Amazon AI initiative tackles defective products
Amazon is using artificial intelligence to scan items for defects before they are shipped.
An Amazon AI model named "Project P.I." (the P.I. stands for "private investigator") uses a combination of generative AI and computer vision technologies to help uncover defects like damaged products or issues like wrong color or size, before products reach customers.
In addition, Project P.I. is designed to help identify the root cause of issues, enabling preventative measures upstream to prevent them from happening again. At the sites where the system is available, Amazon says it has successfully sorted through millions of items that pass through its fulfillment centers each month and accurately identified product issues.
How it works
Before an item ships to a customer, it travels through an imaging tunnel, where Project P.I. uses computer vision to scan the product and evaluate the images to detect any defects, like a bent book cover. If a defect is found, Amazon isolates the product so it is not shipped to a customer, and investigates further to determine if there is a wider issue with similar items.
Human Amazon associates review the items Project P.I. flags to decide whether the item is eligible to be resold at a discounted price as part of Amazon’s Second Chance site, donate it, or find another use for it.
In addition to improving customer satisfaction, Amazon says Project PI also helps improve sustainability and reduce costs in the supply chain by eliminating unnecessary returns of defective products. The technology is expected to expand to additional Amazon fulfillment sites throughout 2024.
In parallel, Amazon is leveraging a generative AI system that uses a multi-modal large language model (MLLM) to investigate the root cause of negative customer experiences.
When the e-tailer learns of a defect from the customer that it failed to identify, it uses that error to understand the cause and continuously improve the system. The system first reviews customer feedback and then analyzes images taken from Project P.I. in fulfillment centers and other data sources to confirm what led to the problem.
For example, if a customer contacts Amazon because they ordered twin-size sheets but received king-size ones, the system cross-references that feedback with fulfillment center images and asks questions like, "Is the product label visible in the image?" and “Does the label read king or twin?”
Amazon also intends to share Project P.I. data on product defects with selling partners to minimize shipping errors in that area of its business. In other recent AI-based efforts to improve the customer shipping process, the e-tailer rolled out the Package Decision Engine, an AI model it designed and built to determine the most efficient type of packaging for each item it learns about.
"We want to get the experience right for customers every time they shop in our store," said Dharmesh Mehta, VP of worldwide selling partner services at Amazon, in a corporate blog post. "By leveraging AI and product imaging within our operations facilities, we are able to efficiently detect potentially damaged products and address more of those issues before they ever reach a customer, which is a win for the customer, our selling partners, and the environment."
"Amazon is using AI to reach our sustainability commitments with the urgency that climate change demands, while also improving the customer experience," said Kara Hurst, VP of worldwide sustainability at Amazon. "AI is helping Amazon ensure that we’re not just delighting customers with high-quality items, but we’re extending that customer obsession to our sustainability work by preventing less-than-perfect items from leaving our facilities, and helping us avoid unnecessary carbon emissions due to transportation, packaging, and other steps in the returns process."