According to CoinMeta, AWS optimized personalized conversion through a context-based multi-arm slot machine (MAB), improving the efficiency of the product acquisition funnel. During a seven-week online A/B test, the final conversion rate for a certain customer group increased by several percentage points. AWS also shared how to apply AI personalization on Amazon SageMaker to overcome selection challenges and provided a code library for users to test. This system optimizes the entire conversion funnel using a linear UCB method, ensuring that each customer receives the best content recommendations.