Amazon tests AI staffing system to cut millions of labour hours

Amazon turned its attention to a new frontier in warehouse efficiency: the movement of employees rather than the movement of packages.

According to internal planning documents reviewed by Business Insider, the company is testing a system known as Full Facility Load Balancing (FFLB), which automatically reassigns warehouse workers throughout the day based on changing workloads and package volumes.

The technology is designed to reduce reliance on manual staffing decisions by managers. Instead, FFLB continuously analyses operational data, forecasts and package flow to determine where employees are needed most. The system reportedly recalculates staffing requirements every three minutes and recommends moving workers between departments when demand shifts.

If implemented across Amazon’s network, the company estimates the technology could recover around $193 million annually in labour costs while eliminating nearly seven million labour hours each year.

Expanding Automation Beyond Packages

The initiative reflects Amazon’s broader effort to automate warehouse operations. While robots and software have long been used to streamline package handling, the company is increasingly applying similar technology to workforce management.

An Amazon spokesperson told Business Insider that the system is intended to help managers make faster and more informed decisions rather than replace human oversight. The company described FFLB as an extension of existing staffing tools used in some facilities rather than a completely new system.

Amazon also challenged the projected savings outlined in the internal documents, saying the figures were based on hypothetical modelling rather than direct measurements of worker productivity.

“The goal is simple: streamline the complex operational decisions that can pull managers away from the floor so they can spend more time with their teams,” the spokesperson said, adding that managers remain responsible for final staffing decisions.

Focus on Container Build Operations

Much of Amazon’s effort is centred on a warehouse function called Container Build, where employees load packages into outbound carts and shipping containers.

Internal documents describe the role as the company’s “single largest labor automation opportunity” because it accounts for a significant share of labour hours in robotics-enabled fulfilment centres.

Amazon believes around 25% of Container Build labour time is linked to overstaffing and estimates FFLB could reduce unproductive labour in the function by roughly 40%.

One internal analysis covering 97 fulfilment centres found that 48 sites were operating below productivity targets, generating an estimated 309,000 excess labour hours each month. Another 266,000 monthly labour hours were associated with workers assigned to stations that had little or no work available.

The company noted that these figures represent only a small fraction of its total warehouse labour hours.

Wider Rollout Planned

Amazon has already started deploying FFLB at dozens of facilities and plans to expand the technology to robotics-enabled fulfilment centres across North America during 2026, according to the documents.

The company is also exploring ways to extend the system beyond Container Build operations, although Amazon said future deployment plans could change.

Internal records suggest some managers have faced challenges adapting to the automated recommendations, with some requesting configuration changes or feature adjustments. However, Amazon views those concerns as part of the normal learning process that accompanies new technology.

As artificial intelligence becomes increasingly integrated into logistics operations, Amazon believes tools like FFLB can help improve staffing efficiency while allowing managers to spend more time supporting employees on the warehouse floor.