Email
sinha.rk.dr@gmail.com
Is the use of AI/ML in Supply Chain overhyped?
We hear a lot about using AI/ML in supply chain design and planning. Many companies are experimenting with it. However, the results are a mixed bag. How should we approach it?
At the outset, we should remember that AI/ML is just a technique which has a large potential to help our supply chain processes. However, if the process design itself is suboptimal, AI/ML can do only a little to improve it. In my opinion, most AI/ML implementations suffer from this shortcoming. It is vital to design the supply chain well and streamline the processes before using advanced techniques like AI/ML.
Once the design is improved and processes are streamlined, there would be a few areas where the current techniques become obstacles in further improvement. These are the apt places to see if AI/ML can help. Let me explain with an example.
We know that supply chains can be made more responsive if we decipher the patterns in consumer demand fast and at the granular level. These should be done wherever customer facing inventory buffers are kept. These product-location combinations could easily run into thousands and millions in a medium sized CPG company. How feasible is it to decipher these million patterns with current techniques without losing time? AI/ML comes in handy in such cases.
This is just one example of how AI/ML can be used well for demand sensing.
We should look out for similar opportunities in other areas if we want AI/ML to deliver significant benefits in implementation.
Would love to hear your thoughts on where you have experienced significant benefits through AI/ML implementation…