Ep 341: Transforming Global Supply Chains with AI

This is the Everyday AI show, the everyday podcast where we simplify AI and bring its power to your fingertips. Listen daily for practical advice to boost your career, business and everyday life.

Whether you know it or not, your life is very much impacted by the supply chain. Yes, something that you probably don't pay much attention to and at all. It's impacting our daily lives in good ways and maybe bad ways, but something impacting the global supply chain and industries obviously bringing in goods from, from all over the country and internationally. AI AI is impacting their day to day operations. And I think today's conversation, whether you work around in, in supply chains and logistics or not, I think today's conversation is one that you're definitely going to want to pay attention to. All right, I'm excited to get into that, but before we do, let me welcome you to the show. This is Everyday AI. My name is Jordan Wilson and I'm the host and Everyday AI it is for you. This is your daily live stream podcast and free daily newsletter helping everyday people like you and me understand what the heck is going on in the world of generative AI and how we can learn what's going on and learn from experts like our guest today and and apply it to grow our companies and to grow our careers. So that sounds like something you want.

To be a part of.

Yeah, learning from experts, growing your career sounds good. Make sure you go to your everydayai.com Sign up for the free daily newsletter where we recap our interview every single day as well as give you all the AI news, tips and tools that you need. So if you're listening on the podcast, thank you. Make sure to check out your show notes for that link to go sign up for the free daily newsletter. All right, before we get into today's conversation, well also a reminder, go check out our Thanks a million giveaway on our website and newsletter. But let's go over the AI news for today. So Microsoft has unveiled its Phi 3.5 AI model. So Microsoft has launched actually three new AI models in its Phi series, marking a significant advancement in AI technology and providing developers with powerful tools for innovation. So the new models include Phi 3.3.5 Mini Instruct, Phi 3.5 Mixture of Experts or MOE, and Phi 3.5 Vision Instruct, designed for various tasks such as co generation, reasoning and multimedia or multimodal processing. So each model shows some impressive performance in benchmark tests, outpacing competitors like Google's Gemini 1.5 Flash and even OpenAI's GPT4O Mini. In certain areas.

So.

So this new model from Microsoft is a smaller language model and is meant to be used locally on devices. So it's available under a Microsoft branded MIT license. These models allow developers to use and modify them freely. That's a big part there, right? The difference between proprietary closed source and open source promoting collaboration. All right, two pieces of OpenAI news so OpenAI has introduced fine tuning for its largest and most powerful model, GPT4O, targeting corporate clients. So OpenAI is making headlines with the launch of its fine tuning feature that allows businesses to tailor its Advanced AI model GPT4O, using their own proprietary data. So this development comes at a time when competition among AI startups is intensifying as companies seek to leverage AI for improved business outcomes. So OpenAI's new customization option in fine tuning enables companies to enhance the performance of GPT4O by integrating their unique data sets by training the model. So the move is obviously strategically timed as businesses are under increasing pressure to demonstrate the effectiveness of generative AI and show a return on investment. Also worth noting, OpenAI did announce this earlier, but developers do have up to 1 million free tokens every single day. That came after Google just announced a billion free tokens a day. Yes, billion with a B. All right, last but not least, OpenAI has partnered with Conda Nast to utilize premium media content. So OpenAI has announced a multi year agreement with Conde Nast, the media company, allowing OpenAI to use content to from renowned brands such as the New Yorker, Vogue, gq, Vanity Fair and Bon Appetit in its AI products. So the partnership highlights the growing trend of collaboration between AI companies and traditional media outlets, aiming to enhance news, discovery and delivery while maintaining quality and accuracy and probably fewer lawsuits as well. All right, so there's a lot more AI news and more in today's newsletter. Make sure you go sign up. But if you're listening, you probably want to hear about how AI is transforming the global supply chain and that's something I'm not an expert in, don't worry. But we do have an expert to help guide us through this because I think it's something that impacts our daily lives. So I'm excited. So please help me welcome to the show. There we go. So we have Evan Smith, the CEO and co founder of Altana. Evan, thank you so much for joining the everyday AI show.

