Quiz about Certification Exam, Microsoft, AI-901: Microsoft Azure AI Fundamentals (Updated Version)

Question 1

You have a Microsoft Foundry project that contains an agent named Agent1.
You need to ensure that Agent1 always calls an Azure function when the agent responds to
user
input.
To what should you set tool_choice for Agent1?
OPTIONS :

Accepted Answer

auto

Answer

none

Answer

required Answer : C EXPLANATION : Microsoft’s Foundry Agent Service documentation states that tool_choice provides deterministic control over tool calling: auto means the model decides whether to call tools. required means the model must call one or more tools. none means the model does not call tools. Therefore: A . auto = Incorrect, because the model may or may not call the Azure function. B . none = Incorrect, because this prevents tool/function calls. C . required = Correct, because it forces the agent to call a tool. The Azure OpenAI function-calling documentation also confirms that tool_choice=\"auto" lets the model decide whether to call a function, while tool_choice="none" forces a user-facing response without a tool call.

Question 2

DRAG DROP
You have a Microsoft Foundry project named project1 that contains an Azure OpenAI
resource named
Resource1.
To Resource1, you deploy a gpt-4.1-mini model by using a model deployment named my-
mini-gpt.
You need to connect to my-mini-gpt from an application.
How should you complete the Python code? To answer, drag the appropriate values to the
correct
targets. Each value may be used once, more than once, or not at all.

NOTE: Each correct selection is worth one point.

$\"Certification$

OPTIONS :

Accepted Answer

See the Explanation. Answer : A EXPLANATION : client = OpenAI( api_key=\"...", base_url="[https://resource1.openai.azure.com/openai/v1/",](https://resource1.openai.azure.com/openai/v1/",) ) response = client.responses.create( model="my-mini-gpt", ... ) For Azure OpenAI in Microsoft Foundry, the base_url uses the Azure OpenAI resource name in the endpoint format: In the question, the Azure OpenAI resource is named Resource1, so the first blank must be resource1. Microsoft documentation for Azure OpenAI v1 endpoints confirms that the endpoint must use the ...openai.azure.com/openai/v1/ path. For the model parameter, Azure OpenAI requires the deployment name, not the underlying model name. Microsoft states that Azure OpenAI always requires the deployment name when calling APIs, even when the parameter is named model. The deployed model is gpt-4.1-mini, but the deployment name is my-mini-gpt. Therefore, the second blank must be: model="my-mini-gpt" So the correct selections are: base_url blank = resource1 model blank = my-mini-gpt

Question 3

What are two purposes of instructions when prompting a generative AI model? Each correct
answer
presents part of the solution.
NOTE: Each correct selection is worth one point.
OPTIONS :

Accepted Answer

defines constraints on the model's responses

Answer

defines the agent's role and behavior

Answer

defines the Azure region where inference occurs

Answer

selects which model to use

Answer

defines the tokens per minute (TPM) allocation for the model Answer : A,B EXPLANATION : Microsoft Foundry Agent Service documentation states that instructions define goals, constraints, and behavior for an agent. Therefore, instructions are used to guide how the generative AI model or agent should respond and behave. Option A is correct because instructions can define constraints the model must follow. Option B is correct because instructions can define the agent’s role and behavior. Options C, D, and E are incorrect because Azure region, model selection, and TPM allocation are configuration or deployment/resource settings, not purposes of prompt instructions.

Question 4

You are developing an application that analyzes voicemail recordings by using Azure
Content
Understanding in Foundry Tools.
You need to extract a transcript and structured information from the recordings.
Which type of analyzer should you use?
OPTIONS :

Accepted Answer

document analyzer

Answer

video analyzer

Answer

audio analyzer

Answer

image analyzer Answer : C EXPLANATION : Voicemail recordings are audio content. Azure Content Understanding analyzers define what type of content to process, including documents, images, audio, or video, and what elements to extract, including transcripts and structured fields. Microsoft’s custom analyzer documentation also shows an audio example based on prebuilt- audio for processing customer support call recordings, which is the same content type as voicemail recordings. Therefore, to extract a transcript and structured information from voicemail recordings, you should use an audio analyzer.

