Should We Trust Drones with Warehouse Management?

A sense of ​“hurry up and wait” has followed the field of supply chain robotics ever since Jeff Bezos famously announced on 60 Minutes that Amazon was developing drone-based delivery technology. As of 2019, that technology has yet to see even a limited rollout, and predictions for a total transformation of logistics have proven somewhat premature. However, that doesn’t mean that the supply chain and logistics industry isn’t seeing big shake-ups from drones, robots and other automation technologies. 

The most important advances in supply chain robotics have largely arrived inside the warehouse and fulfillment center rather than in the last mile. Rather than dropping off packages at customers’ houses, today’s supply chain drones and robots are picking items, counting inventory and scanning barcodes. So even if the automated delivery revolution hasn’t yet arrived, businesses large and small are finding many ways to put automated technologies to work. 

The question remains, however: How trustworthy are these technologies? Can they be counted upon to deliver the day-in, day-out performance and safety required to be a worthwhile investment? Many businesses have already made big moves into automation — are they seeing good results? And how can ERP systems provide a useful complement to drone technologies? 

Delegating repetitive tasks to self-operating systems is nothing new in the world of business and industry. Automation has long ruled the roost in many manufacturing sectors. However, new opportunities are arriving in the field of warehouse and fulfillment center management.

Thanks to the power of machine learning, businesses can create lean supply chain operations that enjoy many advantages over organizations that don’t automate their supply chain processes:

• Reduced resources allocated to low-value tasks. 

• Lower overhead and staffing costs. 

• Improved ability to scale quickly.

• Lower vulnerability to duplication and human error. 

• More proactive approach to inventory management.

• Improved ability to gather data upon request. 

The use cases for these technologies are many and varied, and more are being developed every day. A list of their current and proposed applications includes: 

• Monitoring and cycle-counting inventory through sophisticated aerial cameras and geographic sensors and then reporting through a system that flags potential issues, such as low-stock items and SKUs out of place. 

• Picking items, particularly ones whose bins are on shelves too high for workers to reach by themselves. 

• Self-driving and automation-assisted forklifts.

• Goods-to-person picking systems, in which an automated storage retrieval system (ASRS) brings picked SKUs to employees for packaging and shipping. 

• Assembling and packing customized mixed pallets for individual customers. 

• Fully automating entire warehouses, including heavy-duty tasks like unloading shipping containers.

• Robotic process automation (RPA), the use of software constructs and AI chatbots to improve processes from customer service to accounting. 

Although not all of these technologies are ready for widespread rollouts, there’s no question that the sector is on an upward trend. Supply chain RPA is now a strongly capitalized industry projected to reach $1.7 billion in 2019, with leaders like Automation Anywhere and UiPath raising hundreds of millions. And aerial warehouse drones have their particular champion in the form of California-based PINC, which counts an ex-Walmart executive among its board members and has worked on deploying the company’s drones in Walmart’s distribution centers. 

Thus, signs point to a big push in implementing supply chain automation within the next several years — but what kinds of results are to be expected? 

With investment flowing into the supply chain robotics industry, it’s fair to ask whether the technology appears to be fulfilling its promise. The answer appears to be largely a ​“yes,” but with some significant qualifications.

First, the good news: Available statistics show that these technologies are creating big gains in value and efficiency. A quick scan of the data from around the business world shows that some businesses are getting positive results from automation and machine learning technologies:

• Manufacturers have been using robots for decades, typically with excellent results. In fact, automakers in the United States have increased their production rates by 16 percent annually since 2010, and this increase is attributed largely to their expanded use of robotic technologies. Cobots, or robots designed to work alongside humans, are of particular note thanks to their impressive task flexibility and smaller footprint.

• A 2016 note from Deutsche Bank found that Amazon’s 2012 purchase of robotics manufacturer Kiva resulted in a 20 percent reduction in operating expenses, or around $22 million for each warehouse in which Amazon deployed the robots. Warehouse efficiency saw a big boost, with cycle times of 60 to 75 minutes plummeting to around 15 after the introduction of robots. Robots have also helped maximize warehouse space capacity, with inventory space growing by 50 percent in robot-assisted locations.

• Supply chain management software with machine learning capabilities has demonstrated excellent value in a McKinsey report, greatly reducing stock-outs and helping inventory move more efficiently for early adopters in the retail sector. Between predictive analytics and improved inventory management, AI-enabled supply chain management shows incredible promise — and unlike many robotics technologies, it’s ready for immediate implementation.

Unfortunately, however, reliable performance data is somewhat thin in this area. Robotics technology is still too nascent for hard numbers to be available about its performance on the ground. Moreover, as automation technology continues to grow and change, challenges continue to present themselves — and it’s often hard to predict where they’ll appear next. 

