AI-103 Practice Test 33 — 15 Questions

Question 1

You are configuring a function-calling agent in Microsoft Foundry. The agent must call an internal order-status REST API. Security review requires that the agent never receives the API key directly and that calls are auditable per user. You also need the agent to recover gracefully when the API returns a transient 503. (Choose two.)

Choose an answer to begin — your selection is scored in the full session.

Accepted Answer

Define the retry policy on the function tool so that transient failures such as HTTP 503 are retried with backoff before the result is returned to the model.. Keeping the credential in Key Vault and retrieving it through the agent's managed identity ensures the key never enters the model context, while vault access logging supports auditing. Configuring retries with backoff on the function tool handles transient 503 responses deterministically, preventing the model from re-invoking the function and creating duplicate calls. Together these satisfy the security and resilience requirements.

Answer

Disable function calling and instead have the agent generate SQL queries against the order database.

Answer

Embed the API key as a parameter default in the function tool definition so the model can pass it automatically.

Answer

Instruct the model in the system prompt to retry the function call up to three times whenever it receives an error.

Question 2

You are building a document processing pipeline that reads printed English text from scanned invoices. The invoices often contain noisy backgrounds and low-contrast text. You need to extract the text as accurately as possible using the Azure AI Vision Image Analysis 4.0 Read capability. Which approach should you take?

Accepted Answer

Preprocess the image with contrast enhancement and denoising, then call the Read operation and use the returned words and lines.. The Read capability extracts text from images but performs best when the input is clear. For scanned invoices with noise and low contrast, preprocessing steps such as contrast enhancement and denoising significantly improve recognition accuracy. After preprocessing, calling Read and using the returned words and lines provides the most accurate text extraction, making this the correct approach.

Answer

Call the Read operation on the raw image without preprocessing, then post-process the returned text lines using custom regex.

Answer

Enable the OCR feature with a custom model trained on invoice fonts using Azure AI Vision Studio.

Answer

Use the Dense Captions feature to generate descriptions of the invoice text, then parse the descriptions for key fields.

Question 3

A developer on your team is configuring an Azure AI Foundry agent that must answer questions about a product catalog stored in Azure Cosmos DB for NoSQL. The agent must retrieve the current price and stock for a given product SKU at query time. Which tool should the agent use?

Accepted Answer

A function tool that queries Cosmos DB for the SKU and returns the current price and stock as JSON.. A function tool is the correct choice because it allows the agent to execute authenticated code that queries Cosmos DB for the exact SKU and returns live price and stock values. Function calling is intended for deterministic, structured lookups where a vector index or static file would return stale data.

Answer

The code interpreter tool to run a Python script that connects to Cosmos DB and computes the price.

Answer

The file search tool with a CSV export of the catalog uploaded to the agent's vector store.

Answer

The Azure AI Search tool pointed at an index populated nightly from Cosmos DB.

Question 4

You are deploying a generative AI chat application on Azure OpenAI. Product owners want the model's responses to follow a strict JSON schema so the front end can render structured cards. You need the highest possible reliability that outputs conform to the schema. What should you configure?

Accepted Answer

Use structured outputs with a JSON schema definition on the model deployment.. Structured outputs enforce the JSON schema during decoding, so the model cannot emit tokens that violate the schema. This provides a deterministic contract that the front end can rely on, unlike temperature tuning, sampling adjustments, or prompt examples, all of which only increase the probability of conformance without guaranteeing it.

Answer

Add few-shot examples of well-formed JSON to every request payload.

Answer

Increase the max_tokens value and set top_p to 0.1 to narrow sampling.

Answer

Set the temperature parameter to 0 and include the schema in the system message.

Question 5

You are building a Python application that uses the Azure AI Vision Image Analysis 4.0 SDK to read text from scanned invoices. The invoices are stored in an Azure Blob Storage container, and each blob has a public URL. You need to extract all visible text from each invoice with a single REST API call, without downloading the image bytes to your application. Which API call should you make?

Accepted Answer

POST to the Image Analysis 4.0 endpoint with the URL parameter set to the blob URL and the features query parameter set to read. The read feature is the OCR capability of Image Analysis 4.0, and the url query parameter lets the service fetch the image directly from Blob Storage. Combining them satisfies both constraints: extracting all visible text and avoiding a local download. Caption, denseCaptions, and tags describe or classify the image rather than returning its textual content.