Question 5

Your organization is developing an AI solution to screen job applications automatically. You
want to ensure
the model makes fair hiring decisions across all demographic groups. Which of the
following actions would
best help you address fairness concerns in this AI system?
OPTIONS :

Accepted Answer

Train the model on historical hiring data without any modifications to ensure it learns real- world patterns

Answer

Audit the training data for demographic bias, test model predictions across different demographic groups, and establish performance metrics that measure fairness outcomes

Answer

Remove all personally identifiable information from the model to prevent any discrimination

Answer

Use the largest possible dataset available, as larger datasets automatically eliminate bias

Answer

Deploy the model to a limited user group first and monitor only the overall accuracy metric Answer : B EXPLANATION : The correct approach to addressing fairness involves multiple steps: auditing training data for potential bias, testing model performance across demographic groups to identify disparate impacts, and establishing fairness- specific performance metrics—not just overall accuracy. This aligns with responsible AI principles for fairness in AI solutions. Why other options are incorrect: • Training on unmodified historical data often perpetuates existing societal biases • Removing only PII does not address algorithmic bias in decision-making patterns • Large datasets do not automatically eliminate bias; they can amplify it if biased patterns exist • Monitoring only overall accuracy misses fairness issues that may disproportionately affect specific groups

Question 6

Your organization is developing an AI solution that analyzes customer support tickets to
automatically
categorize them by urgency and assign them to the appropriate team. The solution uses
historical ticket data
for training. You discover that the model performs significantly worse on tickets submitted
by customers
from certain geographic regions. What principle of responsible AI is most directly addressed
by investigating
and correcting this issue?
OPTIONS :

Accepted Answer

Fairness

Answer

Transparency

Answer

Accountability

Answer

Inclusiveness

Answer

Privacy Answer : A EXPLANATION : The scenario describes a model that performs unequally across different customer groups based on geographic region. This is a classic fairness issue—ensuring that AI systems make equitable decisions across all demographic and geographic groups. Fairness in AI specifically addresses algorithmic bias and unequal performance across different population segments. Why the other options are incorrect: Transparency refers to explainability and understanding how the model makes decisions; Accountability is about responsibility for outcomes; Inclusiveness addresses designing AI that serves all users including those with disabilities; Privacy concerns data protection and confidentiality.

Question 7

A healthcare AI system deployed in Azure uses patient medical imaging to detect potential
anomalies. The
model must operate reliably across different imaging equipment manufacturers, patient
populations, and
clinical settings. Which considerations are most critical to ensure the solution meets
reliability and safety
requirements? (Select all that apply.)
OPTIONS :

Accepted Answer

Testing the model's performance across diverse imaging equipment, patient demographics, and clinical environments

Answer

Encrypting all patient medical data both in transit and at rest

Answer

Establishing error thresholds and fallback procedures when the model's confidence is low

Answer

Obtaining explicit consent from patients before the system analyzes their images

Answer

Providing detailed explanations of how the model identified each anomaly to clinicians Answer : A,C EXPLANATION : Reliability and safety in AI systems require rigorous testing across diverse real-world conditions (different equipment, demographics, environments) and establishing appropriate confidence thresholds with documented fallback procedures when the model is uncertain. These measures ensure the system performs dependably and safely. Why the other options are incorrect: Encrypting data is a privacy/security measure, not a reliability consideration; Obtaining consent addresses privacy; Providing explanations to clinicians relates to transparency, not safety. While these are all important responsible AI principles, they don't specifically address reliability and safety.

