Amazon is beginning to use artificial intelligence (AI) to spot damaged items before they are shipped to customers.
The company has implemented the new technology at two fulfillment centers and plans to add it to 10 more locations, The Wall Street Journal (WSJ) reported Wednesday (May 31).
The AI-powered solution is trained with photos of undamaged items so that it can identify damaged ones, according to the report.
The technology is added to an imaging station that the goods already move through and that confirms that the proper items for an order have been picked. That station will now check for damage as well, the report said.
If the AI identifies an item as broken, the item is diverted to a worker who will double-check it.
At locations without this technology, checking for damage is done by the warehouse workers who also pick and pack orders.
This time from a worker who must meet per-hour targets, it’s a task that is not their primary responsibility, and it’s something that can be easily overlooked, because fewer than one in a thousand items are damaged, the report said.
Thus far, Amazon has found that the AI solution is three times better than warehouse workers at spotting damaged items, per the report.
Amazon competitor Walmart has adopted new technology to further refine its omnichannel fulfillment capabilities too.
Walmart acquired robotics automation company Alert Innovation in October 2022, saying it is bringing that company’s capabilities in-house to scale its proficiencies and deploy market fulfillment centers (MFCs) more quickly.
The retailer has been working with Alert Innovation over the past five years to customize Alert’s material-handling technology for use in Walmart’s MFCs, which are compact warehouse built within, or added to, retail stores to help fulfill pickup or delivery orders.
About two-thirds of jobs in those regions are exposed to some level of AI automation, the firm found.