NextGen Technologies: BUILDING THE SUPPLY CHAINS OF THE FUTURE: A new crop of supply chain technologies promise to increase efficiency, visibility and speed. Five technologists share their vision of the future.
Today, a new crop of NextGen technologies is sprouting across the supply chain. Artificial intelligence (AI), robots, blockchain, the physical Internet and the connected supply chain are part of the discussion. As with their predecessors, they too promise increased efficiency, visibility and speed. But no technology, no matter how advanced, exists without people. At least not yet. So, there's also the matter of how people will collaborate with these new tools.
It's also worth noting that none of these technologies is fully baked. Like those before them, tomorrow's NextGen technologies will evolve in three stages.
The first is the "what's possible" stage. That's where most of what's discussed in this article stands today. Next is the "probable use" stage. The third is the "prevalent use" stage. Those two are where many think the technologies discussed here are headed, but we aren't there yet. It's always worked that way in the past, despite our impatience to know the answers today. Now is no different.
In the sections that follow, we asked experts in the field to share their vision of where we are and where we're going with supply chain technologies.
Taylor Smith: Building the connected supply chain
The pace of change in today's supply chain is relentless with ever-increasing speed and complexity combining with continuous business pressures. With these forces at work simultaneously, existing coping mechanisms aren't good enough for today--let alone the future. There has to be a grander scheme in play. Call it the connected supply chain strategy for equipment, people and logistics.
The three pillars here are: visibility, productivity and reliability. Supply chain visibility must stretch from goods availability to equipment assets, demand and orders. Productivity is critical to the delivery stage of each supply chain leg. Meanwhile, reliability and uptime of assets maintains the flow of goods through distribution centers and across the supply chain.
Going forward, major players here include sensors, the Internet of Things to move collected data, and advanced analytics, to name three. There is also voice-directed workflow, artificial intelligence and machine learning. Clearly, this isn't easy and we're only in the early stages of building the connected supply chain.
There's a lot at stake here. Eighty percent of distribution centers today are manually operated. Yet, SKUs continue to proliferate. The average DC manages 14,000 SKUs. And 90% of consumers consider fast delivery to be within two days. But that time frame is shrinking. Meanwhile, 70% of consumers are repeat buyers based on their delivery experience. Not making an expected delivery takes on a new importance.
Connecting assets in the DC is one necessary step in building the connected supply chain. We are well beyond the concept stage on this but nowhere near broad usage. A handful of large, highly automated companies are interested in making their DC equipment fully intelligent and capable of managing that data centrally to improve productivity and reliability. Sensors and data gateways will figure prominently.
The potential is enormous. On the one hand, these connections keep the facility running to fill current orders, avoiding untimely shut downs due to unexpected maintenance issues. But the horizon goes far beyond today's orders.
A highly productive DC, or network of them, opens the door to a reduction of capital expenditures down the road. Most conservatively, it could delay construction of a new facility as business expands. In the best-case scenario, that sixth or seventh facility won't have to be built. The throughput can simply be absorbed by more efficient existing DCs.
Then there's the matter of connected workers in the facility. On the one hand, guided work directs people through their workday using technologies such as voice and software. It's much more productive than worker selected work, resulting in greater efficiency, higher quality and better compliance with company work standards. Again, we're in the early stages of what it can be with only half a million people now using guided work systems. Future developments here will likely include natural language systems that are easier to use.
When it comes to logistics, the connected supply will feature real-time tracking and monitoring of shipments on trucks. For the most part, we are in the dark today as to what happens at any given moment to a truck's cargo between the time it leaves a shipping dock and arrives at its destination.
Early steps to shed some light on what happens in between involve several different types of sensors. Already, temperature, humidity, vibration, shock and light levels can be monitored. But each one is an endpoint in itself.
Going forward, the focus will be to build out narrow-band IoT networks tied to the Cloud. Cellular connectivity will be central here. This will allow sensor data to be streamed in real time and by truck location.
Just as important, these truck-based data gateways will extend tracking beyond today's high-valuable and perishable products. Ultimately, most goods will be tracked as they move between facilities in the connected supply chain.
