AI Pilot Execution and Scaled Deployment CAIPM Practice Questions (Page 9)

Questions 41–45

1. 41application · medium

   An insurance company deployed an AI claims-processing model. The operations team notices that the model's processing time has increased significantly during peak hours, affecting service-level agreements. The model itself is accurate. What should the team do first?

   • ARetrain the model with more recent data to improve efficiency.
   • BScale the inference infrastructure to handle peak load, such as adding auto-scaling or additional resources.
   • CReduce the model's input features to speed up processing.
   • DLower the confidence threshold to reduce the number of predictions.
   Reveal Solution Discussion Select an answer first
2. 42expert · hard · select all that apply

   A company has deployed an AI system for dynamic pricing. The system has been in production for three months. The data science team wants to ensure the model remains accurate and aligned with business goals. Select all that apply. Which actions are appropriate for post-deployment monitoring and optimization?

   • AEstablish a cross-functional review team to evaluate model performance and business impact periodically
   • BFreeze the model after deployment to ensure stability
   • CRetrain the model only when a major business change occurs
   • DTrack model performance metrics such as prediction error and business metrics such as revenue
   • ESet up automated alerts for data drift and model drift
   Check answers Discussion Select an answer first
4. 43expert · hard

   A healthcare provider is scaling an AI diagnostic tool that was piloted in one clinic. The tool uses patient data and provides treatment recommendations. The provider must comply with HIPAA and wants to ensure patient safety. What is the most important governance requirement for scaled deployment?

   • ARely on the tool's CE mark as proof of safety for all populations
   • BOnly ensure that patient data is encrypted at rest and in transit
   • CDeploy the tool in all clinics immediately to maximize benefit
   • DEnsure the AI tool is validated for the new patient populations and that there is a human-in-the-loop for final decisions
   Reveal Solution Discussion Select an answer first
5. 44foundation · easy

   What is the primary purpose of deployment governance in AI projects?

   • ATo ensure the deployment complies with policies, regulations, and ethical standards
   • BTo reduce the cost of the AI solution
   • CTo speed up deployment by ignoring compliance
   • DTo make the AI model more complex
   Reveal Solution Discussion Select an answer first
6. 45application · medium

   A retail company is launching an AI pilot for demand forecasting in three pilot stores. The pilot sponsor wants to prove the model's value before committing to a full rollout. The project manager must define the pilot scope and success criteria. Which approach best aligns with the sponsor's intent?

   • ADeploy the model to all 200 stores to gather the broadest possible dataset for validation.
   • BFocus only on technical model performance, such as precision and recall, without linking to business metrics.
   • CDefine success as a 10% improvement in forecast accuracy across the three pilot stores over a 90-day period, with a clear comparison to baseline.
   • DRun the pilot for 12 months to capture seasonal variations, even if it delays the go/no-go decision.
   Reveal Solution Discussion Select an answer first