Robots have taken their place inside eCommerce and other commerce-related warehouses, and will in the coming years take even larger roles in fulfillment, according to estimates. Hot on their trail are machine learning and artificial intelligence (AI) technology — the software and algorithms promising to reduce the risks of overstocking and understocking, and providing other benefits that can boost retailers’ revenue.
Overstocking costs retailers about $470 billion annually, according to one of the most recent estimates, this one from IHL Group. Understocking is even more expensive — about $630 billion in global annual costs. Algorithms — most notably, the ones used by Amazon — already help assuage both problems by predicting consumer demand “for hundreds of millions of products it sells, often as much as 18 months ahead,” according to The Economist.
That said, increasing use of AI within the supply chain promises even “better predictions will improve inventory management and demand forecasting, too, freeing up cash and storage space,” the report said. It found that one company that provides food storage for grocers and retailers boosted warehouse efficiency by about 20 percent via the use of AI for what’s called “smart placement” — that is, information and predictions about the placement of pallets, a prime source of labor and time consumption in the fulfillment process.
On a basic level, machine learning and AI algorithms place “rarely ordered items at the bottom and frequently accessed items on the top,” reads one recent analysis. “This minimizes the amount of time it takes to complete the majority of the company’s online orders.” Depending on what predictions you believe, most retail warehouses could be fully automated by 2030. And AI will play a major role in that.
“The end goal of automated warehousing is computer vision,” that analysis said. “This is where the AI will learn and evolve as it works to improve operations. Computers will soon be capable of recognizing and organizing inventory, and even administrating quality control for a variety of stock without the need for human oversight. If the company has more than one warehouse, the AI in each location will be able to communicate with each other to find the best logistical solutions.”
Supply Chain Trends
The supply chain is thirsty for technology, including more machine learning and AI, according to survey results from Symphony RetailAI. For instance, it found that “48 percent (of retailers) rate their forecasting technology as average to very poor.” Not only that, but lack of unification when it comes to supply chain and fulfillment stands as a major problem for retailers — which could create opportunities for the use of more artificial intelligence in those areas.
“The challenge for retailers is that they lack connected systems — 36 percent of respondents indicate that they have separate demand planning, replenishment, allocation and order management systems for store and eCommerce orders,” the survey results found. “Combined with the fact that 28 percent don’t manage each of their modules on the same platform, it’s clear that disparate demand replenishment systems significantly burden efficiency.”
But the race toward that AI supply chain future is not an easy one, as 43 of the retailers in that report said they don’t think the technology they are using is keeping up with business demands. That can lead to various inefficiencies and lost revenue — including stocking too much or too little product — with 43 percent of respondents saying they are “challenged by lack of real-time visibility to all supply chain inventory.” As for AI, one in three of the survey respondents said they have deployed AI into supply chain management, though one in four said they are working toward that.
The report did not go into detail about how those retailers and supply chain professionals view the differences between machine learning and artificial intelligence. As demonstrated by recent PYMNTS research entitled “The AI Gap: Perception Versus Reality in Payments and Banking Services,” there are common misunderstandings about those differences — in general, true AI can do unsupervised learning — and, in the financial sector at least, scant deployment so far of true AI. That is certainly no apple-to-apples comparison when it comes to supply chains and fulfillment, but is does indicate that the road to true AI is a long one.
That said, there is no going back, it seems, on the robotic wave hitting commerce-related warehouses. By 2022, the warehouse robotics market will have grown to a value of some $4.44 billion, according to one estimate. But where the robots go, it seems all but certain that much more AI will follow.