--Taylor Smith is president of Honeywell's Workflow Solutions division
Benoit Montreuil: The new world of the physical Internet
It all started as a headline in the Economist in 2006. I picked up a copy at the airport on a trip back to the United States. The idea of a physical Internet sounded interesting. Unfortunately, the articles in the magazine never explained what a physical Internet would be. But the idea was so interesting that I spent the next months shaping what it could be.
The physical Internet (PI) is a supply chain metaphor based on the digital Internet. Quite simply, the physical Internet is a ubiquitous handling and logistics system for moving, deploying and realizing goods through the supply chain. It's a true open network of logistics networks. The concept is not a whole lot different than the digital Internet's movement and management of standard data packets around the world.
Today, there is nothing standard about moving goods at any stage of the supply chain except for maritime containers. The physical Internet changes all of that. It starts with standardized modular packaging, moves up to small modular handling containers and on to modular transport containers. It's a lot like a combination of Lego blocks and Russian dolls.
But it doesn't stop there. The physical Internet is also a meshed multi-party network of hyper-connected facilities that cross-dock and store goods through the supply chain. It's a continuous-flow, multimodal logistics model with a network of open hubs on one side and open distribution and fulfillment centers on the other side.
Open hubs aren't many days of driving distant from each other. In long-haul transportation, they are 100 miles to 250 miles apart. Drivers drop their standard modular transport containers at a hub for another driver to take the next leg. Then the original driver can pick up a different load and return it to that morning's home base. In urban omnichannel logistics, open intra-city hubs are to span the city, a few miles away, allowing to cross-dock handling and packaging containers from open peri-urban hubs and fulfillment centers toward their final destination, or vice-versa, exploiting green vehicles fitted to urban settings.
In the beginning, the physical Internet lived primarily in academic research. While that work continues, industry is now leading the way in the United States, Europe and Asia with a flow of innovative hyperconnected solutions, technologies and business models. This is mostly stemming from logistics and transportation service providers; brands, manufacturers and retailers; technology and equipment providers; and platform service providers.
Around the world, industry probably has tens of explicit PI projects underway. There are hundreds that incorporate aspects of PI. For example, take a close look at the Amazon fulfillment model and it's easy to see key aspects of the physical Internet. Its fulfillment centers are open to any vendor in the same way Amazon does for its cloud computing and storage services. This enables vendors to deploy their products to ensure fast delivery across the United States without investing in facility assets.
In Europe, there is an organization called ALICE (Alliance for Logistics Innovation through Collaboration in Europe). This EU-funded European technology platform is charged with developing a comprehensive strategy for supply chains and logistics systems on the continent. It has made the physical Internet the heart of its European strategic vision with mature widespread implementation targeted by 2030. Many millions of Euros are now spent annually by the European Union and businesses to make PI-based networks a widespread reality by 2030. That's a long way from a headline in the Economist.
--Benoit Montreuil, Ph.D., is the Coca-Cola Material Handling & Distribution Chair and professor at Georgia Tech and originator of the physical Internet concept
CAPS Research: Blockchain's impact on supply management
Yes, blockchain appears to be a technology that will have a large impact on the supply chain at some point in the future. However, like many technologies, blockchain is likely to follow what Bill Gates, former chairman of Microsoft, calls the 2:10 rule. It refers to two years of excited talk and then an additional 10 years before the technology is actually adopted and implemented into the mainstream of business. Blockchain appears poised to follow that pattern.
Blockchain holds great promise for procurement organizations. The first applications of this technology will likely be in areas such as financial services, payment processing and tracking the physical movement of goods. New blockchain applications will evolve and may dramatically change how procurement organizations and their supply networks are managed.
The shared and permanent nature of a blockchain creates a good environment for the transfer of assets through purchasing transactions, including inventory, sourcing, distribution and financial information. Traditional central record systems only reside inside a single firm to establish ownership and trust. However, blockchain's distributed ledger allows all firms involved to exchange information and assets in a secure manner. Instead of placing trust in a single entity in the value chain, veracity of the data is maintained among blockchain network participants through cryptographic proofs and data visibility.
