Accepted Answer

Disable public network access and configure a Private Endpoint for the resource.. To disable public network access while allowing internal applications in a specific Azure Virtual Network to communicate with Azure AI Services, you should disable public network access and configure a Private Endpoint. This creates a private IP address in the VNet for the service, enabling secure private connectivity.

Answer

Configure an inbound Network Security Group (NSG) rule on the AI resource subnet.

Answer

Keep public network access enabled and rely on API keys for security.

Answer

Deploy a VPN gateway inside the Azure AI resource group.

Question 11

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?

Accepted Answer

Microsoft Defender for Cloud. Microsoft Defender for Cloud provides a centralized security posture management dashboard that aggregates security recommendations, vulnerability assessments, regulatory compliance, and threat protection across Azure resources, including Azure AI services. It is the correct service to open when a security officer requests a unified view of security recommendations and compliance posture across multiple resource groups.

Answer

Azure Advisor

Answer

Microsoft Entra ID protection portal

Answer

Azure Policy compliance blade

Question 12

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?

Accepted Answer

Provision a separate Azure OpenAI resource for each tenant.. Provisioning a separate Azure OpenAI resource per tenant provides independent quota limits, separate keys, and distinct billing meters, which together deliver strict cost and quota isolation. This is the recommended multi-tenant pattern when isolation is a hard requirement.

Answer

Use a single Azure OpenAI resource for all tenants and track usage using custom HTTP headers.

Answer

Deploy all tenants in the F0 free pricing tier.

Answer

Rely on Azure Cost Management resource group tags alone without resource separation.

Question 13

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?

Accepted Answer

positive. Azure AI Language sentiment analysis assigns the label corresponding to the highest confidence score among positive, neutral, and negative. Here positive is 0.85, which is the maximum, so the sentence is labeled positive. The scores are a probability distribution that sums to approximately 1.0.

Answer

neutral

Answer

mixed

Answer

negative

Question 14

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?

Accepted Answer

Regex entity type. A regex entity type is designed to extract entities that match a specific regular expression pattern, such as two uppercase letters followed by four digits, a hyphen, and three digits (e.g., '[A-Z]{2}\d{4}-\d{3}'). In Language Studio, you define the regex pattern and optionally provide examples, making it ideal for structured identifiers like serial numbers.

Answer

List entity type

Answer

Learned entity type

Answer

Prebuilt entity type

Question 15

You are deploying an Azure AI multi-service account in the West US region. The finance department requires that all resource deployment costs be tracked by department code using Azure Resource Manager tags. Which Azure feature should you use to enforce the mandatory application of the department tag across all future Azure AI resource deployments?

Accepted Answer

Azure Policy. Azure Policy is the governance service that enforces organizational standards by evaluating resources against rules and taking effects such as deny, audit, or append. A policy with an 'append' or 'deny' effect on the tag requirement enforces mandatory department tags across all future deployments.

Answer

Azure Cost Management budgets

Answer

Azure role-based access control (RBAC)

Answer

Azure Blueprints

Question 16

You are implementing prompt engineering for a customer support agent. To minimize hallucinations when the model lacks information, which Prompt Flow technique is most effective?

Accepted Answer

Instructional grounding to only answer from context. Instructional grounding — explicitly instructing the model to answer only from the provided context and to say it doesn't know otherwise — is the most direct and effective prompt-engineering technique to minimize hallucinations when the model lacks information. It constrains the model's generation to the retrieved context rather than its parametric memory.

Answer

Chain-of-thought prompting

Answer

Increasing the Top_P parameter

Answer

Few-shot prompting with varied examples

Question 17

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?

Accepted Answer

prebuilt-idDocument. The prebuilt-idDocument model is specifically designed to extract key information from identity documents such as passports, driver licenses, and U.S. state IDs. It returns structured fields like name, date of birth, document number, and expiration date. This is the only prebuilt model among the options that targets identity documents.

Answer

prebuilt-tax.us.w2

Answer

prebuilt-W2

Answer

prebuilt-businessCard

Question 18

You are developing a RAG solution using Azure AI Search. Users are searching using domain-specific acronyms that are not well-represented in standard embedding models, resulting in poor retrieval scores. What feature should you enable in Azure AI Search to improve keyword matching for exact acronyms?

Accepted Answer

Hybrid search combining vector search and full-text BM25 search. Hybrid search in Azure AI Search runs vector (k-NN) search and full-text BM25 keyword search in parallel, then fuses the results using Reciprocal Rank Fusion (RRF). BM25 excels at exact term matching, so domain-specific acronyms that embedding models underrepresent will still be matched lexically, dramatically improving recall for those queries. This is the standard remedy when semantic/vector retrieval alone misses rare tokens.

Answer

Increasing the embedding dimension size to 3072

Answer

Enabling Azure AI Content Safety filters

Answer

Setting the semantic ranker threshold to zero

Question 19

Which file format is NOT natively supported for training custom models in Azure AI Document Intelligence?

Accepted Answer

DOCX. Azure AI Document Intelligence custom model training accepts PDF, JPEG, PNG, BMP, TIFF, and HEIF image formats, but DOCX (Word) files are not natively supported as training inputs. To train on Word content, you must first convert DOCX to PDF or render pages as images before uploading to the labeled training set. This is why DOCX is the correct answer to 'which is NOT natively supported.'

Answer

PDF

Answer

JPEG

Answer

BMP

Question 20

You are provisioning an Azure AI Language resource and want to ensure that all telemetry and activity logs are retained for compliance auditing over a 365-day period. Where should you configure this retention policy?

Accepted Answer

In the resource's Diagnostic settings by specifying a retention period on a storage account destination.. Diagnostic settings in Azure allow you to configure the collection and retention of telemetry and activity logs. By specifying a retention period (e.g., 365 days) on a storage account destination, you ensure logs are retained for compliance auditing. This is the correct place to set retention policies for Azure AI Language resource logs.

Answer

In the Microsoft Defender for Cloud security policy blade.

Answer

In the Azure AI Language service Studio configuration panel.

Answer

In the Azure Cost Management budget alert settings.