Amazon tests AI system for moving warehouse workers in real time

Business Insider reported that Amazon is piloting a system called Full Facility Load Balancing, or FFLB, inside robotics-enabled warehouses. Internal planning documents reviewed by BI say the system forecasts package volumes and workload signals, recalculates staffing needs about every three minutes, and recommends moving workers between roles. The documents modeled roughly $193 million in annual labor-cost recovery and nearly 7 million labor hours, though Amazon said those numbers were hypothetical and not guaranteed savings. The story matters because Amazon's warehouse AI is moving beyond package automation into labor allocation, with managers still described as decision-makers but software increasingly shaping where people work.

Key details: Published June 18, 2026 at 09:00 UTC, The system is called Full Facility Load Balancing, or FFLB, Business Insider says it can recalculate staffing recommendations roughly every three minutes, Internal estimates modeled $193 million in annual labor-cost recovery and nearly 7 million labor hours, Amazon said the estimates were hypothetical and managers remain decision-makers.

Why it matters: This is the labor side of warehouse AI: once software starts recommending where workers should move minute by minute, automation affects supervision and staffing even before jobs disappear outright.