As a result, blockchain has the potential to dramatically simplify the integration and automation of procurement processes. In fact, blockchain stands to replace supply chain transactions loaded with unintended redundancy, inaccurate information, manual intervention and interpretation and security vulnerabilities.
We believe that blockchain has the potential to remove redundancy as it simplifies and improves various portions of the procurement process. Such procure-to-pay processes most important to chief procurement officers and chief supply chain officers include: contracting, supplier evaluation/selection/ on-boarding, conflict resolution, catalog and data management, new product development and qualification, foreign currency exchange management, track-and-trace, ethical sourcing, compliance, risk management and cyber security. These may eventually all be part of blockchain-enabled systems.
Most early supply chain use cases have focused on track-and-trace, product genealogy and authentication of sourced materials. By initial indications, blockchain is exceptionally useful in supply chain applications requiring visibility. Early adopters will likely find niche areas where they can deploy blockchain and grow the functionality over time.
Blockchain-based procure-to-pay systems have begun to emerge too. Banks and fintech firms are developing blockchain-based systems that allow procurement to directly link with accounts payable and payment. Expect blockchain-enabled platforms to facilitate procure-to-pay systems. It may even be possible to facilitate decentralized progressive payments throughout the supply network between suppliers and fintechs. Fintech firms incorporate technology to support banking or financial applications and have begun to proliferate to facilitate transactions throughout the supply chain.
The track-and-trace capability is a perfect match for regulatory compliance. Blockchain provides great visibility of material flows in supply networks. Moving raw materials, components and finished goods through a blockchain also enhances traceability.
Blockchain technology is still very early in the development cycle, but it is likely to be an important technology that will evolve in interesting ways in the future. As a procurement or sourcing professional, it is vital to familiarize the organization with this evolving technology. An understanding of the frameworks and tools and ongoing developments and progress will provide opportunities.
This article is based on research conducted by Thomas Y. Choi, Ph.D., executive director of CAPS Research, along with. Dale Rogers, Ph.D. (lead researcher), Todd Taylor and Raymundo Beristain-Barajas. CAPS Research is jointly sponsored by member companies, the W.P. Carey School of Business at Arizona State University and Institute for Supply Management.
Pieter Abbeel: The near-, medium- and long-term future of AI and robotics
We are on the verge of making artificial intelligence (AI) a core competency of robots, and that will significantly change how they perform a broad range of tasks. This is a big deal. It will significantly alter how work gets done in the supply chain. Not only will it change current practices but promises to create new applications not considered possible for robots today.
This will take some time. Our current stage is to give robots eyes. Next, we want to use experts to give robots goals. And finally, we want to make it possible for non-experts to manage robots' activities.
In traditional automation, robots do the same action over and over again. Typically, these are simple, repetitive motions previously programmed to guide the robot. Once instructed, the robot performs the activity with little or no relief. Not very human-like at all.
Around 2012 came advanced computer vision thanks to Deep Learning; this is yielding the possibility of giving robots eyes. Combine that camera input with AI for decision making, and the robot now has a chance to first see and then understand the situation, modifying its actions accordingly.
At this point, we don't need better cameras. In fact, they have been good enough for years. The big challenge now is to take that camera input and use AI to more completely interpret the image data. That information will direct the robot's next action, which it may very well have never previously performed.
Making all of this happen is very difficult and complicated. But we are now at a stage where we can construct a machine learning situation for the robot. Going forward, our path is clear: Give robots the vision processing capabilities of humans.
That's the near-term status of the capabilities of AI and robots in the supply chain. And as should be clear, we are not close to completing this stage of development. But that doesn't mean we don't already know what we want to accomplish next. Which is good, because these developments are occurring somewhat in parallel rather than sequentially.
The mid-term goal is to create systems that allow robots to adjust their behavior on the fly. There are two key phrases here. One is imitation reinforced learning; in other words, imitate for the robot what it is expected to learn. The other term is reinforcement learning; that means giving the robot the objective or goal and let it learn on its own using AI.
