When I think of Artificial Intelligence (AI), the first image that comes to mind is Tony Stark’s “J.A.R.V.I.S” computer. J.A.R.V.I.S could answer any question thrown at it. History, math, computational problems… it was what we wished Siri would be when we pulled our new iPhone out of the box and engaged Siri for the first time. This type of AI is called General AI. Currently, we are years away from general purpose AI at the level of J.A.R.V.I.S, although IBM’s Watson is pushing the boundary. What we experience as AI is just an algorithm (or in Watson’s case a collection of algorithms) that acts on data to give the effect of simulating human intelligence.
For the supply chain, AI holds the promise of being able to optimize every action, every expenditure of time, fuel, resources and labor, to maximize throughput in ways that humans would never intuitively do through better optimization.
ERP does not think
The supply chain generates a huge amount of data. In the beginning, this data was manually cataloged (anyone remember inventory cards?) and retrieved. Electronic automation later added speed and scale to what could be logged and processed for the retrieval of the human overlords. Electronic general ledgers developed into huge business scale calculators called enterprise resource planning systems (ERPs).
When the humans have a question, ERP can pull everything together to answer that question very quickly. Today, ERP systems have evolved into huge knowledge centers, and the ERP itself has a perspective of the supply chain that no human could possibly hold in his head. The scale of ERP is beyond human capacity for accuracy and speed of recall… but the ERP does not think . It just responds to the rules and demands placed upon it by the human.
Here come the machines
Enter Artificial Intelligence. AI loves data. And AI can pull meaning out of data that humans can never pull out. In addition to being faster than humans at optimal decision making, the real benefit of AI is in its ability to find solutions to problems that wouldn’t be apparent to even the most skilled analytics experts. We call this unintuitive optimization . It means making improvements that you didn’t even realize were possible or needed.
Have you ever noticed that at the grocery store the beer is by the diapers? That seems like an odd place for it. Wouldn’t it make more sense to have the beer next to other adult products, or complimentary food items like snacks or meat? Well, an analysis of data from a 1992 study that examined 1.2 million super market purchases identified about 20 common product pairings like beer and diapers, as well as fruit juice and cough syrup. This conclusion took years of data mining and analysis from the best data experts. With AI, this unintuitive optimization would be recommended based on data without anyone even posing the question.
In modern warehouses, there is a lot of waste: wasted movement, wasted space, and cash sitting in inventory. What if every warehouse was optimally placed, stocked with the most efficient mix of product and was laid out to absolutely minimize movement in put-away and retrieval? AI holds the keys to driving down costs in ways that were previously impossible. Automation is coming, but automation holds only part of the future. The robots need a brain, and AI will start to appear at every level from the global supply chain to the optimization of a single action within a single warehouse.
For these intelligent warehouses, the AI connected to their business systems will be thinking about their business every second of every day. By analyzing every data point across the entire warehouse, the AI has far greater visibility into the comprehensive operation than any team member ever could. For pennies, it identifies and solves critical operational problems.
The next appearance of AI
Optimization recommendations : AI will tell you how to load and stock your warehouse for maximum efficiency. And you won’t have to input a bunch of rules into the warehouse management system (WMS) to make this happen, it will always be looking for an optimal product location and direct the users or automated system. This type of “always looking for optimizations” will apply to everything from freight and truck routes to ordering and planning. AI will be able to see patterns and recommend solutions before the humans realize there is a problem.
Human interactions : automating customer service requests while still giving customers a “human” touch. Current AI is learning how to simulate human conversation. This will be used to serve customers in an always on, always friendly and knowledgeable manner.
How to get ready
The promise of AI will begin to manifest, being embedded within products and as standalone services in the cloud. But to take advantage of the next quantum leap in productivity, you need to be ready. Real-time, accurate and detailed data about the state of the warehouse becomes the input to AI. AI must be fed, and the better the input, the more accurate the output.
The foundation of inventory optimization is accurate, real-time inventory. This type of data would be held within the WMS system. You need to get beyond “physical count” thinking into perpetual, real-time accuracy. The WMS is not just concerned about how much inventory you have, it is concerned with how it is handled. This provides the basis for warehouse optimization.
The ERP system will contain order and planning data. The ERP must be kept clean and current. Drive out manual work-arounds and email based processes from your business and systematize them. The rules that must be kept in someone’s head cannot be seen by the computer, and cannot be seen by AI. AI can learn to deduce things from dirty data, but clean inputs will get you better outputs.
Embrace open systems. Current cloud technology allows instant communication with the outside world: customers, suppliers, vendors, service providers, etc. Legacy systems are often closed or difficult to integrate to other systems. Think of your data systems as part of a global supply network and not islands within your organization.
We’re just at the beginning of the warehouse revolution. Already, advanced WMS, ERP and CRM systems are improving workflow within the warehouse and helping alleviate warehouse bottlenecking to allow the entire supply chain to be more efficient. The pace of technology change is getting ridiculous, and it will not stop coming. I am excited to see this change happening and to be a part of the change. Accurate, real-time data combined with AI are going to make a lot of operators a lot of money. Get ready for the wave, it’s coming.