In December 2016, Amazon successfully delivered its first package via Amazon Prime Air to a customer in Cambridge England. According to the company, the entire process took just 13 minutes “from click to delivery.” Amazon Prime Air promised order fulfillment times of just 30 minutes or less by leveraging automated drone delivery.
If you’re an Amazon Prime customer living in 2019, then you know that this promised future never fully materialized. For most Amazon customers, delivery fulfillment still involves a two-day lag time rife with last-mile issues, including thousands of layoffs and multiple reported deaths. Free expedited package delivery, as it turns out, is not without its costs.
Despite these regrettable issues and the lack of nearly instant airborne drone delivery, Amazon has continued to make strides in last-mile delivery in many areas – most notably in cities that are lucky enough to be located near an Amazon distribution center. Amazon delivery truck routes are AI-optimized for maximum efficiency, and are equipped with GPS trackers linked to customer accounts so that Prime members can accurately predict and track delivery times. In some cases, the delivery driver will even send a photo of the delivered package to the customer as proof of delivery, though the results often end up something like this:
Even so, these advances pale in comparison to the leaps and bounds in data-driven efficiencies that have been achieved inside Amazon’s warehouses. The world outside of an Amazon warehouse is full of factors beyond the control of the company – traffic, inclement weather, poorly designed road systems, etc. However, inside the warehouse, Amazon is able to create whatever systems, mechanisms, and procedures that it can dream up. The indoor space of the warehouse is wholly and completely in the control of the company, and they are free to optimize however they see fit. It’s true that this can lead to unsavory, reprehensible outcomes, but it’s also true that Amazon has cracked the code of warehouse efficiency.
To be clear, the vision of an automated warehouse full of AI-machines operating without human intervention is still at least a decade away. Currently, Amazon has 110 warehouses in the US, 45 sorting centers, and roughly 50 delivery stations – all of which employ more than 125,000 full-time warehouse workers. Only a fraction of this overall labor is performed by robots. Currently, robots are simply too imprecise and clumsy to perform most of the tasks required for warehouse operations. But that’s not the point. AI can provide a huge amount of benefit to warehouse operations without a robot in sight, simply by optimizing the procedures and decisions that human workers execute. To learn more about why this is important, visit our primer on Data Science in Retail.
Not every warehouse can or should rise to the level of extreme organization and optimization demanded by Amazon, but the warehouse environment that does not apply machine learning for even basic logistics is the warehouse that wastes dozens of man hours and potentially loses millions in overhead costs. Drop us a line for a consultation on how to apply AI to your warehouse operations.