AI in Retail Quiz

Answers and explanations appear as you take the quiz.

1. Why define data available at the decision moment for a retail model?
   • Recommendations, pricing, forecasts, and fraud reviews have different inputs and error costs
   • Every decision can use later purchase data
   • Data timing never matters
   • All retail models optimize the same outcome
2. A recommender increases basket size but customers complain about irrelevant offers. What should be reviewed?
   • Only the basket-size increase
   • Customer outcomes and whether the proxy metric harms service
   • The color of the recommendation card
   • The number of products in inventory
3. What did the false-decline worked example ask retailers to measure?
   • The average basket size
   • Only the number of blocked fraud attempts
   • Legitimate customers declined, correction time, and safer verification options
   • The model’s response time
4. Why evaluate retail outcomes by subgroup?
   • It removes the need to monitor returns
   • All customers have identical histories
   • Segmentation guarantees better revenue
   • Averages can hide less useful service or unfair treatment for some groups
5. Which outcomes should accompany a recommendation model’s sales results?
   • Only click-through
   • Useful discovery, returns, complaints, and stock availability
   • Only the count of messages
   • The number of model parameters
6. Which retail information does the guide identify as needing protection?
   • Only store hours
   • Only public product names
   • Purchase history, location, and identity information
   • The model’s display color
7. What should an automated system provide when it cannot answer or a transaction is wrong?
   • An alternative route and a way for staff to correct the issue
   • A higher price
   • A hidden score only
   • No explanation or recourse
8. Why monitor new products, promotions, and seasons?
   • They prevent the need for versioning
   • They make all recommendations more accurate
   • They eliminate false declines
   • They can change behavior and cause model or policy drift