AI’s recent and rapid expansion into all parts of the supply chain greatly expands our ability to grab new data from places you’ve never been able to grab data from before. Often, the above is due to lack of time and not having knowledge of the existence of data that could be used.
In addition, you probably have a lot of data waiting to be used just within your organization. If it is useful, structured data that explains supply chain performance from various vantage points, such as Reliability, Responsiveness, Agility, and other standard metric hierarchies found in SCOR, then AI can be put to use internally as well. Once collected, it can also be used for benchmarking purposes.
Even better, it can a) merge internal insights with b) external data explaining important, critical, need-to-know aspects such as what is going on in the economies your supply chains serve, and c) merge that with your customer data and insights. You now have a holistic look at what is going on, what may be percolating, and what may be coming your way in the future.
All the better to plan with.
All the better to make decisions with.
All the better to understand what headwinds your revenue faces from a demand aspect.
However, we know that Newton’s third law holds true in supply chains: for every action there is an equal and opposite reaction.
You should be looking at that opposite reaction, too. In Supply Chain, it’s called Risk.
Here’s where we get to the Board of Director concerns.
The drive is to use qualitative and quantitative metrics to describe high-level strategic Risk. The inputs are operational performance measures. From your supply chains. If you use industry standard metrics in your supply chain, you’ll get to compare your risk against others.
But what risk?
Here’s a simple place to start.
Before you look for AI, determine if you want to be “in the loop”, “on the loop”, or “out of the loop”. For busy supply chain professionals, “out of the loop” sounds pretty nice. That is until you consider that “out of the loop” AI is what ChatGPT is. It goes wherever it wants, when it wants, gathering what it evaluates is meaningful and, in some very spectacular ways, makes up “facts” as it considers made up facts to be “needed”. Big risk here.
While here in the U.S. we can let the courts decide if disinformation from AI is protected under free speech, you don’t have that luxury in your supply chains. Software suffering from “hallucinations”, in scientific terms – the “data” returned is garbage, does you no good.
If you choose to be “in the loop”, you have the most control. AI goes out at your command to do your bidding. Less risk on the AI side, probably – dependent on what you are doing with it. Definitely more time on your side. And that may lead to risk around other activities that are not being attended to.
If you choose to be “on the loop”, you train AI on what you need it to do, review its input, and then provide feedback as to the value and usability of the data and synopsis provided. Over time, the AI is able to perform searches, gather data, and create decision options for you to review for use in problem solving – without forcing you to wonder which options you can or cannot trust.
For example, in procurement, you could use AI to survey the potential supplier market in more detail that you have time to (because that is AI’s only job) and integrate the market info with financial, social, regulatory, geopolitical and economic data that helps you understand better how to build a resilient supply base. A resilient supply base using suppliers you may never have found on your own with the limited time you have available for searches. Or perhaps not using suppliers you may have thought there were no substitutes for. Once the AI software is taught what to pull and deliver, it can continually search for new entrants for your supply chain.
You are definitely “on the loop” when something changes, anomalies occur, or a disruption changes the search parameters. Constant change is true in the AI world just as it is in the supply chain world.
How do you want to handle the search for answers to that change?
And is it the AI that is creating the change?
For example, AI will create new algorithms. Humans won’t really know what those algorithms are. Plus, the AI logic used is not human logic, but programing-based logic (which includes biases embedded, however unintentionally, by the original creators). There is some great science fiction exploring this but know that you will lose control if you don’t have adequate monitoring in place if your AI is “making decisions” . Even if those decisions begin as just coordination based on search findings of how your supply chains are running.
An article from Thomas.net discusses that “AI within the supply chain should be carefully considered within a comprehensive risk, contingency, and mitigation matrix”. You already know how to map risk using a risk matrix.
This is more of the same.
But
You must include data analysts, IT and data governance personnel, as well as your supply chain experts, when working through the risk analysis. Time spent here is time well spent.
Having already anticipated a risk, or a similar risk, contingency plans can quickly be executed. Consequently, saving your supply chain’s performance and your company embarrassment, or worse, loss of revenue.
And about those Boards: as mentioned before, if you provide them with standard, hierarchical supply chain metrics, they can evaluate the very high-level business strategies, integrate your operations-based risk insights, so that combined, the business is able to develop a robust understanding of capabilities, early warning signals, and potential expansion of your market share.
#AI #supplychain #risk #supplychainrisk #metrics #SCOR #riskmatrix #riskmanagement, #ProcessandStrategySolutions #CynthiaKalinaKaminsky
Dr. Cynthia Kalina-Kaminsky helps companies of all sizes and provides public training to those who want to successfully transform supply chains, integrate digital technology solutions, and avoid Risks that jeopardize your strategic goals, regulatory requirements, and customer supply chain performance.
Comments