Supply Chain Drones: Novelty or News?
Drones may seem like toys, but AI is radically improving the value of drone-based data collection for supply chain operations. The key is pinpointing the information you could get from a bird’s eye view of a system you want to understand better.
The US House of Representatives just passed a bill that would ban sales of commercial drones made by DJI, a Chinese company controlling over 70% of global production volume. The ban, if signed into law, would prohibit new sales in the US but would not prevent the use of existing units already in the hands of American operators.
So what?
Whether this matters to supply chain leaders depends on how useful aerial drones could be in operations – and why. I think it does matter, mainly because drone applications for supply chain management depend more on what they can see than what they can carry.
Delivery by Drone Is a Red Herring
Amazon founder Jeff Bezos appeared on CBS’ 60 Minutes in 2013 to pitch the idea of package delivery by air with an octocopter drone. The company has since spent 11 years and untold millions of dollars developing the program, known as “Prime Air,” only to announce heavy layoffs last year. This is despite the fact that, during the same period, total company revenue has grown from $74B to $575B – and archrival Walmart launched a drone delivery service.
Drone delivery is legitimate, and recent FAA approval for flights without a pilot line-of-sight opens the door to more expansion. However, in practical terms, it looks likely to remain a niche application rather than a breakthrough. High-value, lightweight, single-unit shipments like medical supplies or spare parts may make economic sense. Plus, individual consumers will no doubt have the occasional need for a 30-minute delivery by air at a premium price, but serious volumes are not on the cards.
Eye-in-the-Sky Data Is the Killer App for Drones
Meanwhile, the use of drones for data collection and analysis is exploding. With drones, everything from inspecting pipelines to scouting crop health and taking warehouse inventories to spotting sharks off Bondi Beach is much easier. Companies like ExxonMobil, Land O’Lakes, and Maersk are driving innovation with these applications in a market that is projected to grow 26% per year through to 2033.
The underlying technology trends supporting this growth are less about aeronautics or material science and more a product of AI. Visual images are only part of the story, with thermal detection, geolocation tagging, 3D imaging, and volumetric calculations all adding depth to the data that can be retrieved safely and cheaply with aerial drones. Each of these data layers has diagnostic and planning value for operations, for example, yard management, plant maintenance, infrastructure design, and demand planning.
Also important is the use of emerging AI techniques for synthesizing large data sets based on limited data collection. For example, researchers from Los Alamos National Laboratory used an AI model called Perceiver (from Google) to recreate complete information on a large subject (an ocean) based on sparse data observations. The approach is reminiscent of Random Forest Regression techniques used by Nanjing Agricultural University to analyze field wheat health. Applied to drone-based data collection, these techniques could help operations leaders assemble super-detailed digital views of the real world.
Practical Applications Are Closer Than You Think
I recently heard the story of a semiconductor manufacturer using aerial images of parking lots at their competitor’s fab locations to gain intelligence on employee shift counts to better understand production patterns. It reminded me of a similar example where a retailer gathered parking lot data using drones to analyze shopper behavior. The common thread is realizing what kind of data is hiding in plain sight, provided you can systematically collect, scale, and analyze it to answer operational questions.
Drone services are everywhere, and independent operators like Dreamscape Aerials are ready and able to give you a look at whatever you want. It could be a building site, a railyard, a store, or the top shelf of your biggest warehouse. The key is imagining what information you could get from a bird’s eye view of any system you need to understand better.
Not a Novelty
Drones may seem like toys and the DJI ban like no big deal. Looking at the bigger picture, however, it is increasingly clear that drones offer a whole new way to think about operational data.
Supply chain leaders should keep their minds and eyes open.