AI-102 Plan and manage an Azure AI solution Practice Test 6 — 15 Questions

Question 1

Which THREE factors should you consider when selecting the pricing tier for an Azure AI Language resource to handle a high-volume production workload?

Accepted Answer

Transactions per second (TPS) limits. Option A (Transactions per second (TPS) limits) is correct because Azure AI Language pricing tiers such as Free, Standard S, and Standard S0 differ in their per-second request throughput, and a high-volume production workload must select a tier whose TPS ceiling will not throttle requests. Option B (Maximum document size) is correct because tiers impose different maximum document/request size limits (for example, the free tier caps documents at a smaller size than paid tiers), and exceeding that limit causes request failures in production. Option E (Latency guarantees) is correct because higher pricing tiers provide better latency/performance characteristics, which is critical for a high-volume production workload with response-time requirements. Option C (Language support) is not a tier-dependent factor, since the same language and feature availability applies across tiers for a given Azure AI Language resource. Option D (SLA percentage) is not a differentiator between pricing tiers, as the SLA is defined at the service level rather than varying by selected tier.

Answer

Language support

Answer

SLA percentage

Question 2

You are reviewing the response from an Azure OpenAI Service chat completion API call. The finish_reason is 'stop'. What does this indicate?

Accepted Answer

The model completed the response naturally. The `finish_reason` field in the Azure OpenAI chat completion API response indicates why the model stopped generating tokens. A value of `'stop'` means the model encountered a natural stopping point, such as the end of a sentence or a logical conclusion, and completed the response without hitting any limits or filters.

Answer

The response was truncated because the token limit was reached

Answer

The response was blocked by the content filter

Answer

The model is still generating the response

Question 3

You are deploying an Azure AI Language service custom text classification model. You need to ensure that the training data is balanced and representative. What should you do?

Accepted Answer

Split the labeled data into training and test sets, ensuring each class has similar proportions.. Splitting labeled data into training and test sets while ensuring each class has similar proportions (stratified split) is a standard practice for balanced and representative training data. This approach prevents class imbalance from skewing model evaluation metrics and ensures the model generalizes well to unseen data. In Azure AI Language custom text classification, this is critical for achieving reliable performance across all classes.

Answer

Use only the most frequent labels to train the model.

Answer

Oversample the minority classes to match the majority class size.

Answer

Use all labeled data for training and rely on cross-validation.

Question 4

Your company uses Azure OpenAI Service to generate product descriptions. You need to reduce costs while maintaining response quality for common requests. What should you implement?

Accepted Answer

Enable semantic caching to reuse responses for similar prompts.. Semantic caching stores responses for semantically similar prompts, allowing Azure OpenAI Service to return cached results for common requests without incurring per-token inference costs. This directly reduces costs while preserving response quality, as the cached responses are identical to what the model would generate. Unlike other options, it avoids degrading output quality or requiring architectural changes.

Answer

Use a smaller model variant (e.g., GPT-3.5 instead of GPT-4).

Answer

Batch multiple requests into a single API call.

Answer

Reduce the max_tokens parameter for all requests.

Question 5

You need to provide a business analyst with access to create and manage Azure AI Language projects without granting them full subscription-level permissions. What role should you assign?

Accepted Answer

Cognitive Services Language Owner at the resource level. The Cognitive Services Language Owner role at the resource level grants full permissions to create, read, update, and delete Azure AI Language projects and resources, including managing custom models and deployments, without granting any permissions outside that specific resource. This is the least-privilege role that satisfies the business analyst's need to manage Language projects while avoiding subscription-level access.

Answer

Reader at the resource group level

Answer

Contributor at the subscription level

Answer

Cognitive Services User at the resource level

Question 6

Your organization uses Microsoft Purview to govern AI models. You need to ensure that data used for training complies with internal policies. What should you configure?

Accepted Answer

Set up Data Estate Insights to monitor data usage. Data Estate Insights in Microsoft Purview provides monitoring and analytics over data estate health, including data usage patterns. By configuring Data Estate Insights, you can track how training data is accessed and used, enabling you to verify compliance with internal policies on data consumption for AI model training.

Answer

Use Microsoft Purview Data Map to catalog training data

Answer

Apply sensitivity labels to training data

Answer

Enable Microsoft Purview Information Protection

Question 7

You are designing a solution that uses Azure AI Translator to translate customer support tickets. The solution must handle confidential data and ensure no data leaves the specified Azure region. What should you do?