Yeah, I'm looking forward to it.

All right, cool. So Evan, just tell us a little bit about Altana and what it is that you all do.

Altana's built the world's largest data network on the supply chain probably that's ever been created. And we, we render a bottoms up, living, breathing view of the global supply chain using artificial intelligence. So think about it like Google Maps for the world supply chain connecting governments, businesses, insurance providers, global logistics companies all into one common operating picture to do a range of jobs, move goods across a border, target fentanyl precursors that are moving around the world, going to the cartels to just planning supply chains and making them more resilient and shock resistant. So wide range of use cases all on that one data and AI platform.

Yeah. And give us just, just to set the kind of tone here, Evan, give us your, who's your average, you know, client or customer and you know, really what are they using the platform for ultimately?

Well, it's a lot. So we work with governments, we work with Customs and border Protection, with the Department of Commerce, with Department of Defense civil agencies and national agency, national security agencies in the US and abroad. We work with global businesses with physical supply chains. So everybody from like a brand or a retailer to a manufacturer in aerospace or auto or apparel, so really sweeping coverage of global business. And then we work with financial services companies like insurance providers and we work with global logistics providers that are moving goods around. So we're connecting all of those different stakeholders who you know, before us really had no way of seeing the same set of facts. You know, who am I doing business with? Where are these goods going? What's the nature of these goods? We connect them all into one common operating picture where they can have that shared source of truth.

So let's do this because I'm sure there's some people, Evan, tuning in, who work, you know, in logistics and on the global supply chain. And then there's people that maybe don't for everyone else. Right. That maybe doesn't work in or around the industry. How does the supply chain, supply chain affect us every day? I assume it affects probably just about me every time that I, I leave my house and obviously when I'm in here as well. But you know, how does the supply chain impact the average American on a day to day basis?

Well, like in literally every moment of the day. Right. So like the, the supply chain is the fabric of the world's physical economy. Right. That's the way to think about it. It's, it's how all of the goods and even most of the services that we depend on get to us. Right. And it's abstracted away. It's invisible to a lot of us. We're not paying attention to it unless something goes wrong. And you know, a big thing did go wrong in Covid. That was kind of the big wake up call where it was like this outsourced just in time. Supply chain is actually pretty fragile and we need to start to know what that whole network looks like. But supply chain, it's everything. It's your carbon footprint, it's your national security exposure to China, who controls 60 to 98% of the world's critical minerals that all of our semiconductors and weapons and automobiles and everything else depends on. So it's embedded in really every part of our lives. And for a long time it's been sort of a paper based back office function, you know, just, just like make sure that things get there on time, right. And at a low cost. And now it's really rising to the fore as a strategic, in the national security context, like a theater of geopolitical competition. And as a business context, the supply chain is actually becoming a strategic driver.

So I mean, let's just kind of start by answering some of those hypothetical questions that I started out with. You know, what are some of the biggest ways that generative AI? Because yes, you know, in most industries as big as the, you know, global supply chain, we've had traditional AI, machine learning, deep learning in play for many decades. But how has generative AI and large language models most impacted the supply chain? And you know, how is it already transforming it?

So a little bit of, just kind of arc of history on this. So for a long time the supply chain, broadly like the global supply chain was paper based. So all of the transacting parties that are, you know, going from raw materials to an intermediate component to a finished product are transacting and then, you know, moving goods with their logistics providers or declaring something to the government. It was all these paper documents that humans were filling out, right? And that over the last 10 or 20 years has started to migrate to a digitized format. So it's like PDFs, it's, you know, rows and columns, but it's still really messy stuff. You know, it's in Chinese and Spanish and English. And if you've ever seen supply chain data, the listeners have, you kind of know what I'm talking about. So. So really why does that matter? Historically, it's been impossible to really know what that whole supply chain network looks like, right? So you've got to get the data somehow and you've got to process all that text information. Billions and billions and billions of records and no group of Humans can do that. It would take millions of years of people's time. So enter AI. So we can take this unstructured data that documents, you know, semi structured rows and columns and we can now apply large language models and deep learning and we can start to construct a bottoms up view of that global supply chain network for the first time. So, you know, this company is making the cotton at this location, which moves to this mill, which turns into a rolled textile, which moves into this other garment manufacturing facility, which turns into your T shirt, which goes over the ocean to this retailer. And this is the store you buy from. So that whole network of production you can now actually map out continuously and build this like bottoms up view of the physical economy. So that's, that's kind of thing one is like it's now possible for us to actually know and model the whole global supply chain network and, and you.

