Using AI/ML Analytics to Further Improve Efficiencies Within Your Supply Chain

By Dr. Sanjoy Paul, Prof. Hau Lee and Mahesh Veerina

In this series, we put different tracking approaches on a common foundation and categorize them into that we refer to as hard and soft attribute-based tracking. Then we argue why both hard and soft attribute-based tracking are important and how one complements the other, leading to near-optimal visibility. In this blog post, we discuss how to optimize your supply chain using AI and machine learning, and how you can potentially offer that as a service to other companies in the industry and create a new revenue stream. For more on this topic read the whitepaper A Holistic Approach to Supply Chain Visibility.

Further Optimizations Leveraging the Power of AI/ML

Let us revisit the concept of “Inbound Edge” and “Outbound Edge” and specifically focus on hard handoffs. The chances of delay and inefficiency with hard handoffs are higher compared to soft handoffs, assuming every organization is efficient in executing its own processes. Since an “Outbound Edge” triggers a set of actions starting with the “Inbound Edge” of the next downstream functional entity, it would make sense for the Outbound Edge to give “advance” notification to the Inbound Edge and help it get ready ahead of time. This will further streamline the process. For example, if the Outbound Edge, namely the “Delivered” step of Inbound Logistics Process (refer to Figure 1) informs the Inbound Edge, namely the “Received” step of Factory Process ahead of the actual delivery, the receiving process of Factory can be triggered and the equipment and workers at the receiving dock of the Factory can be ready to receive and move the supplies to the appropriate location within the Factory for processing without losing any time.

Figure 1: Activities (steps) within specific functional processes in a Supply Chain

The question is, how does AI/ML help in this process? With the introduction of both hard and soft attribute-based tracking, a tremendous amount of data will be collected, not only within a functional entity/organization, but also across organizations. That data can be analyzed using traditional AI/ML techniques to do predictions with a fair amount of accuracy. For example, it is highly likely that the lead time to deliver supplies by the 3PL entity to the Factory can be estimated with good accuracy, and an ETA can be provided to the “Receiving” side of the Factory so that the appropriate resources can be lined up “just in time.” Of course, the Receiving side should be ready for the Delivery from 3PL but should not be ready so much ahead of time that the resources are tied up when they could be used elsewhere. Thus, “just in time” readiness is important and that is best achieved by making prudent use of AI/ML algorithms for better estimation of ETA.

The above is just an example of estimation. Assume an ETA is being provided proactively at every step within a functional entity as well as across functional entities, and “just in time” readiness is implemented at every step in every process. While this might be a simplistic view of the complex supply chain processes, the fact remains, if we implement the concept of ETA at intermediate steps in an incremental manner, we can realize business benefits “incrementally” by reducing OPEX step by step, and that can be a journey and not a one-time exercise.

Optimized Supply Chains and New Revenue Streams

While we focused purely on cost savings by streamlining the processes within a supply chain, there are opportunities for new revenue generation as well. By virtue of higher efficiency, Suppliers and 3PL companies will be able to serve more customers with the same resources, and thereby generate additional revenue. By the same token, assuming steady demand, Factories will be able to produce more finished goods leading to higher revenues for the company.

Once a company is able to optimize the operations in its supply chain, leveraging the power of hard and soft attribute-based tracking powered by AI/ML, it can potentially offer that as a service to other companies in the industry and create a new revenue stream.

In this series of blogs, we have defined a framework to capture visibility in a supply chain from both hard attribute and soft attribute perspectives, described a process for computing what is called Visibility Index and defined a way of categorizing enterprises into four broad categories based on their Visibility Index. Keep in mind that contextual attributes are also an essential part of how companies manage their supply chains. It’s critical that companies understand what’s occurring both inside and outside of their supply chain so that they can successfully manage suppliers, parts, sites, and products.

Organizations need to pay close attention to news and trends (e.g., industry trends, trade disputes, tariffs, logistics, new technology such as drones, etc.) when identifying and managing risks for each component of their supply chain. Taking hard and soft attributes into account, 90% of companies are in the lower left (Basic) quadrant. But, with the right investments those same 90% can move to the upper right (Advanced) quadrant. If net profit on sales is 5%, net profit can be doubled if supply chain costs can be reduced from 9% to 4% (or from 12% to 7%) and our claim is, it is highly achievable as a company moves from the lower left to the upper right quadrant of the Visibility Index chart.

This is the big attraction and importance of cost reduction in a supply chain: profits can be increased without having to increase sales. In the process of defining the framework for Visibility Index, we laid the foundation for a digital twin of a supply chain that captures and depicts the state of a supply chain in real time, helping in agile decision making and risk mitigation. Finally, we argued that a Visibility Index is a means to an end, where granular and ground-truth based visibility can help enterprises close the gap between planning and execution, creating what we refer to as the Next Generation Supply Chain.

About the Authors

Dr. Sanjoy Paul is an innovator, disruptive entrepreneur, and an industry-recognized expert in AI & IoT.

Prof. Hau Lee is a Professor at Stanford University Graduate School of Business and Co-Director of the Value Chain Initiative.

Mahesh Veerina is a seasoned Silicon Valley entrepreneur, technology executive and investor and is the President and CEO of Cloudleaf.

All blog posts in this series:

1. Hard and Soft Attributes: A Framework for Supply Chain Visibility
2. The Impact of Hard and Soft Attributes in Your Supply Chain.
3. 1 + 1 = 3: The Power of Tracking Hard and Soft Attributes in Your Supply Chain.
4. The Benefits of Tracking Both Soft and Hard Attributes.
5. Using AI/ML Analytics to Further Improve Efficiencies Within Your Supply Chain. (This blog post)