Shaping benefits from emerging technologies
In the e-commerce sector, there is a growing need for reliable technology solutions to tackle challenges like pushing small orders quickly, ensuring efficient and cost-effective supply chain and distribution operations, last-mile logistics challenges and more. While many e-commerce sector companies are facing different results by applying individual emerging technologies, the following article suggests that a blend of these technologies are likely to yield the best outcomes
It’s not only the growth of e-commerce, but also the ever-tightening delivery timeframes that have logistics pros eyeing emerging technologies like artificial intelligence, blockchain and the Internet of Things. The challenge is to determine which of these can alleviate concerns like spiraling transportation costs and pulling together final-mile delivery resources.
The march of e-commerce now has Amazon advertising two-hour food deliveries in select cities. In this environment, more companies face pressure to deliver goods at a faster pace, all while keeping costs under control. To cope, many organizations are pinning their hopes on emerging technologies such as artificial intelligence (AI), autonomous mobile robots (AMRs), blockchain and the Internet of Things (IoT). But when will these “emerging” technologies form practical applications that solve logistics pain points?
To find out how these emerging technologies are evolving into logistics relevance, we talked to top consultants and analysts. We found that AI is already being applied to processes such as transportation routing decisions and freight matching, sometimes within supply chain control towers.
At the same time, IoT is being used for reasons such as predictive maintenance over vehicle assets and in conjunction with AI to better pinpoint estimated times of arrival (ETAs). And while the new technologies still need to mature to shake the “emerging” moniker, it’s also apparent that organizations can start leveraging most of these technologies today.
Al’s many uses
Today it seems that AI is on everyone’s lips. The reason: AI is a technology with broad scope, explains Chris Striffler, a senior manager with Clarkson Consulting. “AI and machine learning (ML) have broad applicability, so we’re seeing a lot of traction around those,” he says. “By contrast, blockchain has strong potential in areas including traceability and contract management, but it doesn’t have the broad applicability that AI has.”
AI can be thought of as next-level analytics that constantly sifts through Big Data to arrive at optimal decisions. A subset of AI is ML, which is able to recognize data patterns, learn from them, and come up with recommendations on ways to save costs or improve service levels. With some AI projects, AI output can automatically trigger decisions to be executed within solutions like a transportation management system (TMS).
Within the TMS domain, AI can churn through transactional history, current pricing data from carriers, along with current weather and traffic data to make better freight and routing decisions, explains Michael Daher, a principal and transportation practice leader with Deloitte Consulting. Daher adds that many clients are layering AI capabilities into supply chain control towers-systems that centralize logistics data and shipment visibility.
According to Daher, AI’s output can automate better decisions around load planning, mode selection and carrier selection, says Daher. “By leveraging AI to help automate decisions in these areas, the managers are able to focus more attention on the strategic nature of their supply chain management network.”
Deloitte sees logistics evolving toward three pillars: connected communities, holistic decision-making and intelligent automation. AI is central to holistic decisions, Daher says, but it also overlaps nicely with supply chain control towers.
“I think the first step on the journey is setting up control tower environments,” he says. “The second evolution is applying AI to be able to fully leverage the data essentially being collected by the control tower.”
Within some larger enterprises, there’s a centralized analytics group working on AI projects and other projects that tap emerging technologies. These enterprises tend to be quite forward thinking about AI, but small to mid-sized companies can also work with consultants to rapidly deploy AI within “proof for value” projects, says Striffler. “Frankly, logistics and freight are a great place to start with a use case,” he says. “That said, it’s smart to view AI not just as a tool for point solutions, but as part of a larger strategy.”
AI is being applied to freight management challenges by C.H. Robinson, a third-party logistics (3PL) provider, according to Tim Gagnon, vice president of analytics and data science for the company. “We have effective AI solutions embedded in our pricing, tender acceptance, freight matching and visibility services,” says Gagnon, who heads up C.H. Robinson Labs, an innovation incubator for the company.
Gagnon explains that AI capability can learn from the data and processes it analyzes, and be used within a control tower. “What excites us most about these solutions is that they can effectively learn from millions of decisions and outcomes to develop more effective judgments of risk and reward, ultimately learning to better efficiency, visibility and reliability for our customers,” he adds.
IoT plus analytics
The IoT, when combined with analytics, is helping organizations with issues such as asset uptime and real-time knowledge over the location or condition of shipments. It’s the combination of analytics with IoT-connected sensors that brings the benefits, says Ryan Kauzlick, a vice president for consulting firm enVista. “If you’re collecting all of this data, you’ve got to understand what it actually means, which is where analytics comes in,” he says.
Top IoT use cases include predicting equipment failures, protecting the safety and quality of “cold chain” distribution along with asset and associate tracking, according to Kauzlick. “Where the IoT can really begin to be value-enhancing is when it can predict when and where these failures will occur, so you can change or refocus procedures,” says Kauzlick.
In some IoT applications, multiple data sources may need to be fed into a predictive analytics platform, says Kauzlick. Fortunately, analytics platforms from Microsoft and others are getting easier for organizations to use without having teams of programmers working on a solution for months.
Kauzlick says that his team often leverages Microsoft’s predictive analytics platform. “These platforms are absolutely lowering the barrier of entry for the analytics side of IoT,” he adds. “They allow an organization to create a data set that is trusted, and that you can then easily reach into for analytics.”
When it comes to all of the details involved in moving goods across borders and recording changes to chain of custody over goods, blockchain is seen as an ideal technology. That’s because it’s a digital ledger that lives in the Cloud where partners can access information easily without corrupting it. These characteristics make blockchain ideal for traceability storing cold chain data, proof of delivery, or contract details involved with global trade, says Striffler.