Know, hey, as a reminder to our live stream audience here, thank you for tuning in, Michael, Woozy, Fred, Monica, Colby, everyone else, if you have any questions on how AI is impacting the global supply chain, please drop them in now. So one thing that, you know, I'm curious about, Evan, is, you know, even what you just said right there that, you know, the industry even in the last, you know, decade or two is just paper based, right? Which, which to me it's like, I assume every single, you know, global industry has been digitized, you know, since the 90s, but I know that's not always the case case, you know, can you talk a little bit specifically even, you know, large language models and you talked a little bit about how one of their greatest strengths is being able to take unstructured data and to bring some structure to it and to be able to tell the story. How have you even seen those impacts firsthand, maybe with either customers or clients? And what does that also mean? Right? Something that is seemingly as simple as, you know, helping turn a, a paper industry with, you know, on, you know, know, unlabeled data, different languages, not, not, you know, formatted properly and organized. How can generative AI on a simple scale even just help the industry communicate better? And then what does that mean?

Well, it, it has a lot of implications, but I'll just pick a few. So if you're, if you're a global business, let's say you're a name brand retailer and you're buying, you know, consumer products and apparel and you're selling those through, through your store because of all that opacity and the data problem that I was speaking to it's historically been impossible to know how the goods were made. To my example, where did that cotton come from? And therefore you can't answer questions like what's the carbon footprint of that T shirt? Or was there forced labor involved in making that T shirt? Or am I going to have a business interruption six months from now because, you know, five or six hops away in the network that made that T shirt, it's called the value chain, we're seeing a massive disruption, There's a flood, there's a factory outage, something like that. So everything until now has been obscured in a fog. You couldn't answer those questions. So that's one. I mean, I think another one we would work a lot on at Altana is the movement of goods across the border. So for non trade nerds listening, you, you have to declare your, your product to a government. This is the oldest form of taxation, actually. It's called a customs duty. So these tariffs and duties when you move goods across the border are how governments manage trade compliance and collect revenue. So the reading of documents and the judgment about what's the nature of the goods, and I have to declare this code and therefore I'm paying a 5% duty to the government. That's happening on trillions and trillions of dollars across border movements of goods. And that industry of licensed humans that's doing that work is about a $150 billion per year business. So these are credentialed licensed customs brokers who are just reading documents, making a judgment about the nature of the goods, and helping file a regulatory attestation so we can actually look to accelerate and even automate a lot of that work. And that's profoundly value creating in the movement of goods.

And I guess where at least my mind goes when, when you say that, Evan, and I'm sure other people's does as well, is it sounds like there's just been these just almost wild inefficiencies. Right, right. In your industry, which makes sense, right, when it is literally connecting the world and everyone has different, you know, access to different technology, everyone has different resources. Right, but how does you know AI and generative AI and large language models, how does it impact the work that humans will be doing, you know, in the future? Because it sounds like so many of these jobs maybe aren't needed in their current state. Right. So I'm sure it's going to change, but how do you even see the jobs and the roles in the long term changing across the global supply chain?