Investing in autonomous warehouse systems still makes many businesses nervous, and their caution is understandable. Many key aspects of supply chain robotics still have significant questions to be answered about their safety and scalability:

• Power players like Amazon may have achieved widespread implementation of robots in their warehouses, but very few businesses have that kind of resources. The challenges of implementing robotics systems for medium-sized and small businesses are significant, since these businesses are less able to take advantage of economies of scale. With the falling price of basic robot models and, especially, the advent of cobots, implementing automation at a smaller scale has begun to be a more viable option — but the field still clearly favors the bigger players. 

• Having autonomous robots working alongside humans in a warehouse is a challenging safety question. Some businesses, such as Amazon, have chosen to address this challenge by having robots work in separate areas of the warehouse. Other companies that have cobots in their facilities choose to make their robots slower in the interest of safety. No matter which strategy a business chooses, all stakeholders should be familiar with OSHA’s regulations on robots.

• Warehouse drones haven’t seen sufficient widespread implementation to collect real data on their safety. Many companies err on the side of caution and have drones running only when workers aren’t in the warehouse, or in areas where workers can’t go. Well-known UAV company AutoModality has designed navigation software that it claims allows drones to safely avoid workers, but public data isn’t yet available on its efficacy. 

• Laws governing indoor drones are largely non-existent, but liability is still an issue. Most laws governing drone use focus on outdoor applications, since they pose a potential danger to airplanes and helicopters. Since the FAA doesn’t consider indoor locations to be navigable airspace, it doesn’t regulate indoor drone use. Perhaps surprisingly, neither does OSHA. However, businesses that use drones without taking proper safety precautions face potential legal liability. 

• Finally, most of the stable and successful technology in supply chain automation focuses on what’s called ​“narrow AI.” Narrow AI is built to do one task or one set of tasks extremely well, such as scanning inventory or answering customer questions through a chatbot. The more narrow the tasks, the better current AI solutions tend to perform. A self-driving car, for example, counts as narrow AI, but the task involved is so complex that AI engineers have yet to create a truly road-ready build. This stands in contrast to ​“general AI,” the field that seeks to create AI that can easily learn new tasks, navigate novel obstacles and develop its own solutions. Most automation professionals agree that general AI still has a long development road ahead of it. So, while businesses can expect AI to continue to develop at a rapid pace, expectations of its capabilities should be kept realistic.

As businesses navigate the ups and downs of automation, many have discovered the importance of laying the groundwork first. The quality of inputs has a huge effect on the efficacy of an automated system, and drones and robots won’t fix problems that are already present in a company’s warehouse operations — so it’s key to improve those factors before introducing automation. 

A business without adequate inventory management systems, for example, will still be at greater risk for out-of-stocks and misplaced items no matter how often warehouse drones take inventory. A system that feeds inaccurate data into its automated systems is unlikely to see much improvement, so it falls on businesses to ensure that they’re using the most accurate data and the most appropriate platforms available. 

For any business that’s considering adding a significant automated element, ERP systems and warehouse management software platforms can provide the necessary data improvements. Many ERP systems now offer tools to synchronize automated warehouse processes through their platforms, giving credentialed employees an easy way to modify and optimize them. RPA is a particularly opportune area for ERP integration, as its processes often require easy access to data. 

Consider some of these use cases for integrated robotics and ERP: 

• Since robots can be synchronized with ERP systems, they can create real-time data flows that update the system as they scan and pick. 

• Managers can set target inventory levels for various SKUs, and robots and drones taking inventory can automatically flag and reorder items with low stock. 

• Some value-added services such as CNC machining may be able to be automated through ERP software. 

• RPA systems can ease the transition from legacy ERP to cloud-based models by performing many of the ​“busywork” tasks that occupy employees’ time.

• Chatbots can be integrated with RPA systems on a common platform to give customers up-to-the-minute information on whether a specific product is in stock and approximately how long it would take to deliver.

In general, AI and ERP integration is still in its infancy. Not every ERP system may be built for every automation function, and many vendors are currently working to expand their automation capabilities. Thus, a business investing in one or both should be sure to talk to the software vendor first and to research unbiased information on warehouse and distribution software they’re interested in.

Warehouse and logistics management are more competitive than ever before, and technology has only increased the relentless need to keep up. Drones, robots, cobots, RPA and other automated warehouse management tools are definitely the wave of the future — but most businesses should take care that they’re riding the wave rather than chasing it: 

• Focusing on how automation can address notable pain points or bolster core competencies, rather than seeking out novel use cases, is usually a much safer bet when investing in automation.

• Before implementing automated systems, ensure that there is a solid base of data to use, and check whether ERP vendors can offer automation solutions. An ERP information database like the one offered by Top 10 ERP is a great place to start. 

• Businesses shouldn’t rush into robotics investments when they don’t have the capital. Instead, they should explore the many potential benefits of smaller-scale automation options like RPA. 

• Maintain a strict focus on safety when implementing warehouse robot and drone technologies. 

The brave new world of supply chain automation is waiting. But all businesses should remember that it’s an extremely young set of technologies — so don’t force what doesn’t fit, and remain skeptical until value has been demonstrated.