Transforming Value Stream Mapping With AI (Part I of III)

There will hardly be any supply chain graduate who did not take a value stream mapping course. A fundamental methodology in lean manufacturing, value stream mapping essentially facilitates analysis of value streams, and then leveraging those mappings, we can identify opportunities to optimize the processes by eliminating non-value added aspects.

Key process steps and data (like throughput time) can be visualized leveraging a value stream, and can help provide a good overview of the current state, not only in terms of process steps but also in terms of key metrics and data pertaining to the process steps. If done right, a value stream map becomes a powerful communication tool as well, to highlight improvement opportunities and build collective agreement on them.

Traditionally, value stream is a very manual process. Some experts believe that it is the manual aspect of the process that makes it so effective. That may be true three decades ago, but is definitely not true now (in my opinion). Anyways, in the traditional format, the data for value stream mapping is collected on the shop floor and maps are typically hand drawn, on paper. Walk into any plant and you will find hand drawn value stream maps of lean boards.

But it is the manual collection and hand drawing exercise that, in my opinion, limits the value. While the first aspect that may come to your mind is that building these maps is manual and hence inefficient. That I believe is the minor part of the problem. The major challenge is that with the paper version, the interpretation, insights, and subsequent actions are limited to certain people and the view of the current state becomes static.

On the one hand, only individuals with expertise in leveraging these maps and making sense of the insights embedded in these maps can make the best use of these hand drawings. The other aspect is that this manual process creates static current state maps, and this approach seems ancient in the context of frequently varying processes.

Extensive digitalization of manufacturing processes, and data mining of these processes are already allowing companies to uncover insights into their manufacturing processes that they did not have access before. True smart manufacturing capabilities in the context of Industry 4.0 take the concept of digitalization to a whole different level. The critical question to ask then is, in light of all these developments, is the current process of value stream mapping not archaic?

The answer to the question above is imminent to many. Many different postulations to digitalize the process of VSM have been proposed. Process mining is the underlying foundation of most of these postulations. A challenge is that though most of the proposed processes are promising, they are not all feasible in the current state. However, it is definitely possible to build a solution that automates VSM process and then uses that to democratize lean manufacturing related analytics.

In the second part of this article, to be published on 03/05, we will understand what value stream mapping is about, the critical aspects that need to be captured, and how process mining forms the foundation for many digital VSM tools. Then in the third and final part, we will explore shortcomings in the process mining approach and understand how AI can help address those shortcomings.

Stay tuned !



Leave a comment