Well, I'm certainly Not the first guest to tell you that it's going to change the nature of work. Right. So some of these, some of these manual tasks, certainly the repetitive rote tasks are going to be automated. A lot of the more sort of judgment oriented and creativity oriented jobs will I think be co piloted with AI. I think humans will stay in the loop for a very long time, if not forever. And so the nature of work and the supply chain is going to be around value added, creative, you know, judgment oriented and even I would say collaborative workflows. Right. So, so it's not just me in my silo typing, retyping data into a new form for the, you know, trucking carrier who needs to pick up the load or the customs broker needs to declare the goods to the government. All that stuff, all that data processing is going to happen in the background and then the humans are going to be involved in. Okay, so how do I actually make a more resilient supply chain? How do I make a more sustainable supply chain? How do I better connect not just my direct buyer and supplier, but I'm going to go to the tier two, the tier three, the tier four, and we're going to work together as a network. So those are the things that are going to be possible here with AI doing the heavy lifting on what has historically just been these document oriented processes.

I think that was a really good viewpoint on the near future and how industry is working on the supply chain might be impacted. Evan, I'm, I'm wondering, you know, because there are probably a lot of people that, that work in this industry tuning into this episode, you know, aside from, you know, oh, go check out Altana and you know, things that you offer. How can company, you know, people who are in these positions, aside from, you know, a complete overhaul of, of their data and their operations. Right, because I understand sometimes that's easier said than done. So for everyone else that's maybe, you know, struggling to kind of come into this generative AI area, who works in these industries? What can they be doing or what should they be looking at? What are some of the biggest, you know, kind of low hanging fruit, so to speak, around the, you know, logistics and supply chain industries, Are you still.

Running in circles trying to figure out how to actually grow your business with AI? Maybe your company has been tinkering with large language models for a year or more, but can't really get traction to find ROI on Genai. Hey, this is Jordan Wilson, host of this very podcast. Companies like Adobe, Microsoft and Nvidia have partnered with Us because they trust our expertise in educating the masses around generative AI to get ahead. And some of the most innovative companies in the country hire us to help with their AI strategy and to train hundreds of their employees on how to use genai. So whether you're looking for ChatGPT training for thousands or just need help building your front end AI strategy, you can partner with us too.

Just like some of the biggest companies in the world do.

Go to your everydayai.com partner to get in contact with our team or you can just click on the partner section of our website. We'll help you stop running in those AI circles and help get your team ahead and build a straight path to ROI on gen AI.

Well, I don't think anything is, is that easy where you can just, where anybody without organizational buy in can, can just get to work. I shouldn't say that there will be, there will be parts of supply chain and logistics that you know, you can feed a. Here's a good one. So if you're processing PDFs, stop, you know, feed it to one of the large language models have, that accelerate your workflow, right? Like return structured data. So that's, that's something anybody can just go do now from their own desk. I think, you know, bigger projects do require more organizational buy in. So you've got to make your company's data available to some AI system. You've got to organize that data and get it one place. You've got to, you know, either yourself or through a vendor. You've got to begin to structure and apply that data into training and inference workflows. So the, the, the, the sort of business transformation through AI does require organizational buy in.

So you know, one thing that you mentioned a little bit earlier, Evan, is you know, how AI right now can help, you know, companies or people in logistics really better understand, you know, things like, oh, their carbon footprint or you know, was there forced labor involved somewhere in this process? Right. And I'm sure that those things were all available, that information was available but was maybe harder to track down or you know, just took more people. Right. What are some of the, some of those biggest advantages that the rest of us may not necessarily see in the, in the products and services. Right. But what are maybe some of the biggest advantages or new capabilities that the industry is now seeing with generative AI and large language models?

Well, so the, the, one of the ones I'm really excited about is being able to pioneer new kinds of insurance. So if you think about global property and casualty insurance for Businesses, right. So it's everything from like cargoes to factories. You have, you're insuring against the loss of the goods or of the capital equipment. The factory burns down, I get paid. But these policies also have what's called a business interruption limit. So I'm going to transfer some of that business interruption risk through insurance. And so the global insurance industry all has this, this business interruption risk exposure through the supply chain. So if, if my factory burns down or there's some interruption there and then I can't, I can't operate. Well, you know, now I, now I have like a billion dollar business interruption loss. If my supplier's suppliers, factory burns down, I have a contingent business interruption loss. So you're looking at, you know, trillions of dollars of, of loss or rather of insured value and hundreds of billions of dollars of insurance premium, all with this hidden risk exposure through the supply chain that nobody's been able to model or underwrite until now. So as we, as you think about like where does global business go and how do you apply AI in the supply chain? And I'm a little bit talking about work we're doing, but I think it's more general because we can now process these billions of data points and model these kinds of impacts that cascade through a network. We can ensure against supply chain losses. We can model those supply chain losses so businesses can get better coverage, insurance providers can do better underwriting and risk selection. The whole ecosystem is going to benefit.