According to Daher, one of the biggest applications for blockchain is removing friction from Custom clearance. A blockchain could hold common, easily verifiable data on details like bills of lading, certificate of origin, insurance or invoicing. “Right now, there is a lot of time and administrative resources invested in Custom clearance that blockchain could help reduce,” says Daher.
Blockchain has various pilot projects or consortiums working to prove its supply chain value, including Walmart’s program to trace green leafy produce supplies, and the Blockchain in Transport Alliance (BiTA). A bit more time and broader participation is needed for blockchain to really take off, says Striffler. “Organizations are starting to dip their toes into the water with blockchain, but it needs that critical mass to be fully effective. We might be a couple of years away from that point.”
The growth of same-day-and in some areas two-hour-delivery windows is making “last-mile” processes a focal point. In terms of software for last-mile, the trend is toward a more holistic approach that looks beyond tactical scheduling, routing and management of deliveries, according to Bart De Muynck, research vice president for analyst firm Gartner.
“Last-mile solutions were in the past more purely focused on the transportation part, but now the evolution is toward a more holistic view that can be termed last-mile delivery orchestration,” says De Muynck. “As part of this evolution, software vendors are offering systems that can help improve the customer experience.”
Vendors with monikers such as “delivery experience management” address this need to gather information on what customers expect from e-commerce deliveries. The idea is to fine-tune the experience, says De Muynck, to keep customers for the long term. “It becomes like a feedback loop where you can manage everything involved with the delivery so the next time a customer places an order, you’re able to provide a much better experience.”
Traditional routing and scheduling solutions for delivery fleets have been around a long time and continue to be used, but the last-mile landscape now also involves “crowd-sourced” delivery drivers, as well as established parcel delivery companies. More use of drop-off kiosks and pickup locations adds further options to last-mile scenarios.
The added complexity is giving rise to vendors that offer what amount to “last-mile delivery orchestration platforms,” says De Muynck. These solution sets span multiple functions including managing crowdsourced drivers, with the aim of helping companies figure out the best last-mile options. “From a technology perspective, we’re starting to see a new type of application, which is these last-mile orchestration solutions,” he says.
And, e-commerce players-as well as online grocery companies-are experimenting with small autonomous delivery robots, while Amazon and others have experimented with small aerial delivery drones. Currently, says De Muynck, such use of delivery bots and drones remains experimental and would face regulatory hurdles and cost effectiveness concerns, but it’s inevitable that companies will need to continue to test such technologies to drive down costs as delivery windows get tighter.
“Every time delivery expectations get shortened, we see increasing costs, so everyone is looking for those technologies that can either provide a way to either bring down that cost of delivery, or provide the customer with a better experience,” says De Muynck. “Ultimately, if you can provide a better customer experience, some customers are willing to pay for it.”
Autonomous trucks continue to draw interest from investors as well as companies that need to find ways to get goods to consumers quickly and cost effectively. Companies developing autonomous truck technologies include Alphabet/Waymo, Embark, TuSimple and Starsky Robotics.
Rather than fully replacing humans with fleets of “driver-less” trucks, says De Muynck, a more likely near-term scenario will be to pair autonomous trucks with human drivers to extend range and lower costs. The human driver could handle urban environments and loading/unloading interactions, while the autonomous vehicle handles the driving for long stretches of highway.
Assuming regulations over driver rest could be ironed out, this pairing of human drivers and autonomous trucks could keep assets rolling to effectively speed up transit times. In some regions, this might eliminate the need for additional warehouses to serve customers for next-day deliveries, notes De Muynck, which would offer a major cost savings.
“This team model, in which one team member is a human driver and other the autonomous truck, could help move more products more quickly and at a lower cost than is possible with conventional trucks,” adds De Muynck.
The new robotics normal
With the time pressures and extensive picking of small items involved in e-commerce, warehouses are under intense pressures to accurately get orders out the door rapidly. The labor shortage has made this extremely difficult to do under manual or semi-automated methods.
These two key challenges are why autonomous mobile robots (AMRs) are widely seen as one of the main technologies that help supply chains with e-commerce. Bringing in AMRs is simply seen as a tactical necessity, explains Remy Glaisner, a research director for analyst firm IDC. “Labor availability is a big challenge,” he says. “Basically, many operations simply can’t find enough people at the local level, which drives operations to look at robotics.”
Glaisner estimates that across all DCs, the current penetration rate for AMRs is somewhere between 1% and 2%, making it early days for AMRs. That said, close to 60% of companies are considering robotics, he adds.
Once an operation does roll out robotics, it’s typical that key metrics are improved by roughly 20%, adds Glaisner. This creates what he calls a “new normal” in which the operations begin to see steady benefits in areas like order accuracy and more predictable cycle times.
“Initially robotics help solve the issue of not being able to find enough staff, but after they’ve been in use for a while, it tends to set up a new normal for the performance of an operation,” Glaisner says. “That is when managers realize that, yes, we can actually grow our business or serve customers better thanks to robotics being deployed in the operations environment. Mobile robotics can be transformative in that they can help move the warehouse environment from being more purely a cost center concern to being a source of value generation.”
Ultimately how various emerging technologies can help with the challenge of e-commerce is more than deploying each one as an isolated technology. They tend to overlap in a good way. IoT needs predictive analytics and AI, while AI and ML are also baked into AMRs and autonomous trucks.
You don’t to pick and choose from a list of emerging technologies – you might blend ML with IoT to understand the operational implications of massive data sets, Striffler points out. “There can be strong synergies from implementing these technologies together,” he adds.