We are moving through the first phase quickly right now. Using virtual reality, an expert demonstrates an action it wants the robot to perform. What has typically required 90 or 100 demonstrations is shrinking rapidly to a few and even a single demo in certain isolated instances. This is a good start, but not where we want to wind up.
Where we want to get to is only science fiction today. Now, we need experts--computer programmers, roboticists and others--to manage robotic actions. The key word here is experts.
Where we are headed is to get beyond experts; we want to make it possible for anyone to teach a robot. We can't do it yet, but there are plenty of people working toward that goal. When we arrive, robots will have a much different role in the supply chain than they do today
--Pieter Abbeel, Ph.D., is a professor at the University of California at Berkeley. He is also founder/president/chief scientist of covariant.ai, previously known as Embodied Intelligence
Daniela Rus: Integrating the workforce with NextGen technologies
We are at the front end of a transformation in how the work of supply chains gets done.
On the technology side, we have robotics, artificial intelligence (AI) and machine learning. Robotics puts computing in motion to move items in the facility. AI gives robots the ability to reason as to what they should do in various circumstances. And machine learning blends the capabilities of robots and AI.
These NextGen technologies, along with many others, will automate many aspects of operations. In many cases, technology will collaborate with people, with machines doing what they are best at, such as moving with great precision, and people doing what they are best at, such as strategic planning. Machines will take on the role of assistants to people and gradually they will be able to do increasingly more tasks.
Don't look for this transformation to occur overnight. Jobs have been changing for a long time, and they will in the supply chain. Just look at agriculture, which once accounted for more than 40% of U.S. jobs and today for less than 2%. There is no question that people will shift their approach to work in the supply chain. But there are some caveats.
Technology adoption requires that the needs of the workplace are better understood. This will lead to technology that can better meet those needs. And then people will be required to develop a much better understanding of technology's capabilities. In the end, this is a great opportunity for the workforce to develop new skills and uptrain. We will need to put in place people who can get the maximum benefit from NextGen technologies.
The front end of that shift is the emergence of data scientists and Cloud computing managers, to name two on the leading edge today. These types of jobs did not exist a few years ago. Today there are many people employed as data scientists and computing managers, and many job openings in this space.
Getting to the point where the job market needs match well the skills of the workers will require a fresh approach to education and training. Key players will be academia, industry and government. Collaboration will be key and no one has a silver bullet. This will take not only on-the-job training but also different types of training. In fact, we will have to fundamentally change our view of the relationship between education and work. Traditionally, education and work were sequential. Complete the first phase then put that knowledge to work in the supply chain. In the future, there will have to be a much more parallel relationship between education and work, with people learning and improving their skills continuously.
One potential solution is education through MOOCS--massive open online courses. These courses exist today and offer people an opportunity they haven't had previously. With MOOCS, people can become educated on subjects for which such education was just out of reach previously.
Just as important will be a better understanding by all of the impact of technology on work. For instance, in 2017 MIT hosted a symposium on AI and the Future of Fork (AIFUW). This symposium has become an annual tradition, with the 2018 meeting planned for November 8, 2018. AIFUM provides an opportunity for people to better understand the transformative capabilities of technology on the supply chain process but the impact on the workforce.
While the transformation of work due to technology will not happen tomorrow, AI is advancing at a rapid pace. The new capabilities put in place just in the past few years are already extraordinary. Beyond improving the supply chain process, NextGen technologies will make jobs more rewarding and economically positive.
--Daniela Rus, Ph.D., is the Andrew and Erna Viterbi Professor of Electrical Engineering and Computer Science and director of the Computer Science and Artificial Intelligence Laboratory at MIT
Gary Forger is the special projects editor for the Supply Chain Group. He can be reached at email@example.com.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Supply Chain & Logistics Technology|
|Publication:||Logistics Management (Highlands Ranch, Co.)|
|Date:||Nov 1, 2019|
|Previous Article:||2020 Trade Update: MORE COMPLEXITY IN COMPLIANCE: Most countries impose legal control on the export of goods from their jurisdictions, while...|
|Next Article:||LS2: Cranking up the SPEED of delivery: How the world's second-largest motorcycle helmet manufacturer put a scalable, on-demand warehousing and...|