Question 8

You are building a generative AI customer service chatbot using Azure AI Foundry. You
want the model to
respond politely but concisely, prioritize resolving common billing issues, and never suggest
refunds that
exceed company policy. What are the most effective ways to enforce these behaviors?
(Select all that apply.)
OPTIONS :

Accepted Answer

Include detailed instructions in the system prompt that define the desired tone, priorities, and policy constraints

Answer

Rely exclusively on the model's base training to understand company policies and billing procedures

Answer

Use temperature and top-p parameters to control the randomness and diversity of responses

Answer

Post-process model outputs to validate that responses comply with policy before sending to customers

Answer

Implement a separate legal review step that approves every response before it reaches users Answer : A,C,D EXPLANATION : Effective prompt engineering for generative AI involves clear system prompts (to define behavior and constraints), careful tuning of model parameters like temperature and top-p (to control randomness), and validation or filtering of outputs to ensure policy compliance. These are practical, scalable approaches used in production systems. Why the other options are incorrect: Relying solely on base training is insufficient because models don't reliably understand specific company policies without explicit guidance; Requiring legal review of every response is impractical and not a standard implementation pattern for customer service chatbots.

Question 9

Your team is implementing an AI solution that extracts structured data from handwritten
medical forms using
Azure AI Foundry's Content Understanding capabilities. The solution must handle forms in
multiple
languages, varying handwriting styles, and poor image quality from photocopies. Which
approach best
addresses these challenges?
OPTIONS :

Accepted Answer

Use Content Understanding's multimodal capabilities to process the image and apply post-processing logic to handle language variants and image quality issues

Answer

Deploy only optical character recognition (OCR) because handwritten forms require language-specific models

Answer

Manually transcribe all forms to ensure perfect accuracy before using the extraction system

Answer

Train a custom machine learning model from scratch on your specific form types

Answer

Require all submitted forms to be rewritten using typed text instead of handwriting Answer : A EXPLANATION : Azure AI Foundry's Content Understanding is designed to handle complex, real-world documents including handwritten content, multiple languages, and poor image quality. It uses multimodal AI capabilities that can interpret visual input intelligently, then post-processing logic can address specific edge cases. This is the most practical, scalable solution for this scenario. Why the other options are incorrect: OCR alone cannot reliably extract from handwritten forms; Manual transcription is not scalable; Training a custom model from scratch requires significant data and effort when a pre-built solution exists; Requiring rewritten forms eliminates the purpose of automating form processing.

Question 10

An organization is implementing a speech-enabled AI application using Azure AI Foundry
that allows
customers to report issues verbally. The application must: convert spoken words to text,
understand
customer sentiment, extract key details from the speech, and generate an appropriate
response. Which AI
workloads and capabilities should the solution combine to meet these requirements?
OPTIONS :

Accepted Answer

Speech recognition, sentiment analysis, information extraction, and text generation capabilities

Answer

Computer vision and image generation

Answer

Only speech synthesis and natural language understanding

Answer

Generative AI models alone without specific NLP workloads

Answer

Keyword extraction and entity detection only Answer : A EXPLANATION : The scenario requires converting speech to text (speech recognition), understanding emotional tone (sentiment analysis), identifying important information (information extraction), and generating an appropriate response (text generation/generative AI). Speech recognition converts audio to text, sentiment analysis interprets emotion, information extraction identifies key entities and details, and text generation creates the response. Why the other options are incorrect: Computer vision and image generation are for visual tasks, not speech; Speech synthesis alone generates audio but doesn't understand spoken input; Relying on generative AI alone without specific NLP workloads omits the crucial steps of speech recognition and sentiment analysis; Keyword extraction and entity detection are components of information extraction but don't cover the full requirement set.

Quiz AI-901: Microsoft Azure AI Fundamentals (Updated Version)

Free Test - Simulator AI-901: Microsoft Azure AI Fundamentals (Updated Version)

AI-901: Microsoft Azure AI Fundamentals (Updated Version) Practice test unlocks all online simulator questions

What to expect from our AI-901: Microsoft Azure AI Fundamentals (Updated Version) practice tests and how to prepare for any exam?