Amazon Payments Uses SageMaker AI Contextual Bandit for Funnel Personalization, Reports High-Single-Digit Lift
AWS says Amazon Payments used a multi-objective contextual bandit on Amazon SageMaker AI to personalize an acquisition funnel, achieving a high single-digit conversion lift for one audience. The case study also identifies content—not the model—as the constraint.

Full Report
AWS has described how Amazon Payments used a multi-objective contextual bandit on Amazon SageMaker AI to personalize content in an acquisition funnel. The post notes that generative AI has made it inexpensive to produce personalized content at scale, while leaving teams with the task of choosing which variation to show each customer.
The approach delivered a high single-digit conversion lift for one audience. AWS also says the work revealed that content, rather than the model, was the constraint. The result points to both the selection method and the range of available content as relevant parts of a personalization effort.
Why This Matters
The case pairs a reported result with a specific scope: the high-single-digit lift applied to one audience, and AWS identifies content as the limiting factor. For teams evaluating personalization, that makes the available creative variations a practical consideration alongside the model.
Evidence
Start with the primary evidence, then review supporting sources and context.
Primary Evidence
1Uplifting conversion across the acquisition funnel with personalization using contextual bandits on AWS
Chidi Prince John
“Generative AI makes it cheap to produce personalized content at scale, but which variation do you show each customer? Amazon Payments used a multi-objective contextual bandit on Amazon SageMaker AI to personalize an acquisition funnel, achieving a high single-digit conversion lift for one audience, and learning why content, not the model, was the constraint.”
Recommended signals
Recommended by brand, topic, and recent related coverage