What are the use cases of AI in Inventory Management in OEMs?
715 views1 Upvote1 Comment
Sort by:
Content you might like
Sales/Marketing23%
Competitive Analysis35%
Product Innovation23%
Training8%
Business Forecasting6%
Other (please note in the comments)1%
Once a year56%
Twice a year35%
Every quarter8%
I would look at the following use cases.
AI if done correctly can be good at statistical analysis and in the case of OEMs that can be used for determining correlations/forecasting. You could build a tool to help forecast the best stock level for the firm and hopefully get some value that way.
The challenge is that there could be a push-pull between the machine vs. the humans involved. At its best you'll get a better answer, at its worst they'll work against each other. A machine will stick to the numbers (usually) a person will stick to the story. We humans are hard wired to find the story and if we can't find it we'll invent it. We've started to train AI to do that (AI hallucinations) which is pretty scary. We may have built a tool to exploit a significant weakness that we have.
There are things that you will need to keep in mind should you go down this route. You will have to integrate with other areas, i.e. if you sell 20 year old whiskey and you see an increase in sales today such that you need more whiskey, your lead time for 20 year old whiskey is 20 years. You may increase production today only to realize that in 20 years there is a slump in whiskey. Similarly there will be a lead time for your inputs. If you're a bottler of 20 year old whiskey and the whiskey is an input then your lead time for supplies is 20 years. There are other considerations as well such as the fact that you may be regionally limited, Burgundy and Bourdeau are only so large. You can't just "grow" the area.
You might be able to use AI to try and forecast the prices for inputs etc. but you should have some override. Even though an input is cheap you may have other considerations such whether that capital may be better used elsewhere, shrinkage, loss, theft, etc. All of this is to say that you won't be able to sell the AI only as a service, you might be better off including some human knowledge as well. The AI may give you a solution but the real test will be whether adding some humans to the process makes things better or worse.