Amazon AI technology is being put to a new use – the ability to detect damaged goods before they are shipped to a customer.
This isn’t a big issue, but it’s also a massive issue. The WSJ reports Amazon say fewer than one in a thousand items are damaged, so unless you’re ordering 3 items a day from the marketplace it’s unlikely you’ll ever receive a damaged product – especially as those packing the products do give them a cursory glance and pick up some faulty products. However Amazon ship in the region of 8 billion items a year and that means the one in a thousand start to add up.
Amazon AI is being rolled out at a small number of their warehouses to screen products for damage before they go to customers. Why you might ask are those packing the products not catching them? Well when you have a packing target to meet and you’re packing hundreds of items it’s hard to spot that one damaged product, especially as one minute you might be packing a book, followed by a toy, followed by drill bits, followed by a tent… the sheer range of products Amazon sell is immense!
The Amazon AI, already live in two warehouses and due to be rolled out in 10 more in the US and Europe, works by scanning products placed into bins to check for any wear and tear. If a product is identified as potentially damaged, it will be physically examined by an Amazon worker who will make the final decision as whether it can be shipped or should be moved to faulty stock. Having trained their AI to compare images of damaged products to pristine products, it has learned how to compare features to identify broken items. Amazon say their AI is three times more likely to spot a damaged item than an employee.
This is an interesting use of AI, often when we think of AI in ecommerce we’re considering how to automate processes such as optimising ad spend, managing picking routes in a warehouse, or perhaps generating better product description content. Amazon AI detecting damaged goods is an interesting development which highlights how technology can work to make our lives better and lower the burden for workers, in this case freeing them pack goods as fast as possible to satiate our never ending desire to receive purchases as fast as possible.
