JDA Software and KPMG LLP recently published a wide-ranging survey regarding supply-chain technology. The main takeaway: end-to-end visibility is the No. 1 priority. But in order to make this a reality, the survey also notes that AI (Artificial Intelligence), machine learning (ML) and cognitive analytics will be critical.
Some of the ways in which AI can be used to impact Supply Chain:
The System Must Read Signals and Manage Billions of Pieces of Information: You need to process as many signals as possible to get a complete picture, such as weather events, temperatures, social trends and so on. For example, by using weather forecasts and port congestion data, it’s possible to predict the impact on freighters in route and determine which shipments will be late — and the captain may not even know what’s happening!
Or take another example: Let’s say an ice storm halts traffic on I-75 in Ohio. By using AI signals, you can answer questions like: What is every possible transit alternative and at how much added time or cost will there be? How will expediting some deliveries during the storm mess with the rest of the supply chain?
The System Must Look Into The Future: Rules-based approaches are too brittle to provide solid forecasts. In fact, these systems may do more harm than good.
“To help companies draw the right conclusions from the data they gather,” said Michael, developer of NeuroBayes algorithm, “businesses need to apply ML and AI technology designed to grasp the oncoming impacts of what’s happening everywhere in the moment and predict how demand and supply will look in the future. That means having algorithms that can evolve over time.”
He points to the following: Suppose you are doing assortment planning in a retail business. The traditional approach is to forecast sales based on prior history and trends. “A retailer may always send one style of athletic shoe to the Midwest because they know the sales history and the product does well there,” said Michael. “But with ML and AI, there is now the ability to blend external and internal data to predict demand and areas for growth. If retailers take an index and predict where customers are most concentrated, that data can help them figure out where to ship the athletic shoe to maximize their sales.”
The Technology Must Overcome Human Nature: So long as the data is correct and the algorithms appropriate, then an AI system will learn and react to ensure that the orders and price points remain in line with a probability that keeps a business both stocked and efficient.
“However, as humans, our instinct is to fix things ourselves, especially if it’s an area we have been tasked with overseeing,” said Michael. “The autonomous supply chain requires us to discard pride, ego and personal bias and trust the technology. As trust in the system’s recommendations increases, a greater and greater portion of decisions can be made automatically by the system, without human intervention. This will allow the professionals to focus their time and effort on problems that only they can solve.”