The requirement for fast decision making and the reality of increased data load make a case for cognitive automation.
Companies have done an impressive job figuring out how to automate supply chains to prepare and ship customer orders with speed, accuracy and precision. Amazon, the bellwether for such innovation, employs more than 200,000 mobile robots to move products from warehouse shelves, into boxes and on to customers. Pitney Bowes, meantime, has a massive fulfillment, delivery and returns Super Center in Greenwood, Indiana where 63 small robots on wheels help process up to 44,000 parcels per hour.
In some ways, this type of automation is the easy stuff. It’s what industry has always done. We offload time-consuming, manual tasks to the latest machinery to get products to customers as fast as possible.
But in nearly every supply chain process, human beings are still required to monitor everything and respond to arising issues. To make decisions. And we mostly do that well – except when we don’t.
Let’s face it, in a global economy, supply chain management has grown increasingly complex. Planners have to contend with so much data coming their way on a minute-by-minute basis it’s often impossible to keep pace. As human beings, we make an average of 35,000 conscious and subconscious decisions every day between our personal and professional lives. We really don’t have the capacity to handle much more. So, at some point, we either put ourselves on automatic pilot and make decisions least likely to get us fired, or we succumb to decision fatigue where we’re more prone to costly errors.
Each of these scenarios is avoidable. Indeed, a growing number of companies are minimizing such risks by using cognitive automation to enhance supply chain decision making.
Built on artificial intelligence (AI) and machine learning (ML) technologies, cognitive automation software works by crawling across enterprise systems to find, index and augment relevant supply chain data. It then tees up recommendations to advise decisions for improving core supply chain disciplines, including forecasting, ordering, inventory, logistics and procurement.
Experts say the need for greater visibility across systems continues to be one of the highest investment priorities for businesses. In fact, a JDA Software and KPMG LLP survey last year found 82 percent of supply chain executives polled planned to deploy cognitive analytics by the end of 2020 and 62 percent were investing in AI and ML. Together, 80 percent of respondents view AI and ML as the most important technology of all “given its wide applicability and promise of addressing complex business problems across the value chain,” the companies reported.
There are numerous ways in which cognitive automation technologies can be applied to not only help shorten the clock speed of decisions but also help planners make them more effectively. Here are five of the most common applications in use by organizations today.
Supply chain planners typically build work schedules according to the normal things one might expect such as how many people and trucks they have access to, current demand and where orders are headed. From there, though, it can get tricky as real life happens. People get sick and can’t come to work. Wildfires, floods, hurricanes or other natural disasters muck with transportation routes and timelines. Spikes or drop-offs in demand related to world events, such as the global pandemic, necessitate on-the-fly adjustments across the supply chain.
For small and nimble organizations, deciding how to address such changes might not be difficult. But for busier mid-sized to large companies, any one of them could be a tipping point for disaster.
Cognitive automation technologies can help by flagging impending problems and providing data-based labor scheduling ideas to get ahead of them – like locating additional in-house or outside drivers to fulfill orders or scheduling more or fewer employees stocking shelves or loading trucks as demand fluctuates.
Many third-party logistics (3PL) and fourth-party logistics (4PL) providers struggle to determine exactly how many delivery vehicles they need to support ongoing capacity. In the ideal world, they’d be able to adjust fleet levels every half hour or so to keep pace with the speed of commerce. But nobody can possibly assess all of the data and execute smart decisions that rapidly. So, planners default instead to ordering vehicles they think they’ll need on a more manageable weekly or monthly schedule.
The trouble is: what happens if demand fluctuates and they either over-ordered or under-ordered what they needed? In those situations, you either have delivery vehicles the company paid for sitting idle or not enough of them resulting in potentially ticked-off customers and lost revenues.
Cognitive automation can help by predicting variability and recommending real-time improvements to maximize efficiency in capacity planning.
Just as airlines do their utmost to avoid having planes departing with too many empty seats, supply chain planners detest when containers go out with any remaining airspace. But it happens all the time.
Part of the problem has to do with the fact shipments tend to be scheduled for specific days of the week. If containers aren’t loaded in time or there aren’t enough orders justifying the shipments, it doesn’t matter. They still go out, much like those half-empty airplanes we all love when we get a complete row of seats to ourselves.
The bigger issue, though, is that companies simply lack the visibility and transparency to actually know with any degree of confidence what’s being loaded or not at any given moment. Only 6 percent of them have full visibility into their supply chains, a Geodis survey found. They have a general sense of what’s happening, of course. Just not the up-to-the-minute specificity that would improve overall operational efficiency and minimize the chance that orders fail to ship because they didn’t get into a box on time.
Cognitive automation can help by collecting and correlating the most accurate data possible about what, where and when goods are being loaded. It then correlates, analyzes and offers recommendations that planners can use to make sure every container is as full as possible when leaving loading docks.
Companies have relied on a logistical method known as “cross docking” since the 1930s. This is where products are distributed to a customer or retailer with little or no handling or storage time. The name relates to how products are received at an inbound dock then moved over to another dock to be handed off to carriers. It’s worked for the most part. But with the advent of e-commerce and direct-to-customer being more common these days because consumers want what they want sooner, it’s become outdated.
Direct shipment is now the way to go for modern merchants. It’s faster, involves fewer hands along the way, and is thought to offer savings of $5 to $8 per shipment compared to cross docking. However, to do direct shipment right, companies need very tight control and clear visibility of end-to-end processes.
Cognitive automation can help by optimizing shipping networks, uncovering opportunities for more efficient transportation, identifying bottlenecks, and automating direct shipment delay resolution to ensure customers receive their complete shipments – on time.
Most companies have a disaster recovery plan that touches on supply chain contingencies. And for many organizations, that plan is likely old, outdated and gathering dust on a digital shelf somewhere.
In a day when global health, economic and political issues can disrupt business as we know it on a moment’s notice, agility is more critical than ever. You must know, for example, how to reroute delivery vehicles around natural disasters. If you’re delivering wine or fish or other spoil-ables in 100-degree temperatures, you’ve got to make sure you’re using refrigerated delivery vehicles or delaying shipment for cooler weather. Also, when you discover a supplier is suddenly out of commission, you need an immediate backup option.
Disaster recovery, by definition, is about having the ability to get back on your feet when the unexpected occurs. But studies show few businesses – especially small businesses – have formal recovery plans. More alarming, 93 percent of companies without one fold within a year of a major disaster.
Cognitive automation can almost serve as a disaster recovery plan. It provides the apparatus a way to collect and evaluate data from a variety of sources and make recommendations to adjust the process to compensate for unforeseen events.
As business leaders, we spend substantial time examining how we arrived at decisions. The thought processes and considerations going into them. At the end of the day, though, decisions are less about what happens in our brains and more about how we process the heavy loads of data coming our way every day. Taking that into account, it’s clear we sorely need the help cognitive automation can provide.