Amazon has announced an update to its “Just Walk Out” checkoutfree technology. The update introduces a new multi-modal foundation model that uses transformer-based machine learning to increase accuracy in physical stores.

The new AI system analyzes data from cameras and sensors throughout the store simultaneously, rather than in a linear sequence. It processes multiple inputs, including cameras, weight sensors, and other data, prioritizing factors that help determine the variety and quantity of items selected. The system is trained on a 3D store map and an image catalog of merchandise for visual recognition.

Amazon claims the update makes Just Walk Out faster, easier to deploy, and more efficient for retailers. The company states that the new model reduces the need for retraining in unfamiliar shopping scenarios and can handle complex situations with variables such as camera obstructions, lighting conditions, and the behavior of other shoppers.

Just Walk Out is currently available in over 170 third-party locations across the US, UK, Australia, and Canada.