Answer

POST to the Image Analysis 4.0 endpoint with the URL parameter set to the blob URL and the features query parameter set to caption

Answer

POST to the Image Analysis 4.0 endpoint with the URL parameter set to the blob URL and the features query parameter set to denseCaptions

Answer

POST to the Image Analysis 4.0 endpoint with the URL parameter set to the blob URL and the features query parameter set to tags

Question 6

You are implementing a function-calling flow in an Azure AI Foundry agent. The agent must call a get_order_status function, but during testing it sometimes invents an order ID and calls the function with a fabricated value instead of asking the user. You need the agent to request the missing order ID from the user before invoking the function. What should you do?

Accepted Answer

Mark the order_id parameter as required in the function's JSON schema and provide a clear parameter description, then let the model request it when absent.. In function calling, the JSON schema of the tool defines what the model must supply. Declaring order_id as required with a descriptive field gives the model the signal it needs to ask the user for the missing value instead of inventing one, which is the correct fix for the observed behavior.

Answer

Set the agent's temperature to 0 so the model becomes deterministic and stops fabricating order IDs.

Answer

Wrap the function call in a try/catch that retries with a random order ID when the first call fails.

Answer

Add a system message instructing the model to never guess values, and remove the order_id parameter from the function schema.

Question 7

You are using Azure AI Vision Image Analysis 4.0 to analyze a batch of images. You need to extract both the caption and the detected objects with their bounding boxes. You are writing code that calls the Analyze Image API. Which two parameters must you include in your request? (Choose two.)

Accepted Answer

visualFeatures=caption. To obtain both captions and object bounding boxes from a single Analyze Image call, the visualFeatures query parameter must include both caption and objects. These features are requested together, and the response will contain the corresponding sections. Other parameters like details, language, or model-version are optional and do not enable these specific outputs.

Answer

details=landmarks

Answer

language=en

Answer

model-version=latest

Question 8

You are developing an application that uses Azure OpenAI. You need to ensure that the application can authenticate to the Azure OpenAI service without storing credentials in code. What should you use?

Accepted Answer

Azure Active Directory (Azure AD) token-based authentication using managed identities.. Microsoft Entra ID authentication with managed identities allows your application to authenticate to Azure OpenAI without any stored credentials. The managed identity is assigned to the application's hosting resource, and Microsoft Entra ID issues tokens that the application uses to access the service. This eliminates secret management and is the recommended practice.

Answer

Shared Access Signatures (SAS) generated for the Azure OpenAI resource.

Answer

Azure OpenAI API keys stored in Azure Key Vault and retrieved at runtime.

Answer

OAuth 2.0 client credentials flow with a client secret stored in Azure App Configuration.

Question 9

You are deploying a generative AI chat application built on Azure OpenAI in Azure AI Foundry. The application must stream tokens to the browser, and you need to configure it so that responses begin appearing to users as quickly as possible while remaining correct. Which two actions should you take? (Choose two.)

Accepted Answer

Set stream to true on the chat completions request so tokens are delivered incrementally as they are generated.. Streaming delivers tokens incrementally so users see output immediately, and choosing a lower-latency model such as GPT-4o mini reduces the time to first token. Together these actions improve perceived responsiveness without compromising correctness, whereas token limits, temperature, and content filtering do not affect latency in the desired way.

Answer

Enable content filtering at the highest severity threshold to reduce the tokens the model must generate.

Answer

Increase the max_tokens value to the model's maximum so the response can be as long as possible.

Answer

Set temperature to 1.5 so the model explores more possibilities and returns answers faster.

Question 10

You are developing a web application that allows users to upload images and receive a list of detected objects with their bounding boxes. You want to use the Azure AI Vision Image Analysis 4.0 SDK for Python. Which method should you call to analyze an image from a URL and retrieve object detection results?

Accepted Answer

analyze_image. The analyze_image method is the primary entry point for image analysis in the Azure AI Vision SDK. By specifying visual features such as objects, it returns detected objects with their bounding boxes and labels. Other methods like generate_thumbnail, get_area_of_interest, and tag_image serve different purposes and do not provide object detection with coordinates.

