An organization is deploying an AI agent that uses a custom model to detect fraudulent transactions. The agent must be tested for reliability and performance. The team has a limited testing window and must decide what to test first. Which testing activity should be prioritized?

Test the agent's integration with Dynamics 365 Finance to ensure data flows correctly.

Validate the custom model's precision and recall on a recent, representative dataset.

Conduct user acceptance testing with a few business users to get feedback.

Run performance tests to ensure the agent can handle peak transaction volumes.

A company is deploying an AI solution that integrates Dynamics 365 Sales, Marketing, and Customer Service. The solution uses an agent to nurture leads and escalate issues. The testing team needs to design test scenarios that cover the entire customer journey. Which approach is most effective?

Use a single test scenario that covers only the most common path.

Focus on the marketing application only, since it is the entry point.

Test each application's agent functionality independently to isolate issues.

Create end-to-end scenarios that simulate a lead from marketing through sales to a support case, verifying data consistency at each stage.

When developing a strategic approach for creating test cases using Copilot, which step is most important?

Providing clear and detailed requirements to Copilot

Generating test cases only for happy path scenarios

Using Copilot to generate test cases without any human input

Avoiding the use of examples in prompts to Copilot

A team is evaluating an AI agent that automates invoice processing in Dynamics 365 Finance. They want to measure how often the agent correctly extracts data and completes the process without human intervention. Which combination of metrics should they track?

Model training time, dataset size, and number of features.

Number of API calls, server uptime, and database latency.

Accuracy, task completion rate, and user satisfaction score.

Number of user logins, session duration, and page views.

A data science team has developed a custom AI model to detect fraudulent transactions in Dynamics 365 Finance. The model has high precision but low recall. The team is concerned about missing fraudulent transactions. What validation criterion should they focus on?

Accuracy, because it measures the overall correctness of the model.

F1 score, because it balances precision and recall.

Precision, because it measures the proportion of flagged cases that are truly fraudulent.

Recall, because it measures the proportion of actual fraud cases that the model correctly identifies.

Microsoft Certified:Agentic AI Business Solutions Architect

Domain 3Objective 2

Manage the Testing of AI-powered Business Solutions AB-100 Practice Questions (Page 3)

Part of the Deploy AI-powered business solutions domain, which accounts for 40–45% of the AB-100 exam. Microsoft does not publish an official question count, but from its 100-minute exam (~40–65 total, ~16–29 in this domain), expect 4–7 from this objective — we provide 44 practice questions to prepare you well beyond it. (estimate)

44questions here

9free pages

6concepts

40–45%of the exam

Questions 11–15

11expert · hard
An organization is deploying an AI agent that uses a custom model to detect fraudulent transactions. The agent must be tested for reliability and performance. The team has a limited testing window and must decide what to test first. Which testing activity should be prioritized?
Select an answer first
12application · medium
A company is deploying an AI solution that integrates Dynamics 365 Sales, Marketing, and Customer Service. The solution uses an agent to nurture leads and escalate issues. The testing team needs to design test scenarios that cover the entire customer journey. Which approach is most effective?
Select an answer first
13foundation · easy
When developing a strategic approach for creating test cases using Copilot, which step is most important?
Select an answer first
14application · medium
A team is evaluating an AI agent that automates invoice processing in Dynamics 365 Finance. They want to measure how often the agent correctly extracts data and completes the process without human intervention. Which combination of metrics should they track?
Select an answer first
15expert · hard
A data science team has developed a custom AI model to detect fraudulent transactions in Dynamics 365 Finance. The model has high precision but low recall. The team is concerned about missing fraudulent transactions. What validation criterion should they focus on?
Select an answer first

Finished these 5 questions?

Review the revealed explanations, or continue through the curriculum.

Next 5 questions →Back to AB-100

Free Basic Practice is a study aid with revealable answers — not a scored exam. Examers.io is independent and not affiliated with or endorsed by Microsoft. “AB-100” is a trademark of its owner, used for identification only.