How the Amazon.com Catalog Team built self-learning generative AI at scale with Amazon Bedrock

The Amazon.com Catalog is the foundation of every customer’s shopping experience—the definitive source of product information with attributes that power search, recommendations, and discovery. When a seller lists a new product, the catalog system must extract structured attributes—dimensions, materials, compatibility, and technical specifications—while generating content such as titles that match how customers search. A title isn’t a simple enumeration like color or size; it must balance seller intent, customer search behavior, and discoverability. This complexity, multiplied by millions of daily submissions, makes catalog enrichment an ideal proving ground for self-learning AI.

In this post, we demonstrate how the Amazon Catalog Team built a self-learning system that continuously improves accuracy while reducing costs at scale using Amazon Bedrock.

The challenge

In generative AI deployment environments, improving model performance calls for constant attention. Because models process millions of products, they inevitably encounter edge cases, evolving terminology ...