Manage the Testing of AI-powered Business Solutions AB-100 Practice Questions (Page 4)

Questions 16–20

1. 16foundation · easy

   When crafting a prompt for Copilot, which element is most likely to improve the quality of the response?

   • AKeeping the prompt as short as possible
   • BAsking multiple questions in a single prompt
   • CUsing technical jargon to demonstrate expertise
   • DIncluding an example of the desired output
   Reveal Solution Discussion Select an answer first
2. 17application · medium

   A company is integrating an AI agent with Dynamics 365 Marketing and Dynamics 365 Customer Insights to personalize email campaigns. The agent must segment customers, generate content, and track engagement. What should the end-to-end test scenarios include?

   • ATesting only the email content generation in Marketing, without checking data sources.
   • BTesting only the engagement tracking reports, assuming the earlier steps work.
   • CTesting only the segmentation logic in Customer Insights, without verifying the email send.
   • DVerifying that customer data flows correctly from Customer Insights to Marketing and that the agent uses the right segments.
   Reveal Solution Discussion Select an answer first
4. 18foundation · easy

   Which validation criterion is most important for ensuring a custom AI model performs well in real-world business scenarios?

   • ARelevance to the business use case
   • BNumber of training epochs
   • CPopularity of the algorithm used
   • DComplexity of the model architecture
   Reveal Solution Discussion Select an answer first
5. 19application · medium

   A team is developing an AI agent that predicts inventory shortages in Dynamics 365 Supply Chain Management. They need to validate the custom model's predictions before deployment. Which validation approach should they use?

   • ACheck that the model's predictions match the training data exactly.
   • BCompare the model's predictions against historical outcomes on a validation dataset.
   • CAsk business users to review the model's predictions for face validity.
   • DDeploy the model and monitor its performance in production.
   Reveal Solution Discussion Select an answer first
6. 20application · medium

   A QA lead wants to use Copilot to generate test cases for an AI agent that handles employee onboarding in Dynamics 365 Human Resources. The agent must collect documents, verify eligibility, and create user accounts. What is the best way to use Copilot for test case creation?

   • AHave Copilot create test cases from production data without any human review.
   • BUse Copilot to run the test cases automatically and report failures.
   • CAsk Copilot to generate test cases based only on the agent's name and purpose.
   • DDescribe the onboarding process, including edge cases like missing documents or duplicate accounts, and ask Copilot to generate test cases.
   Reveal Solution Discussion Select an answer first