Amazon Personalize is an AWS service that offers businesses a fully managed machine learning engine to build custom recommendation systems for their own websites, apps, emails, and marketing campaigns. It has been generally available since June 2019 and is based on the same recommendation technology used internally by Amazon.com.

Amazon Personalize takes user behavioral data (clicks, purchases, page views, signups), item catalogs, and optional user demographic data as input, and produces ranked recommendation lists as output — including personalized product or content recommendations, similar-item recommendations, personalized search result ranking, and next-best-action suggestions. The service automatically processes data, selects algorithms, trains and hosts models, and updates them in near real time as user behavior changes. As of 2025, it supports Transformer-based v2 recipes, a Content Generator feature using generative AI to create themed narratives around recommendation lists, and integration with Amazon Bedrock for AI-generated item descriptions and metadata enrichment.