The hype machine for AI and its cousin Machine Learning is on full blast, that’s for sure.
It’s all AI all the time, even if there remains a lot of confusion as to what the real uses cases are in the supply chain.
Gilmore Says.... |
 |
| The Journal says that companies using Uber Freight’s chatbot can query which routes often have delays and how their service levels compare with their peers. |
 |
What do you say? |
|
| Click here to send us your comments |
| |
|
|
|
But the hype of AI surely contains many truths, and the potential sounds pretty good to many CEOs happy to have a virtual machine replace a human being to perform a task.
A number of years back now, our former columnist Dr. Mike Watson of Northwestern University wrote that soon virtually any advanced analytic development would get categorized as AI, and he seems to have been very right.
Some relevant data points:
In Gartner’s recent list of the top supply chain technology trends of 2024, one was not just regular AI but something called Composite AI. What’s that? Gartner said it involves “the combined application of multiple AI techniques to improve the efficiency and accuracy of learning to broaden the level of knowledge representations, and ultimately, to solve a variety of business problems that drive supply chain performance improvements.”
More on that at some point, after I study up.
But not only did AI of course get a technology trend on its own, it was cited by Gartner as part of other technology trends as well.
I attended a webinar this week on use of AI in the supply chain. It was OK, but the use cases were at a very high level, such as predictive analytics in demand planning software and opportunities in warehouse automation, but without a lot of detail – and apparently not backed with much in the way of successful real-world examples.
A recent article by Liz Young in the Wall Street Journal on AI and the supply chain tried to address that.
First, the article notes that in logistics, initial use of AI has included deployment of chatbots that can handle customer-support functions such as tracking shipments and booking loads.
Nice but not exactly breakthrough capabilities.
It gets better.
The Journal cites the example of German software firm Celoni’s work with food company Mars to use generative AI to combine truck loads to reduce freight costs or achieve faster delivery.
Previously, the article says, “Mars had manually evaluated factors such as the weather to determine which shipments could be combined and whether it needed to use refrigerated trucks for its freight.”
 So I am not sure what to make of that example. Traditional Transportation Management Systems (TMS) have been combining loads using true mathematical optimization and many constraints (e.g., maximum number of stops for a truck) literally for decades. Weather as a factor has been developing for several years.
All without AI.
So the questions are obvious: How much better of an optimization can be achieved with AI versus linear programming? Does AI handle dynamic changes in execution better or faster than traditional TMS? Are these performance improvements large enough to justify the cost of replacing the TMS you have with an AI-powered one?
The Journal piece also cites the example one unidentified company that is using AI to compare its contracts with suppliers against its invoices to make sure it isn’t missing out on rebates or discounts.
That technology replaced a previously time-consuming manual process of doing the same thing.
Ok, that seems like right out of the AI standard text book.
The article also references apparel retailer ThredUp that it says has been using AI in its distribution centers to improve throughput and productivity -but alas that’s all we get. I would very much like to understand how AI does this better than an advanced Warehouse Management System without AI.
Another one from the Journal: Uber Freight and FourKites, which tracks freight shipments in real time, have both created chatbots that allow shippers to pose conversational questions about their freight,
For example, the Journal says that companies using Uber Freight’s chatbot can query which routes often have delays and how their service levels compare with their peers. Uber say it is working on building the capability to make recommendations for shippers to reduce costs and speed up shipments.
Ask a question and quickly get a correct answer back. That is seems clear will become ubiquitous across virtually supply chain process and that will change the world.
So I think I will end it here. I see it like this:The AI hype is far outpacing the reality. The gains in many cases will be incremental at best. Some applications are developing
But the supply chain computers will be talking to us soon enough.
What is your reaction to the current state of AI? What would you add? Let us know your thought at the Feedback section below.
Your Comments/Feedback
|
