Recent research by IDC indicates that more than 80% of companies cannot respond to disruptions within a day. In fact, on average the response time to disruptions is five days. Can you imagine, the potential loss that can result in 5 days and goes unaccounted for as simply the cost of doing business! Delays in response are direct consequences of having a S&OP solution that relies on bucketed weekly and monthly planning. Disruptions cannot wait to be corrected once a month no matter how much orchestration and collaboration you perform. As noted by some leading analysts, the world is changing too fast for S&OP. It takes a world class S&OE solution to respond fast enough to incoming disruptions and messages. Otherwise, planners are overwhelmed with messages and cannot possibly deal with them at a reasonable pace.

The fundamental problem is that the reason it takes so long to address the disruptions is because it is a manual process. It is a manual process because the S&OP solutions cannot make decisions in real-time or fast enough. And the reason why they cannot do it is not because it lacks AI, it is because it does not know enough about the operations and understand the capabilities in the supply chain with enough detail.

By far the most SCP solutions (S&OP) model the supply chain at a high level that does not reflect the actual operations. An example is weekly bucketed capacities, and not knowing what the status of operations are soon enough. In other words, the model is far from being a digital twin. You can add all the AI power that you want but if there is no understanding of the details and the choices that you may have in the event of a disruption, then how do you expect the system to make fast decisions. You might decide to take an umbrella if you see heavy clouds. If you don’t see or know that, then you cannot make that decision. There is a lot of talk about visibility. Visibility starts at home and your own operations and then extended to suppliers and customers.

To avoid decision latency, you need to solve the problem of data latency. This means having a true digital model. This can only be achieved with S&OE solutions that have a real-time understanding of the operations and have the intelligence, AI or otherwise, to make fast real-time decisions or make recommendations to the users. Thus, avoiding manual processes of scenario analysis based on a few trials out of thousands of scenarios that the system can easily handle in real-time.

AI/ML techniques can only help to make good decisions fast enough, if they have a detailed understanding of the world they are in. Otherwise, they are making decisions in the dark which leads to manual, inefficient and time-consuming manual intervention. For more information on how digital twins and AI/ML can help to make your supply chain more resilient and fault-tolerant visit Adexa.