Today. I'm excited to talk about how we can, change and revolutionize supply chain, management using AI and ML. I'm Shodonna Ashway. I'm, product and program leader.With, twenty years of experience, bridging business and technology. Let's jump to it. Why supply chain resilience requires prediction? We are all aware that supply chain are under, immense pressure. And from goal global disruption to, shifting customer expectation, resilience isn't, optional anymore. It's a strategic necessity. But traditional systems, aren't built for pace or complexity, we face today. Let's look at the global chip shortage during pandemic. Automated occurs underestimated, like, the risk and couldn't adapt quickly resulting in billions in lost sale. Traditional systems couldn't flag, this risk early enough because they lack predictive insight and real time signals from upstream supplier. This expose how reactive and fragmented many supply chains still are, even for most advanced manufacturer. Why traditional tools fall short?

System like Coupa and Rapid Rating, it we use in supply chain, many companies use, give you useful insight, but they rely heavily on backward looking data. That creates blind spots. They don't capture context, from unstructured sources or for, disruptions before they happen. Take a company using Coupa and rapid rating. These tools provide supplier health and scores. But when one supplier's ESG score drops due to a factory labor violation, for example, it doesn't trigger any alerts because the system is only reviewing quarterly reports. On a structured signals like new headlines or audit findings are missed entirely. As a result, the business caught off guard by, reputational fault and has to scramble for damage control.

So at this part, I let me see if I can see if there is any okay. So let's continue, from evaluation to prediction because usually the supply chain is evaluated. This is where predictive modeling shines. Instead of only evaluating performance after the fact, we can forecast trends, catch early warning signs, and shift from reactive firefighting to proactive prevention. Traditionally, a supplier would only get flag after missing multiple deliveries or bringing or or bridging a contract. But what if we could see the signals earlier? For example, in one project, a supplier, on time delivery rate dropped just as slightly. But paired with an, uptick in supply tickets and a slowed, invoice cycles, a predictive model caught this pattern and flagged the supplier as at risk weeks before a critical failure occurred.

That early insight gave the business time to engage, course correct, and avoid costly production delays. Now artificial intelligence and machine learning in action can be like predictive risk scoring. You can use machine learning models that combine a structured data with, conceptual insight from reports, communication, and other sources. This enables us to assign dynamic risk scores and surface actionable insight earlier in the decision process. In recent implementation in a recent implementation I was involved, we layered a predictive model on top of a structured data from an ERP system supplier, emails, and audit findings. The model identified that the regional logistic provider had a pattern of failing during weather related events. The system begin assigning dynamic risk scores during a storm seasons and even suggested rerouting orders. That's the kind of adaptive, intelligence that adds a real business value.

Now with generative AI and LOMs, we can extract insights from ESG documents and, audit reports and even emails. We can simulate supply chain shocks and generate mitigation strategies in real time. This takes intelligence to the next level. LLM can be used not just to extract data, but to simulate disruption scenarios. For example, by feeding a hypox an event like a port closure or regulatory shift, the model can surface downstream effects, and trends strategy or even generate proactive communication. This turn this turn artificial intelligence from a reporting tool into a decision making partner. Now let's see how you can use AI. This diagram shows how LLM can be used to simulate disruption scenarios. Let's take a real, example.

Imagine a major port in China is suddenly shut down due to a COVID resurgence. You, feed that scenario into the alarm. OOM. Let's like, for example, a prompt, a tier one supplier in a city that is unable to ship component for two weeks. What are the potential supply chain impact for a North America EV manufacturer? The LOM can analyze, patterns from past disruption, ESG reports, and operational documents to predict likely impacts such as production delays, increased freight cost, or regional shortages. It can then suggest mitigation strategies, like rerouting, shipments, actively alternative supplier or adjust order quantities in key regions. This allows leaders to basically proactively prepare, not just react, using AI to simulate real world complexity in a matter of minutes. So here I have, like, a prompt.

You can like, I thought that maybe we can test this. So this is scenario. I I wanna give this to Check GPT. So you also can do that. Like, tools like Check GPT can be used today to simulate disruption scenarios without building a cost MLM. By providing a structural prompt, and business contacts, teams can use existing models, for rapid what if planning early signal and or even drafting contingency reports. For deeper automation or domain specific logic, organization can layer that GPV on top of a structured data systems to scale, this scenario when you need to build your own model. Okay. So let me see how much time they have. Okay. I, so we don't have that much time. So I'd like to say that AI for strategy advantage, like, it create cost reduction, improve compliance, and support, supplier diversity goals.

And we always can slice, like, start to be the slice delivery and look at the MVP and just create a team of slice and then a scale. So always start small and then a scale. Start with the use case and then a scale. So predict the unpredictable is the future that is not about more data, but it's smarter insights. And using AIMO and GenAI will define the next era of supply chain.