Accepted Answer

Create a Translator resource with a private endpoint and disable public network access. Using a private endpoint ensures that all traffic to the Translator resource stays within the Azure virtual network and never traverses the public internet, while disabling public network access enforces that no data can leave the specified Azure region. This combination meets the requirement for handling confidential data with regional data residency guarantees.

Answer

Enable customer-managed keys and disable logging

Answer

Use the global Translator endpoint

Answer

Use the S1 pricing tier to ensure regional processing

Question 8

Which THREE considerations are important when planning to use Azure OpenAI Service in a production environment?

Accepted Answer

Configure content filtering to block harmful outputs. Option A is correct because Azure OpenAI Service applies content filtering by default, and in production you must configure and tune these filters (for categories such as hate, violence, sexual, and self-harm) to block harmful outputs and meet responsible AI requirements. Option C is correct because Azure OpenAI enforces per-deployment tokens-per-minute (TPM) and requests-per-minute (RPM) quotas, so you must plan quota allocations and rate-limit handling to avoid throttling under expected production load. Option D is correct because data residency and processing-region requirements determine which Azure regions can host your Azure OpenAI resources and where prompts/completions are processed, which is a key compliance consideration. Option B is not correct because Azure OpenAI provides pre-trained models accessed via the service; you do not train a foundation model from scratch, and fine-tuning is optional rather than a core planning consideration. Option E is not correct because Azure OpenAI is accessed as a managed Azure service endpoint and does not support being deployed inside a customer VNet like IaaS workloads; network isolation is instead handled through Private Link/private endpoints, which is a connectivity choice rather than a fundamental planning consideration.

Answer

Train a custom model from scratch using your own data

Answer

Deploy the service within a virtual network (VNet)

Question 9

Refer to the exhibit. You are reviewing an Azure OpenAI Service API request. The deployment-id is 'gpt-4o'. The user asks 'What is the capital of France?' The response is cut off mid-sentence. Based on the parameters, what is the most likely cause?

Accepted Answer

The stop parameter is causing early termination. The `stop` parameter in an Azure OpenAI API request defines a sequence of tokens that, when generated, causes the model to stop producing further output. If the `stop` sequence appears in the generated text, the response will be truncated at that point, even mid-sentence. In this scenario, the cut-off response is most likely due to the `stop` parameter matching a token in the generated output, not because of token limits or temperature settings.

Answer

The max_tokens value is too low

Answer

The temperature setting is too high

Answer

The system message is missing a context

Question 10

You are the AI architect for a global e-commerce company. The company uses Azure AI services to power its product recommendation engine. The current solution uses Azure AI Language to extract product attributes from descriptions and Azure AI Search to index and retrieve products. The company is expanding to new markets and must comply with data residency regulations in the European Union and Asia. Additionally, the solution must handle a 10x increase in traffic during holiday sales without downtime. You need to design a solution that meets these requirements. The solution must use Azure AI Language and Azure AI Search. You have the following options: A) Deploy separate Azure AI Language and Azure AI Search resources in each region (EU and Asia) and use Azure Traffic Manager to route users to the nearest region. Use geo-replication for Azure AI Search. B) Deploy a single Azure AI Language resource in the US and use Azure AI Search with geo-replication to handle traffic. Use Azure Front Door for global routing. C) Deploy Azure AI Language resources in each region and use a global Azure AI Search resource with replication. Use Azure Load Balancer to distribute traffic. D) Deploy Azure AI Language and Azure AI Search resources in the US only, and use Azure CDN to cache responses globally. Which option should you choose?

Accepted Answer

Deploy separate resources in each region with Traffic Manager and geo-replication.. Deploying separate Azure AI Language and Azure AI Search resources in each region (EU and Asia) ensures compliance with data residency regulations by keeping data within regional boundaries. Azure Traffic Manager routes users to the nearest region for low latency, and geo-replication for Azure AI Search provides high availability and disaster recovery, handling the 10x traffic increase during holiday sales without downtime.

Answer

Deploy a single US resource with geo-replication and Front Door.

Answer

Deploy US-only resources with CDN caching.

Answer

Deploy Language in each region with a global Search resource.

Question 11

Your organization is deploying an Azure AI service for document translation. The solution must support custom glossaries and real-time translation. Cost optimization is a key requirement. Which pricing tier should you choose?

