Analyze, Monitor, and Tune AI-powered Business Solutions AB-100 Practice Questions (Page 3)

Questions 11–15

1. 11expert · hard

   A global e-commerce company's AI customer-service agent is deployed across multiple Azure regions. The team needs to monitor the agent's performance and set up alerts for anomalies, but they must minimize cross-region data transfer to comply with data-residency policies. What monitoring architecture should they recommend?

   • AUse a single Azure Monitor workspace in the headquarters region for all telemetry.
   • BUse Azure Application Insights only in the largest region and ignore others.
   • CExport all telemetry to a third-party monitoring tool hosted outside Azure.
   • DDeploy Azure Monitor in each region, collect telemetry locally, and aggregate only summary metrics to a central Log Analytics workspace.
   Reveal Solution Discussion Select an answer first
2. 12foundation · easy

   What is a primary purpose of analyzing user feedback for an AI-powered solution?

   • ATo determine the exact cause of every technical error
   • BTo automatically fix all reported issues
   • CTo replace the need for telemetry data
   • DTo identify specific features users want added
   Reveal Solution Discussion Select an answer first
4. 13foundation · easy

   A team collects user feedback on an AI-powered chatbot. Which approach is most effective for identifying areas for improvement from this feedback?

   • AIgnore negative feedback to focus on positive comments
   • BCategorize feedback into themes such as accuracy, usability, and response time
   • COnly consider feedback from internal testers
   • DRead each comment individually without grouping
   Reveal Solution Discussion Select an answer first
5. 14expert · hard

   A financial advisory firm's AI investment assistant provides portfolio recommendations. The firm wants to increase user trust and reduce the number of users who abandon the assistant after receiving a recommendation. Telemetry shows that users often leave when the assistant expresses low confidence in its recommendation. What tuning technique should they apply?

   • AHide the confidence score from users to prevent them from leaving.
   • BIncrease the confidence threshold so the assistant only makes high-confidence recommendations.
   • CRetrain the model to always output high confidence scores.
   • DAdjust the assistant's responses to provide more context and explain the reasoning behind low-confidence recommendations.
   Reveal Solution Discussion Select an answer first
6. 15expert · hard

   An AI agent that handles product recommendations receives feedback that it 'always recommends the same products' and 'doesn't consider my past purchases'. The backlog shows that users who provide this feedback have a higher churn rate. The team wants to improve personalization. What should they do?

   • AAdd more products to the recommendation catalog.
   • BAnalyze the feedback and backlog to identify user segments that are underserved, then tune the recommendation logic to incorporate purchase history.
   • CRetrain the model on all user feedback to improve overall satisfaction.
   • DIncrease the temperature of the model to generate more diverse recommendations.
   Reveal Solution Discussion Select an answer first