One thing you said there is just talking about business interruption. Right. And one thing that I think many of us outside of the industry, you know, I think Covid was maybe the first time at a large scale that many of us, you know, felt kind of what it means when there's large scale disruption or you know, business interruption in the global supply chain. Right in the timing of it. Right. So it was about a year or so later many businesses started to gain access to, to generative AI, right. Via whether it was via chat GPT or you know, other companies starting to use the early tech GPT technology. How do you think kind of this, this combination of, you know, so many suppliers were going through a lot during that period with generative AI and kind of these things happening almost back to back. How do you think that that has changed the global supply chain in, in the long run or maybe did it? Not at all.

Well, I think it's changing it. So the technology now exists to get ahead of these problems and not just see a problem earlier as it's happening through a network and then act on it, but to actually design and model a more resilient supply chain or value chain network. And this is a corporate board level initiative for just about every global business right now they have a supply chain resilience or business resilience initiative. So you know, global businesses are trying to solve this problem. Now that requires a lot of business transformation and new ways of doing things. And you've got to not just model it in your software, but you've got to then do the much slower work of, okay, now I've got to go work with new factories in new geographies, build compliant and on spec products and bring those online. So the business transformation is going to take years to build these more shock proof resilient supply chains. But every C suite and board that I talk to has some version of a top three priority focused on this. Because if you think that the world's going to get less predictable and more volatile in the 21st century, which I certainly do, then you can expect that supply chain dislocation, climate dislocation, geopolitical dislocation is going to increase and not decrease. This is something you have to get ahead of.

A great, great question here from Cecilia. So thanks for this. So she's asking Evan, have you seen the use of AI for supplier diversity strategy? Effectively?

Yeah, so I. Supplier diversity just for the audience can mean two different things. One is what it like basically having redundancy in your supplier base so you're not being sole sourced to a single supplier. And another way that that term is used in the industry is with respect to the workforce and the kind of human rights profile of the suppliers themselves. Was it, you know, is it, is there child labor? Is it an all male workforce? Are you sourcing from veterans or people of color? So those are used in, in two different ways in the industry. So the answer to both is yes. And I think it's easier said than it's, it's easier to accomplish on the first category where you can look across a whole network and you know, tier two, tier three, tier four, and then you can start to model those single points of failures and those dependencies. In the case of, you know, the sort of firmographic profile of the companies you're doing business with, that really requires a lot of underlying attestations and data points to come together to build that profile. Right. So are they telling me the truth about the nature of the workforce? So that's where AI can play a role. You know, you can scrape LinkedIn, you can look at the, you know, the the people who work for the company. You can look at different regulatory filings, you can look at different public declarations that are made about that company or you know, chatbot. So there are AI companies and strategies and products built around profiling the workforce and the DEI sort of posture of these companies and then framing that in terms of supplier diversity.

So one thing I think that we can all see the huge possibility and the tremendous upside of generative AI when it comes to the global supply, supply chain. But maybe are there unexpected ways that AI might disrupt, you know, traditional supply chain dynamics maybe that we're not thinking of? Maybe companies are a little, you know, too quick to try to sprinkle AI on everything without even having a strategy. So maybe is, is there a rush to this that companies maybe are, are missing, that are maybe creating some new vulnerabilities that we might not have been able to predict?

