Insight Partners
More By The Author
Back
Article
Q&A with Insight Partners’ Arjun Mehta and Exiger’s Brandon Daniels: AI’s impact on supply chain risk management
Insight Partners | November 18, 2024| 4 min. read
This article was originally published by Exiger and does not necessarily reflect the perspective of Insight Partners. Insight Partners is an investor in Exiger.
Ahead of ScaleUp:AI ’24, Exiger CEO Brandon Daniels and Arjun Mehta, Principal at Insight Partners, joined for an in-depth chat on AI impacts on supply chains as part of a recent series of LinkedIn Live Soundbytes. This series of discussions sets the stage for the global conference bringing together leading AI builders, executives, luminaries and innovators.
Some insights from the full conversation are excerpted in the Q&A below, and you can watch the full session below:
Q: How is AI impacting supply chain risk management? What are some of the use cases that you’re seeing as game changers?
Brandon Daniels: One of the things that generative AI is doing is it’s making it so procurement and supply chain people are able to bridge a skills gap. For instance, when we flag that a supplier to our customers has a risk related to potentially something like human trafficking or forced labor, the first thing that they want to understand is how does this impact us? Do we have current products that are being used that are using this supplier? Are those products prioritized? Do we have products that are alternatives to that supplier that we can utilize?
So the first thing it’s doing for procurement and supply chain people is it’s helping them to understand how risk can impact their supply because their genius is in performance, in schedule, in cost, in creating value out of supplier negotiations; but they don’t have the background and training in compliance and regulatory standard practices. Artificial intelligence is making it so that we can bridge that skills gap.
Further reading: The next industrial revolution: How software is shaping the future of production
AI is also helping people to do more with less. So people who are trying to just purchase a set of software or components or computers from a vendor, they have this issue where they’re trying to assess for risk — but there just may not be enough resources to do it. So, for example, I can still assess my supplier for the things that I and my team has built up skill sets and resources to do, but it’s also helping us extend outside of our domain and to essentially add virtual resources. These AI agents that we’ve built, these domain-specific capabilities, they are basically additional resources in your department helping you to solve these challenges that are net new, and you otherwise may not be able to tackle.
Q: So, how do you think about the use of AI and hallucinations in our end markets and where is human oversight necessary sometimes to kind of manage what is a nuanced supply chain risk analysis?
Daniels: When I think about hallucinations, I think how do you combat this issue of trusting the solutions? Well, the way that we’ve looked at that is we try to compete models, we compete models in order to ensure that we’re getting closer to the best answer. Because if we can identify the places where we’ve got that source of truth and all of the models point to the fact that this is a fact, then we can move toward high probability, high consistency of decision-making. And where there are conflicts, that’s where we need to insert a human.
And that’s what we’ve been doing with our models, and it’s been beneficial. We compete, for instance, our AI adjudication of a risk with what we’re seeing coming from a generative model and then comparing the two and coming out to a decision. And if those decisions match, then we know that we can just print that forward to our customers with high confidence. And that’s where this 10x of capability and capacity really comes out and provides results.
“We’re using AI first to create the data that you otherwise don’t have, because your organization no longer stops at your doorstep.”
Q: What makes your approach to AI different from what others are bringing to the market? What’s unique about Exiger’s deployment of AI?
Daniels: Our differentiated approach is that we’re helping you to create the data you don’t own, so that you can see that third-tier supplier and you can see the facility that they’re producing, let’s say, a polymer out of. We’re creating the same data that you would want in your ERP in your system for that supplier. So we’re using AI first to create the data that you otherwise don’t have, because your organization no longer stops at your doorstep.
The second thing is then we’re helping you to utilize that information inside of a multi-tier orchestration platform. So in most platforms that exist today, you’re using that direct supplier collaboration and control to actually effectuate change. What we’re doing is we’re making it so that you can actually orchestrate, with a forecast demand plan, your multi-tiered supply chain. So you can look out 1, 2, 3, 4, 5 years at your demand plan and see where your tier-five supplier just simply isn’t buying enough titanium to meet your need. That’s where our AI is really helpful.
And then also that last thing of bridging the skills gap. So we boil everything down to risk scores, and you can set your risk appetites like it’s still the human saying, this is the risk level that I can’t tolerate. But we boil everything down to risk scores so that you don’t have to be an accountant to know that there’s a financial health issue with a vendor.
Q: What will AI look like in supply chain risk management in five, 10 years? If you put on that lens and think about the future, what does that look like to you?
Daniels: If I have my druthers, it will break down the silos between engineering and finance and procurement and ops and compliance, so that anyone can ask the question, “How does this impact me?” Because that’s the promise of AI. Our XSB tool set allows you to take a part, break it down into all of its part attributes, and then find other suppliers in your environment that could manufacture that same part. Imagine that as you build a new product, finding all the potential supply chains that you could start that you don’t have access to right now.
Imagine a world in which finance can look at those materials and say, “Oh, wow, we’re actually buying 25 billion tons of aluminum, and we also have to reduce our carbon emissions. There’s a supplier in aluminum that we could give a contract to. They’ll give us a 7% discount, we can give them a 1% discount. We can give them 1% of that discount back to implement an ethical standards policy for all of their underlying mining and ore extraction. And now we’ve been able to track and trace our social responsibility and environmental responsibility to our underlying supplier and reduce our total landed costs by 6%.”
That kind of universal use of supply chain data is where AI is taking us in the future.
