Product

Amazon SageMaker AI

by Amazon

Also known as: Amazon SageMaker AI

Articles
1
Coverage
Aug 31, 2026
Type
ML Platform

Themes

Articles

View in news feed →
Personalization and customer data drive conversion gains

Amazon Payments deploys contextual bandits to lift conversion funnels

Fintech / Payments › Depository Credit Intermediation › Commercial Banking
DataLLM

Amazon Payments used Linear UCB contextual bandits on Amazon SageMaker to personalize content across a three-stage acquisition funnel, achieving high single-digit relative lift in final-funnel conversion for one customer segment in a seven-week test. For commerce teams, this demonstrates how multi-armed bandits can optimize entire customer journeys simultaneously rather than isolated metrics, avoiding the trade-off pitfalls of stage-by-stage tuning.

Retailers adopt AI for demand forecasting and optimization

Decathlon scales demand forecasting with Chronos-2 on AWS

B2B Distribution › Miscellaneous Durable Goods Merchant Wholesalers › Sporting and Recreational Goods and Supplies Merchant Wholesalers
AnalyticsData+1

Decathlon deployed Amazon's Chronos-2 time series foundation model to forecast demand across tens of thousands of products across multiple continents, replacing a hybrid system of DeepAR and exponential smoothing. For commerce teams, foundation models now offer production-grade accuracy with lower operational overhead—no weekly retraining required.

General AI in Commerce

AWS Bedrock Data Automation optimizes document extraction accuracy in minutes

Data

Amazon Bedrock Data Automation now includes blueprint instruction optimization, which automatically refines extraction instructions using three to ten example documents with ground truth data to improve accuracy in minutes rather than weeks. For commerce teams processing invoices, contracts, and enrollment forms across multiple vendors, even small accuracy gains translate directly into reduced manual review and faster throughput.

Machine learning infrastructure accelerates AI-powered business solutions

AWS and NVIDIA Enable Distributed Robot Learning on SageMaker AI

Data

AWS released a solution for scaling robot reinforcement learning with NVIDIA Isaac Lab on Amazon SageMaker, offering both managed HyperPod clusters for long-running jobs and ephemeral Training Jobs for rapid iteration. This removes infrastructure overhead for robotics teams, letting them focus on policy development while compressing months of real-world training into hours of GPU-accelerated simulation.