I, I, I would say not yet. Just my, my first hand experience with these organizations, you know, in supply chain logistics, sustainability, compliance, they tend to not be very fast moving. So I, I don't think we're in danger of some runaway unintended consequence anytime soon. But I think, you know, if you fast forward into the future will be interesting is a lot of the routing and what's called planning, you know, coordinating supply and demand across buyers and suppliers and doing that across the network. You can imagine a future where that's pretty lights out. Like the, the analogy would be Google Maps and Waze, right, where it just starts to automatically route you. So that's a future that will probably exist. And in that future where there's an overlay of big disruptions, dislocations, I can imagine some unintended consequences and some fragilities that the thing we built, the automation we built to adapt to the thing is actually going to create some of those same unintended consequences.

Evan, maybe what's a viewpoint that you have on, on you know, AI's impact on the supply chain that maybe some of your peers might look at you and be like, oh, that's interesting, I wouldn't have expected that out of you.

So I have two. So if you're running a global business, if you're doing business across borders, then if you're not reconceiving of your business as a value chain network. So in other words, it's not just who I sell to and who I buy from directly, but it's my supplier, supplier, supplier back to the soil. It's my customers, customers, customers all the way to the end use then you're not going to succeed in the 21st century. I really believe that. So the world is changing in these ways that we're going from a outsourcing paradigm in globalization to this new thing that people are talking about in terms of deglobalization, near shoring, resilience. You have great power competition that's really refactoring the whole supply chain network because of China, us, US competition. So if you're not organizing your business as a active managed value chain network and you're just, you know, in the traditional mode of I have suppliers and I have customers and that's my worldview, good luck. So that's, that's one take and I think the other one. You, you, you told me to give a, a gen AI hot take. So the, so I was invited to the US Senate to give testimony on AI safety. I was part of that group. And my contribution to the discussion I think surprised a few people in the government. So it was a bipartisan panel and it was very well intentioned. It was actually a great conversation. But one of my contributions was I actually think that the way we're having this conversation where it's all about defending the public and the government and government data and public data from abuses of AI and from bias and other, you know, issues that come up is actually the wrong conversation to be having. And the conversation we need to have as the United States is how do we win the AI race in the 21st century? Because these are, these are going to be the commanding heights of, of geopolitics and the government is one of the holders, largest holders of data. And if you think about the positive applications of AI in education, in healthcare, in medical research, oncology. Right. In being able to manage global supply chain so that, you know, Department of Homeland Security is one of the biggest data holders in supply chain in the world. So I would, I would like us to reframe this and say how can government proactively cooperate in partnership with industry to leverage data in order to achieve these beneficial, safe, positive, net, net positive outcomes of artificial intelligence and ensure that we win the economic and geopolitical competition of this century?

Such, such a great call out there. Yeah, we're looking at you government work, work with private industry a little bit more. Make the industry better for, for everyone. All right. I don't know about everyone else. I am feeling much more informed now. But Evan, maybe what is your one biggest takeaway? You know, because you probably piqued a lot of people's interest. So maybe for those who work in the industry. What's your one biggest takeaway on how people can start using generative AI in large language models to really move their piece or their part of the global supply chain forward?

So I would say you can now know and manage your value chains, raw materials, intermediate goods, finished products, customers at whole chain. It's possible to know it and it's not just something you inherit blindly.

Love it. Love to see it. All right, this was a great one. I, I think, Evan, that you really just, at least for me even, right, I didn't know a ton, I didn't know a ton about, you know, how AI is impacting the global supply chain. But I think you just helped walk me and the rest of our audience through it. And we thank you so much. So thank you, Evan, so for joining the Everyday AI Show. We really appreciate your time.

Thanks, Jordan. This was fun.

And hey, as a reminder everyone, there was a lot there, a lot of value. So if you haven't already, make sure you go to our website, your Everyday AI dot com. We're going to be recapping today's conversation with Evan, putting in a lot of additional resources and some takeaways for you because yeah, we covered a lot there. So make sure you do that and then make sure you join us tomorrow and every day for more Everyday AI. Thanks, y'.

All.

And that's a wrap for today's edition of Everyday AI. Thanks for joining us. If you enjoyed this episode, please subscribe and leave us a rating. It helps keep us going for a little more AI magic. Visit your everydayai.com and sign up to our daily newsletter so you don't get left behind. Go break some barriers and we'll see you next time.