Integrating AI into Enterprise Systems CAIPM Practice Questions (Page 4)

Questions 16–20

1. 16application · medium

   A government agency is integrating an AI system to automate the processing of citizen service requests. The existing system is a legacy mainframe application that stores data in a hierarchical database. The agency wants to minimize disruption to ongoing operations. What should the integration approach be?

   • AReplace the legacy mainframe with a new system that has built-in AI capabilities.
   • BBuild an API wrapper around the legacy mainframe to expose its data and functions to the AI system.
   • CMigrate all data from the legacy database to a modern relational database before integrating the AI system.
   • DRun the AI system in parallel with the legacy system and manually transfer data between them.
   Reveal Solution Discussion Select an answer first
2. 17foundation · easy

   Which activity is a key component of a change management plan for AI adoption?

   • ALimiting communication about the change to top management only
   • BProviding comprehensive user training and support
   • CImplementing the AI system without any pilot testing
   • DDisabling all user feedback mechanisms
   Reveal Solution Discussion Select an answer first
4. 18expert · hard

   A financial services firm is integrating an AI system for algorithmic trading. The system must comply with financial regulations that require audit trails for all trading decisions and the ability to explain any AI-generated trade. The firm also wants to minimize latency to remain competitive. What should the governance and compliance framework include?

   • AStore audit logs in a separate system that is only accessible after a regulatory request.
   • BDisable logging to reduce latency and rely on periodic manual reviews of trading activity.
   • CLog all AI-generated trading decisions with the input data and model version, and implement a real-time monitoring system to flag unusual trades.
   • DUse a black-box model that cannot be explained, but ensure it is highly profitable.
   Reveal Solution Discussion Select an answer first
5. 19application · medium

   A media company is integrating an AI content-tagging service into its content management system (CMS). The CMS is a monolithic application that does not expose APIs. The AI service is a separate microservice. What architectural pattern should be used to connect them?

   • AReplace the CMS with a new system that has built-in AI integration.
   • BUse a message queue to asynchronously send content to the AI service and receive tags.
   • CHave the AI service poll the CMS database directly for new content.
   • DDirectly modify the CMS code to call the AI service's API.
   Reveal Solution Discussion Select an answer first
6. 20application · medium

   A financial services firm wants to integrate AI into its loan approval process. The firm's IT architecture is highly regulated, and the AI model must be explainable to auditors. The program manager needs to define an integration strategy. What should be the first step?

   • AImmediately begin a pilot project in a sandbox environment.
   • BAssess the current loan approval process and identify where AI can add value while meeting regulatory requirements.
   • CSelect the most advanced AI model available and plan its integration.
   • DHire a data science team to build the model independently of the IT department.
   Reveal Solution Discussion Select an answer first