Supply Chain Networks are evolving to become increasingly autonomous, letting intelligent software agents make simple decisions.
Full Article Below -
Untitled DocumentIn Part 2B, we looked at integration, how the networks can be extended, and a comparison of enterprise applications vs. integrator networks vs. Real-time SVoT networks. Here in the final installment of this series, we explore the future of autonomous supply chain networks and conclude with the ‘why’ behind adopting supply chain networks.
Is Autonomous Our Future?
There is an ongoing discussion about autonomous supply chains. No, that does not mean robots will be sitting at our desks. But it does mean that some of the drudgery of dealing with disparate and overwhelming data volumes can be reduced, opening up our time and minds to explore more possibilities.
Let the System Decide—Intelligent Agents
In a multi-party process, there can be hundreds of variables that influence even the most common decisions. Weather can impact independent demand—and dependent demand.
A closed bridge in a foreign land can affect on-time delivery a world away.
One can imagine the cascading dependencies and changes that might have to be made. Humans, who are already managing hundreds of shipments or thousands of products, may not be able to manage it all.
Autonomy does not have to be intimidating or beyond the reach of our comprehension. Simple but powerful improvements, like autonomously updating and improving lead-times or safety stock levels, can provide huge value.
Machine learning today can sit in the background watching, analyzing, and understanding. Intelligent agents that can replan, alert, reorder, and so on, based on massive data analyses, can determine the best option. This can take the drudgework away from users who can then focus on unique situations and person-to-person interactions.
Machine Learning, IoT, blockchain and other technologies have become part of the common language at supply chain conferences. However, the road to making them a reality varies according to their value and the community’s readiness.1
For example, Machine Learning/AI requires a deep base of data over a long enough period to actually have something to learn from. Then the appropriate Intelligent Agents (IA) need to be available to take action based on the best possible solution.2
IoT requires an instrumented supply chain and interoperability to rapidly access and utilize device, machine level, and location data. IoT can draw on machine learning to monitor and proactively avoid negative events, or at least apply predictive analytics based on device/equipment performance history.3
For many, the promise of a tunable system is important. Supply chain is getting so complex, and with expertise limitations, many companies want a lot more automation. We want to ‘tune’ routines and process flows to reflect exceptions or to evolve quickly with as little human intervention as possible. Within environments with isolated or restricted data instances, a mutual tuning to maintain interoperability may be challenging.
As we hope we have hit home by now, the inclusion of a network-wide ecosystem of participants and their data is foundational to achieving a smarter supply chain. With a smarter supply chain, we can leverage these more advanced capabilities, which can learn. With learning, the community can use and trust the system. With that trust, we can let the system make some decisions and operate autonomously across the network.
The technology is ready. The question is, are we?
Conclusions—Going It Alone or in a Network
Today, many users are still just trying to optimize their enterprise task using stove-piped systems such as inventory or TMS, working on each task independently. The rare enlightened supply-chain department might even have the logistics person talk to the procurement person and try to create a balanced plan.
Yet, over the years, our view of the supply chain, which encompassed limited functional views of one up/one down4 has needed to expand to a multi-functional, multi-stage view.5 We know our physical network includes all the elements that support fulfillment—seen and unseen. It’s not just partners, but also the environment that can have an impact on the flow and integrity of product.
As the demands of supply chain expand to include more and more partners and modern data, the complexities of synchronization become overwhelming.
Some organizations still harbor the ‘go it alone’ philosophy of building their own environment with ‘piece parts’ of technology—B2B/EDI, gobs of custom application interface programs, database tools and multiple applications. The challenge is that in the supply chain arena there is constant change and ever-expanding data. You are not in control of much of this. Thus, you need to work with organizations who have the scale and depth to support you. But this is not technology as usual, where the customer signs up for months of customizations and lots of new integration code to support for years, as well as managing all that data, while continuing to have detached/disjointed information with trading partners.
Improving overall performance across the whole chain is the quintessence6 of network objectives today. Leading companies—and that is not necessarily determined by the size of the company—who want to continue to innovate in their supply chains, participate in truly collaborative interactions and processes, and open themselves to more customers, partners, and markets, can leverage the power that networks have to offer.
To meet the supply chain challenge and seize the opportunities, there hardly seems a choice anymore. We need networks.
1 For example, there are several initiatives for blockchain, but rollout has become problematic. Some network providers do offer blockchain capabilities and, if called upon, they can rise to the occasion and provide them. -- Return to article text above
2 IA (Intelligent Agents) act on operational data just as traditional application code does. The methods and deployment of IA is an important topic but is beyond the scope of this article. -- Return to article text above 3 IoT is also a substantial topic, beyond the scope of this article. -- Return to article text above 4 Such as customer/supply product/inventory planning, or logistics, coordination of transport and inbound receiving. -- Return to article text above 5 Most systems today are still designed, sold, and implemented in these functional stovepipes, but that approach won’t get us to the goal. -- Return to article text above 6 Definition of quintessence: typical example of a quality; exemplar; stereotype; epitome; paragon, picture; prototype. -- Return to article text above
To view other articles from this issue of the brief, click here.