Answer

get_area_of_interest

Answer

tag_image

Answer

generate_thumbnail

Question 11

A financial services company uses a custom neural model in Azure AI Document Intelligence to extract fields from loan applications. After retraining the model with 30 additional labeled documents, field accuracy for the 'loanAmount' field dropped sharply, although the field was previously accurate. The new documents were labeled by a different team. You need to identify the most likely cause of the regression before retraining again. What should you investigate first?

Accepted Answer

Whether the labeled documents use inconsistent field labels or value formatting for 'loanAmount' compared with the original training set. Custom neural model quality depends heavily on consistent, accurate labels. When multiple teams label the same dataset, differences in field naming, value formatting, or region placement introduce conflicting training signals that disproportionately harm the affected field. Auditing label consistency for 'loanAmount' across both datasets is the correct first step before retraining.

Answer

Whether the service endpoint region was changed between trainings

Answer

Whether the model's confidence threshold was raised in the calling application

Answer

Whether the training dataset exceeded the maximum page count allowed for custom neural models

Question 12

A retail analytics team wants to count how many people enter a store each hour. They plan to analyze frames from an existing ceiling-mounted camera feed using Azure AI Vision Image Analysis 4.0. They need the count of detected people and the bounding box for each person, without storing the images long term. Which approach should they use?

Accepted Answer

Call the Image Analysis 4.0 analyze endpoint with the features parameter set to people, and read the peopleResult object from the response. The people feature is purpose-built for detecting humans and returns both a count and bounding boxes in peopleResult. That maps exactly to the hourly footfall metric and the per-person coordinates. The other features describe objects, tags, or suggested crops and do not provide dependable person-level detection or counting.

Answer

Call the Image Analysis 4.0 analyze endpoint with the features parameter set to tags, and count tags named person

Answer

Call the Image Analysis 4.0 analyze endpoint with the features parameter set to smartCrops, and use the crop regions as person locations

Answer

Call the Image Analysis 4.0 analyze endpoint with the features parameter set to objects, and filter the results for the person tag

Question 13

You are deploying an Azure AI Document Intelligence custom extraction model for a regulated workload. Documents are uploaded to Azure Blob Storage in a private container, and no public network access is allowed. Your application runs in an Azure virtual network. You must ensure the Document Intelligence resource can access the documents for training and analysis without exposing them publicly. Which configuration should you implement?

Accepted Answer

Configure the Document Intelligence resource with a private endpoint and grant it access to the storage account through a managed identity with the Storage Blob Data Reader role. For private, no-public-access scenarios, the supported pattern is a private endpoint for the Document Intelligence resource plus a managed identity granted Storage Blob Data Reader on the storage account. This keeps traffic on the virtual network and avoids shared keys or SAS tokens. IP firewalls and SAS tokens do not satisfy the no-public-access constraint.

Answer

Copy the documents into the Document Intelligence resource's own storage and reference them by internal path

Answer

Generate a long-lived SAS token for the container and embed it in the training request

Answer

Enable public network access on the storage account but restrict it by IP firewall to the Document Intelligence outbound IP range

Question 14

You are planning to deploy an Azure AI solution that includes multiple Azure Cognitive Services resources. You need to ensure that the resources can be managed together and that costs are consolidated. What should you use?

Accepted Answer

Use an Azure Cognitive Services multi-service resource.. An Azure Cognitive Services multi-service resource provides a single Azure resource that can be used to access multiple cognitive services, such as Computer Vision, Text Analytics, and Speech. It simplifies management, offers a single key and endpoint, and consolidates billing for those services. This is the recommended approach when you need to use several AI services together.

Answer

Deploy each resource in a separate subscription and use Azure Cost Management to aggregate costs.

Answer

Configure each resource to send diagnostic logs to a central Log Analytics workspace.

Answer

Create each resource in the same resource group and subscription.

Question 15

A media monitoring company subscribes to the Azure AI Language service and needs to extract named entities from news articles, then link each entity to its corresponding Wikipedia article when one exists. Which capability should you use to return both the entity text and its linked URL in a single call?

Accepted Answer

Entity linking. Entity linking is the capability that both detects entities and resolves them to canonical entries with associated URLs, such as Wikipedia pages. Named entity recognition and custom NER provide types and spans without links, and key phrase extraction returns salient terms with no knowledge base association, so entity linking is the only fit for a single call returning text plus URL.

Answer

Key phrase extraction

Answer

Named entity recognition (NER) alone

Answer

Custom named entity recognition (Custom NER)