Amazon Reduces Packaging Waste with Artificial Intelligence

In the quest to increase its sustainability standards, e-commerce giant Amazon has used artificial intelligence (AI) and machine learning to maximize its efficiencies for decades. 

The company developed one such AI tool, the Package Decision Engine, to reduce packaging waste and optimize material use. It’s so far helped Amazon avoid over 2 million tons of unnecessary packaging and reduce packaging weight per shipment by 43 percent on average in the United States, Canada and the European Union. It's also helped Amazon avoid 80,000 metric tons of single-use plastic packaging globally since 2020.

Unfortunately, Amazon does not disclose how many tons of packaging material it uses each year. But in 2023, it delivered over 7 billion units by same-day or next-day delivery to Prime members. More than 4 billion of those deliveries took place in the U.S. — a 65 percent increase over 2022 — and more than 2 billion were in Europe.

For such huge volumes of products and an infinite number of order combinations, identifying the optimal packaging solution to keep each order safe during transit while minimizing the amount of packaging represents a significant challenge. The Package Decision Engine’s AI model is built to select the most appropriate packaging based on the product’s size, fragility and material type.

“The Package Decision Engine uses a combination of deep machine learning, natural language processing and computer vision, and is continuously learning about Amazon’s ever-evolving packaging options,” Kayla Fenton, senior manager of sustainable packaging at Amazon, told TriplePundit in an email. “It can predict when a more durable product like a blanket doesn’t need protective packaging, or when a fragile item like a set of dinner plates might need a sturdier box.” 

The AI tool learned that certain keywords associated with the product are important when making packaging decisions. “For example, a padded mailer with limited cushioning might not adequately protect an item with the words 'grocery,' 'screen,' or 'stoneware' in the description,” Fenton said. “So the model would recommend a sturdier option, such as a box. The model has also learned that keywords like 'multipack,' 'bag,' 'shrink,' and 'pack' are also associated with lower damage rates in the mail, and so indicates that the product might already have protective packaging and not need additional protection.”

The Package Decision Engine was conceptualized in 2015 as part of Amazon's broader sustainability goals, which include reaching net zero carbon emissions by 2040. Now, it's expanding to other markets including India, Australia and Japan. But “the tool will need to learn new languages and incorporate products specific to the new markets in order to continue rolling it out internationally,” Fenton said.

One of the ways it learns is by collecting information in near real-time from customer feedback reported through Amazon’s Online Returns Center, product reviews and other customer feedback channels. 

“After compiling customer feedback with visual information and other text-based data about the item, the model produces a score that predicts the best packaging type to use,” Fenton said. “The packaging selection is remembered by the model and used to understand future packaging needs.”

This feedback loop ensures that the packaging meets sustainability standards and customer expectations regarding product protection and waste reduction.

The company’s commitment to reducing packaging waste extends beyond its internal operations. Another AI initiative in the works that revolves around reducing waste is its pilot with Glacier, a recycling tech company based in San Francisco. 

“Glacier uses AI-powered robots to automate the sorting of recyclables and collect real-time data on recycling streams for companies, which can help reduce landfill waste and increase the use of recycled materials in packaging,” Fenton said. “Amazon is currently testing Glacier’s technology to sort and recycle new types of packaging that is not yet recycled at scale today, including bio-based and biodegradable plastics.”