The views expressed here are solely those of the author and do not necessarily represent the views of FreightWaves or its affiliates.
This installment of the AI in Supply Chain series (#AIinSupplyChain) is the last article in the six-month-long weekly series that began on July 7. With one or two exceptions, every week since that first column, we have studied an application of artificial intelligence and machine learning in an area of supply chain, with an emphasis on how the innovation highlighted in each article solves customers’ problems.
For this final column, I want to synthesize three lessons I have learned over the course of the past six months. Starting in January, I will feature an article on #AIinSupplyChain in this column about once every month or so.
Customers just want their problems solved — and technology that’s easy to use
One theme that has arisen consistently during the conversations I have had is that the average customer is more interested in an efficient and cost-effective solution to that customer’s problem and LESS interested in the details and specifics of a particular technology product.
That said, there’s also the danger that insufficient sophistication in understanding the basics of AI and machine learning leaves customers at risk of wasting time and money. I had to pull one story because the more research I did, the more questions arose about how well the product works. Since I could not devote the resources of an investigative journalist to finding out what was going on in that case, I had no choice but to kill the article.
This leads to the conclusion that, before a company, or any organization, implements an AI and machine learning initiative or product in its operations, it must first run at least one small and relatively inexpensive pilot in an environment that allows it to assess the reliability and accuracy of the product it is considering relative to the status quo.
Only once such pilots have been run and deemed successful should such a product be adopted more broadly.
In some instances quality of data is a stumbling block, but it is also an opportunity
It seems obvious from the conversations I have had that, for certain forms of AI and machine learning, the availability of data of sufficient quantity and quality for decision-making remains a major stumbling block to the application of AI and machine learning for the solution of most problems in the supply chain; old data is messy and is stored in silos in non-uniform formats. But, that old data is also necessary for training AI and machine learning models and algorithms. Companies that have a history of treating data as a strategic asset are in a better position to adopt AI and machine learning than their peers that have not.
I have come to the conclusion that AI and machine learning startups that develop proprietary and scalable approaches to solving this problem for their customers might have an edge over their peers.
The products and platforms that will survive must solve problems at 3 levels
In any company, decisions are made at three levels: strategic (long term), tactical (medium term) and operational (day to day). Companies must select the AI and machine learning startups that they partner with to match the scope of the problem they seek to solve. Also, it is important to ensure that a system that solves problems in one category can provide outputs that facilitate decision-making in the other two. For example, a pick-and-place robotics system that is implemented as part of a customer’s day-to-day operations must be able to provide data that supports tactical decisions about warehouse design, which in turn informs strategic decisions about whether capital should be invested in building more warehouses or if that capital should be spent in some other way.
Another thing I have learned while writing this column is that, indeed, The Economist exaggerated its claim that only the internet giants are benefiting from AI, and that the benefits for other businesses remain too far out in the future to inspire confidence.
There’s a lot happening that is perhaps still too early and nascent for publications like The Economist, the Financial Times, The Wall Street Journal and others of that stature to take notice — that is the uncertain and uncomfortable space in which early-stage investors like me spend all our time and energy.
At one time, today’s internet giants were also a mere glimmer in their founders’ eyes, and they too were ignored and deemed irrelevant by these same publications.
There is no shortage of opportunities for AI and machine learning to be applied to the solution of big problems in global supply chains. We will continue to follow and report about what’s happening.
Send me tips and suggestions.
If you are an academic working on such problems, do not hesitate to get in touch with me. If you are a company setting out on this journey and want to compare notes, don’t hesitate to connect with me either. I am really easy to find online, and always eager to talk to other people who find the intersection of supply chain, technology and early-stage innovation as fascinating and intellectually rewarding as I do.
If you are a team working on innovations that you believe have the potential to significantly refashion global supply chains, we’d love to tell your story in FreightWaves. I am easy to reach on LinkedIn and Twitter. Alternatively, you can reach out to any member of the editorial team at FreightWaves at email@example.com.
Dig deeper into the #AIinSupplyChain Series with FreightWaves
Author’s disclosure: I am not an investor in any early-stage startups mentioned in this article, either personally or through REFASHIOND Ventures. I have no other financial relationship with any entities mentioned in this article.