Accepted Answer

Standard S1. The Standard S1 tier is the correct choice because it supports custom glossaries and real-time translation, which are required for the document translation solution. Additionally, S1 offers a balance of features and cost, making it optimal for production workloads where cost optimization is a key requirement, unlike the Free F0 tier which has limited throughput and no SLA, or the Premium S2 tier which is designed for high-volume scenarios at a higher cost.

Answer

Free F0

Answer

Basic S0

Answer

Premium S2

Question 12

You are planning an Azure AI solution that processes sensitive customer data. Compliance requires encryption at rest using a customer-managed key (CMK) stored in Azure Key Vault. The AI resource must also be accessible only from specific virtual networks. Which two configurations are necessary? (Choose two.)

Accepted Answer

Enable customer-managed key encryption for the Azure AI service.. Option C is correct because the compliance requirement explicitly mandates encryption at rest with a customer-managed key (CMK) stored in Azure Key Vault, which is enabled by configuring CMK encryption on the Azure AI service (the service uses the key via a Key Vault reference and requires a managed identity with get/wrap/unwrap permissions on the key). Option E is correct because restricting the AI resource to specific virtual networks is achieved by disabling public network access and creating a private endpoint, which places a private IP from the chosen VNet/subnet into the service via Azure Private Link and blocks access from the public internet. Option A is not required by the scenario itself: a managed identity is needed to access the Key Vault for CMK, but the question asks for the two configurations that satisfy the stated requirements, and the CMK and private endpoint configurations are the direct answers (identity is an implementation detail of CMK, not a separate requirement). Option B is unnecessary because Azure Firewall inspects outbound traffic and does not restrict inbound access to the AI resource. Option D is incorrect because allowing public access from selected IP addresses still exposes the service over the public internet, which contradicts the requirement that the resource be accessible only from specific virtual networks.

Answer

Assign a system-assigned managed identity to the Azure AI service.

Answer

Deploy Azure Firewall to inspect outbound traffic.

Answer

Enable public network access from selected IP addresses.

Question 13

A company uses Azure AI Document Intelligence to extract data from invoices. Recently, extraction accuracy dropped for new vendor formats. Which strategy should you implement to improve accuracy without retraining the entire model?

Accepted Answer

Train a custom extraction model using labeled examples from the new vendor formats.. Training a custom extraction model using labeled examples from the new vendor formats allows Azure AI Document Intelligence to learn the specific layout and field patterns of those invoices without affecting the prebuilt model's performance on other formats. This approach leverages the custom model training capability, which is designed to adapt to unique document structures while preserving the existing extraction logic for previously supported formats.

Answer

Increase the confidence threshold for extraction results.

Answer

Switch from Document Intelligence to Azure AI Language service.

Answer

Increase the number of transactions per second (TPS) limit.

Question 14

Your team uses Azure AI Foundry to deploy a custom chat model. The model must meet compliance by explaining its reasoning and citing sources. Which feature should you enable?

Accepted Answer

Groundedness detection with citation. Groundedness detection with citation is the correct feature because it directly addresses the compliance requirement for the model to explain its reasoning and cite sources. This feature, available in Azure AI Foundry, evaluates the model's responses against the provided grounding documents and automatically generates citations, ensuring that the output is factually supported and traceable.

Answer

Fine-tuning with domain-specific data

Answer

Content filtering

Answer

Prompt engineering

Question 15

Which TWO actions should you take to ensure high availability for an Azure AI service deployed in a single region?

Accepted Answer

Configure autoscaling for the AI service.. Option A (Configure autoscaling for the AI service) is correct because autoscaling dynamically adds or removes compute capacity in response to demand or load, preventing a single instance from becoming a bottleneck or failing under traffic spikes, which directly improves availability within the region. Option C (Deploy the AI service in a second Azure region) is correct because multi-region deployment provides geographic redundancy, so if the primary region suffers an outage, traffic can fail over to the secondary region and the service remains reachable. Option B is not appropriate because a read-only replica does not provide write/full-service availability and is not a standard high-availability mechanism for Azure AI services. Option D is not the right fit because Azure DDoS Protection mitigates volumetric network attacks but does not address compute failures or regional outages. Option E is not correct because Azure Backup protects data for recovery, not continuous service availability.

Answer

Use a read-only replica of the AI service.

Answer

Enable Azure DDoS Protection.

Answer

Enable Azure Backup for the AI resource.