Core AI Concepts and Business Applications CAIPM Practice Questions (Page 5)

Questions 21–25

1. 21expert · hard

   An insurance company wants to use AI to automate claims processing. They have two candidate applications: a natural language processing system that extracts information from claim forms, and a machine learning model that predicts the likelihood of fraud. The company wants to reduce processing time and detect fraudulent claims. They have a limited IT budget and must choose one application to start with. Which application should they prioritize?

   • AStart with the NLP system to automate data extraction, which reduces processing time and provides structured data for future fraud detection.
   • BStart with the fraud detection model because it directly addresses financial losses.
   • CImplement both simultaneously to achieve both goals at once.
   • DStart with neither because AI is not mature enough for insurance claims.
   Reveal Solution Discussion Select an answer first
2. 22application · medium

   A media company wants to automatically generate short news summaries from long articles. The system should understand the content and produce coherent summaries. Which AI approach is most appropriate?

   • AA rule-based system that extracts sentences containing keywords.
   • BComputer vision to analyze the layout of the articles.
   • CNatural language processing using a transformer-based model trained on text summarization tasks.
   • DRobotic process automation to copy and paste the first paragraph of each article.
   Reveal Solution Discussion Select an answer first
4. 23expert · hard

   An e-commerce company wants to use AI to personalize product recommendations. They have a large dataset of user behavior, but the data contains many missing values and inconsistencies. The company wants to launch the AI feature within three months. The data science team is small and has limited experience with complex models. What is the most realistic approach to meet the launch deadline?

   • AImmediately build a deep learning model and rely on it to handle missing data automatically.
   • BSpend the first two months cleaning and enriching the data, then build a simple recommendation model in the remaining month.
   • CLaunch the AI feature with the current data without any cleaning to save time.
   • DHire a large team of data scientists to build a complex model from scratch.
   Reveal Solution Discussion Select an answer first
5. 24expert · hard

   A bank wants to use AI to detect money laundering. They have two candidate solutions: a machine learning model that flags unusual transaction patterns, and a rule-based system that flags transactions above a certain threshold. The compliance team requires that all flagged transactions be explainable to regulators. The bank also wants to reduce false positives. Which approach best meets both requirements?

   • AUse both systems in parallel and manually review all flags from both.
   • BUse a machine learning model with explainability techniques (e.g., SHAP) to provide reasons for each flag.
   • CUse a deep learning model without explainability because it has the highest accuracy.
   • DUse the rule-based system because it is inherently explainable, even if it has more false positives.
   Reveal Solution Discussion Select an answer first
6. 25application · medium

   A logistics company is considering an AI system to optimize delivery routes. The operations manager wants to reduce fuel costs and delivery times, but the company's data is stored across multiple legacy systems with inconsistent address formats. The IT team has limited experience with data science. What should the company do first to maximize the chance of a successful AI initiative?

   • AHire a team of data scientists to build a custom deep learning model from scratch.
   • BClean and standardize the address data, then start with a pilot project on a single delivery region.
   • CReplace all legacy systems with a modern ERP to ensure data consistency before any AI project.
   • DPurchase a commercial route-optimization AI tool and integrate it directly with the existing systems.
   Reveal Solution Discussion Select an answer first