Amazon has recently deployed 'Project P.I.' or 'Private Investigator', where it uses artificial intelligence (AI) to detect damaged goods and other issues before sending them out to customers. Amazon hopes that this technology will not only result in a more "positive customer experience" but will reduce unwanted returns, which "lead to wasted packaging and unnecessary carbon emissions from additional transportation."
The system uses computer vision to check a package's physical condition before the package is sealed, and flags an item if anything seems amiss. This package will then be forwarded to an Amazon associate who will then decide if the item in question will be discounted, donated, or disposed of. They could also look out for any false positives and keep the shipment moving forward if the system makes a mistake.
Aside from the on-site AI that tracks packages going out of Amazon's fulfillment centers, the company is also deploying a Multi-Model LLM (MLLM) to track negative customer feedback and returns and cross-references it with data from its source. This MLLM system will track any potential issues in its warehouses like mislabeled products or even find a bad batch of items before they're all sent out.
Amazon says that this system will help the millions of small and medium businesses on its platform, while simultaneously adding a second pair of eyes for its associates to catch damaged items before they are delivered. The company already deployed the AI tech across several fulfillment sites in North America, with more warehouses getting it this year.
