CU Underwriting Technology and Data Analytics Flashcards
6 cards from real CU practice questions. Tap to flip, then mark Knew It or Still Learning — missed cards come back until you master them.
Read the first 6 CU Underwriting Technology and Data Analytics flashcards as text
What is 'catastrophe modeling' and why is it important in underwriting?
Answer: A simulation technique that estimates potential losses from natural or man-made disasters to guide pricing and accumulation management
Cat models use scientific and statistical methods to project probable maximum losses from events like hurricanes or earthquakes, informing both underwriting decisions and reinsurance purchasing.
Which term describes using insurance data analytics to identify accounts that are likely to non-renew before they do so?
Answer: Lapse propensity modeling
Lapse propensity models analyze behavioral and policy data to flag insureds at high risk of canceling or not renewing, enabling proactive retention efforts.
What does 'algorithmic underwriting bias' refer to?
Answer: Unintentional discrimination embedded in automated underwriting models that leads to unfair treatment of protected classes
When models are trained on historical data that reflects past discriminatory practices, they can perpetuate those patterns, raising regulatory and ethical concerns.
An insurer deploying an AI-based underwriting model is required to ensure the model is 'explainable' primarily because:
Answer: Regulators and applicants may require justification for adverse underwriting decisions such as declinations or surcharges
Explainability allows insurers to provide regulators and consumers with specific, articulable reasons for coverage decisions, supporting compliance with adverse action notice requirements.
Real-time IoT (Internet of Things) sensor data from commercial buildings can help underwriters by:
Answer: Providing continuous monitoring of hazard conditions such as water leaks, fire suppression status, or electrical anomalies
IoT sensors enable underwriters and loss control professionals to monitor risk conditions in near real time, identifying hazards before they cause losses.
In data-driven underwriting, the term 'model drift' refers to:
Answer: The degradation of a predictive model's accuracy over time as real-world conditions change from those in the training data
Model drift occurs when the population or risk environment shifts away from the historical patterns the model was trained on, reducing its predictive power and requiring recalibration.