Berkshire Grey has unveiled major AI-driven software enhancements across its warehouse robotics portfolio, using a decade of real-world operational data and billions of automated picks to help enterprise logistics operators navigate packaging variability and labor constraints.
Developed and continuously refined through a decade of real-world warehouse operations, BG Pilot software draws on insights from billions of robotic picks and one of the industry’s largest proprietary datasets of real-world robotic picking in commercial warehouse operations.
The company said BG Pilot addresses one of the most persistent and costly challenges in robotic warehouse automation: variability. The platform enables robots to handle diverse packaging and mixed SKUs while helping warehouse operations adapt to unexpected conditions, changing workflows and fluctuating labor availability.
Long before today’s surge of interest in physical AI, Berkshire Grey has been deploying intelligent robotic systems in complex, high-volume environments for customers including FedEx, Meijer, and Fanatics. That production experience is embedded in BG Pilot, enabling robots to perceive, decide, and act with enterprise-grade reliability, without turning customers’ operations into a training ground.
The enhanced AI capabilities come amid growing demand for Berkshire Grey’s robotic automation solutions, with new business opportunities increasing by more than 65% over the past six months.
“Growing demand for warehouse automation and accelerating interest in Physical AI are creating significant opportunities for Berkshire Grey,” said Dave Paratore, CEO of Berkshire Grey. “BG Pilot software is central to our ability to meet that demand, enabling us to continually advance our AI capabilities and help enterprises deploy robotic automation against high variability workflows with greater confidence.”
Berkshire Grey advances AI capabilities across the warehouse
BG Pilot is deployed in customer operations today, providing the foundation for new AI-enabled enhancements across Berkshire Grey’s Scoop, Stride, and Core products.
- Scoop — advanced perception: A multiheaded AI model processes visual and 3D data in real time to identify trailer walls, guide robotic picking, detect obstructions and items requiring human intervention, enabling more adaptive trailer unloading.
- Stride — intelligent orchestration: AI-powered orchestration dynamically coordinates product flow, station assignments and labor resources based on changing demand, shipping priorities, and operational constraints to improve fulfillment efficiency.
- Core — precision packing: Advanced machine learning, state estimation, and robotic control enable intelligent box selection, item placement and high-density packing, maximizing space utilization while accounting for product protection and handling constraints.
Together, these capabilities demonstrate how the software combines advanced machine learning with a decade of real-world operational experience to deliver intelligent automation built for the demands of modern warehouses.