Here’s How Amazon Uses AI to Detect Water and Energy Waste

Amazon wants to put the “AI” in sustAInability.

The e-commerce giant announced Tuesday it has internally developed three systems, called FlowMS, Base Building Advanced Monitoring (BBAM) and Advanced Refrigeration Monitoring (ARM). Each system gives Amazon the ability to monitor its facilities for anomalies in energy and water usage by using comparative data points. 

The company built each of the three systems in house. They all use artificial intelligence to prevent energy and water from being wasted, in line with Amazon’s Climate Pledge commitments. 

While ARM applies most directly to Amazon facilities charged with storing groceries and perishable food items, BBAM and FlowMS use technology to identify issues with HVAC systems and water infrastructure, respectively, in standard Amazon distribution centers and fulfillment facilities globally. Those two systems have already been integrated into 120 of the company’s fulfillment centers globally, with plans to bring that number to 300 by the end of 2025. 

Already, each of the newly announced systems have made waves in Amazon facilities. 

In Glasgow, Scotland, at a logistics facility, FlowMS alerted engineers to an underground water leak that led them to a faulty pipe operated by a local water supplier. After identifying the issue, employees fixed the pipe’s valve, in turn preventing 9 million gallons of water from being wasted that year. 

BBAM helped New York-based Amazon employees to understand that a utility meter in the building was miscalibrated, which was causing it to appear that the facility had been using five times as much energy as other, nearby sites. It also aided employees in Spain in identifying a faulty air conditioning unit—the machine learning models powering the system flagged that the machine’s output did not match expected exterior conditions and demand in real time. According to the company, that helped mitigate the issue before it negatively impacted fulfillment center workers. 

Amazon expects to use BBAM to notify employees when dock doors aren’t properly shut in fulfillment centers and delivery sites after a truck loads or unloads its cargo. It anticipates that, when deployed at scale, this will “lead to substantial energy savings.” 

ARM, which Amazon expects to implement into 150 facilities in North America, Europe and India by the end of 2025, is focused on analyzing refrigeration units to ensure perishable products are stored at safe, optimal temperatures. AI also helps ARM to predict issues with the machinery in facilities like these, which can both prevent food waste and help fix refrigeration components before they fully fail. 

FlowMS, BBAM and ARM rely on sensors feeding data to AI systems, which alert humans when they flag any anomaly in the water or energy use throughout a facility. The human operator receives that notification via Slack message or email, and then must further investigate.

In some facilities, just one sensor helps operate FlowMS, but in other, larger fulfillment centers, Amazon employs up to three sensors. For BBAM and ARM, Amazon uses more than four sensors per data type, per room. That means, for instance, that one room could have four sensors monitoring air and four simultaneously monitoring zone temperatures. Upon receiving a notification of an abnormality—actual or projected—in the facility, engineers use the information provided to get to the root of the problem faster than they otherwise could. 

Because of the type of recommendations they provide to the user, Amazon’s BBAM and ARM would not technically be considered a digital twin in the traditional sense. Mature digital twins typically have multiple functions to them, and Amazon’s are primarily used for predictive and prescriptive recommendations focused on remedying issues preemptively. 

The company currently does not have a plan to deploy FlowMS, BBAM and ARM in its corporate office buildings, which have other sustainability measures in place. 

Kara Hurst, Amazon’s chief sustainability officer, said the company continues to iterate to improve its carbon and water footprints, and these systems are just the most recent example of that. 

“At Amazon, we’re innovating with AI to help us find new ways to decarbonize even faster, including inventing new solutions that continue to make our buildings more energy- and water-efficient,” Hurst said in a statement. “This is just one example of how Amazon is leveraging our decades of experience in AI development and sustainability to think big about decarbonizing our business and operating more efficiently.”