AI credit card cohort loss curve narrative

End-of-lesson check

15 questions · take it digitally for instant feedback at tendril.neural-forge.io/learn/quiz/end-finance-ai-credit-card-cohort-loss-curve-narrative-adults

1. Which of the following best describes what the AI surfaces for the credit committee?

   1. The credit score thresholds that should be adjusted immediately
   2. The exact dollar amount of expected losses for the quarter
   3. The segments (by product, channel, or risk tier) showing worse-than-expected loss rates
   4. The specific borrower names driving portfolio deterioration

2. What type of recommendations might AI provide to support credit committee discussions?

   1. Ultimate net loss predictions for the full life of the portfolio
   2. Replacement of the chief credit officer's role
   3. Credit policy levers worth discussing, such as tightening applicant criteria
   4. Final and binding credit policy changes

3. According to the capabilities and limitations of AI in this context, which action is beyond AI's capability?

   1. Approving credit policy changes based on loss curve analysis
   2. Identifying patterns of deterioration in early months
   3. Suggesting segments that are underperforming expectations
   4. Comparing current vintages to historical base case performance

4. Why is AI unable to predict ultimate net loss for a credit card vintage?

   1. Credit card portfolios do not generate enough transaction data
   2. The data is stored in incompatible formats
   3. Early-life vintage data is volatile and patterns can change dramatically as the cohort seasons
   4. The algorithm lacks sufficient computing power

5. What role does the chief credit officer play that AI cannot fulfill?

   1. Providing final judgment on risk appetite and policy decisions
   2. Running the loss curve calculations
   3. Generating the initial loss curve charts
   4. Entering transaction data into the system

6. Why should the credit committee be cautious when AI flags early deterioration in a new vintage?

   1. The data has likely been entered incorrectly
   2. The credit committee is not authorized to review AI outputs
   3. Early-life vintage data is inherently volatile and may not persist
   4. AI always provides incorrect information in the first months

7. What risk arises from relying too heavily on AI's pattern matching when analyzing new vintages?

   1. AI will delete the historical comparison data
   2. The AI might recommend loosening credit standards inappropriately
   3. AI pattern matching can over-react to volatile early-life data and flag false deterioration
   4. The system will automatically approve all applications

8. What context does the credit committee consider that AI lacks the ability to integrate?

   1. Individual borrower transaction histories
   2. Macro economic conditions and broader policy context
   3. The exact interest rate charged on each account
   4. The specific algorithm used for scoring

9. What is the primary function of the credit committee in this workflow?

   1. Training the AI model on new data
   2. Approving individual loan applications
   3. Reviewing AI-generated narratives and making policy decisions
   4. Calculating loss curves from raw data

10. What is the purpose of vintage analysis in credit card portfolio management?

    1. To set the interest rate for existing cardholders
    2. To determine the credit score of each new applicant
    3. To track the performance of loans originated in the same time period
    4. To calculate the total interest revenue for the quarter

11. In vintage analysis, what does the term 'seasoned base case' refer to?

    1. A mature vintage's expected loss trajectory based on historical performance
    2. A newly originated cohort with no history
    3. The most recent month's loss rate
    4. The oldest vintage in the portfolio

12. When drafting the AI-generated narrative, how should inferred causes of deterioration be marked?

    1. With a [hypothesis] marker indicating uncertainty
    2. In italics to emphasize importance
    3. With a confidence score of 100%
    4. As definitive facts without qualification

13. Which of the following is an example of a 'credit policy lever' that AI might suggest for committee discussion?

    1. Changing the company's logo
    2. Adjusting minimum credit score thresholds for approval
    3. Renaming the credit card product
    4. Hiring additional customer service staff

14. Why is the chief credit officer's judgment considered essential despite AI's analytical capabilities?

    1. The AI system requires manual approval for all outputs
    2. Regulations prohibit AI from operating without human oversight
    3. The AI has not been certified for production use
    4. The role involves integrating quantitative analysis with strategic risk appetite, macro context, and institutional knowledge

15. What is the main risk if the credit committee relies solely on AI output without adding macro or policy context?

    1. The committee will lose the ability to read loss curve charts
    2. The committee could make policy decisions that ignore economic conditions and strategic priorities
    3. The committee might approve loans for applicants below minimum thresholds
    4. The AI system will automatically implement any recommended changes