[00:00:10] Speaker A: Welcome to the Loadstar. In this sponsored podcast, we're busting some myths about AI in the supply chain. Is the hype. Well, just hype. What does AI integration even mean? And how can any of this help you better serve your customers and stakeholders? Guiding us on this journey today is Raft CMO Greg Kefer.
[00:00:32] Speaker B: TMS planning systems, the ERP systems are starving for what I call executional data. They're starving for it, right? And oftentimes they're not getting it for months after the product was delivered. And then after that they realize, oh my God, we lost money on that because all of this happened that we couldn't foresee, like the Red Sea or rail strikes. Imagine if that was all happening in real time and AI could be replanning constantly.
[00:01:02] Speaker A: Hello everybody, I'm Mike King. You're listening to a Loadstar podcast. In this episode, we're going to be trying to unravel exactly what AI means for supply chains and for the shippers, forwarders and carriers that do their best to keep everything humming along. Well, look at what companies are already using AI for. And what might all this new tech be capable of in the future? Can indeed AI be like a GPS for both cargo and company efficiency, always finding the best route for the best possible outcome. Well, we'll try and find out. As trailed helping me separate AI myth from AI reality today is an executive who has held leadership positions at global enterprise software companies, cloud tech startups and advertising agencies, including CMO roles at Slync and Lifelink Systems, AI focused startups that disrupted the status quo in the supply chain and healthcare industries. He also spent 17 years at the helm of GT Nexus, where he led all marketing and communications functions as the company grew from startup stage through a $670 million acquisition. He's now the CMO of Raft, which prides itself as the world's largest logistics AI platform. Greg Kefer, welcome to The Loadstar podcast.
[00:02:17] Speaker B: Thanks Mike. Good to be here.
[00:02:19] Speaker A: You're welcome, Greg. Where do we start? There's so much noise around AI in the supply chain and forwarding industry, and it's so difficult to block out the Blarney and zero in on how this new technology is actually changing our industry. So let's try and find some historical context around what AI is in terms of its disruptive power for our industry, if we may. You tell me, is this a new Internet? Is it a new cloud? Is it a new blockchain? Where does it sit in terms of how much of a game changer it is?
[00:02:52] Speaker B: Yeah, I think it sits in the realm of the Internet or the PC, the semiconductor. It's of that class of disruptiveness, if there's such a thing. I don't think it's maybe a blockchain. I mean, I think one of the issues with blockchain was it was very hot and hyped, but nobody can understand it. You don't see really a viable blockchain market out there today. I mean, yeah, there's bitcoin and that kind of stuff, but nobody understood it. If you remember back in the early days of Intel, there was a thing called Moore's law, where basically Moore thought that two times the components every two years without adding cost to its ship, so that it was basically to show that technology would just continue to get better without getting more expensive. That actually held true, but now look at Nvidia, which is the third most valuable company on earth behind Apple and Microsoft. Why is that? Because their chips have increased computational capacity by a thousandfold.
You've gotten to this point where technology can actually learn and self-code. I think the big one is it can talk, it can read, it can behave in a more humanistic way, where we're moving away from keyboards and menus and mice. And I think that's what's really, really got the world's attention, is that AI is a broad thing. It's a handle, kind of like cloud was for the Internet. But below that, there's generative AI, there's large language models, there's small language models, there's deep learning, there's robotics, which manifests into things like self-driving cars. And write me an essay, chat, GPT, or dolly, draw me a picture. On and on and on and on. Right. So it's just also very early, and I think we have to remember that before we get too much disillusionment, that it's not what it was expected to be.
[00:04:39] Speaker A: Note to self, I need to know more about Moore's law. Clearly. I'll look that up later. So, Greg, what does all this then mean for our supply chain listeners today? When I talk to people, they'll say AI. It's about automating jobs. It'll improve productivity, but often without really specifying exactly how. Some also say AI tech will root cargo more efficiently. Others say you can't cut out the human element from a business like forwarding. So I guess my question to you, Greg, is what are AI's applications to global supply chains? And specifically the business of moving cargo from A to B, or lots of cargo from lots of suppliers through hundreds of movements to create a product and get it to market. If I may throw that at you like that.
[00:05:25] Speaker B: Yeah. So there was a lot in that. I would say that, first off, I think AI is finding its way into the supply chain in many different places that, you know, I think whether you're doing planning or you're running a network of a fleet of trucks, there's a lot that AI can bring to the table. I think when I think of our audience and the nature of this podcast, and you mentioned freight forwarding and moving freight, I mean, look, this industry, from a workflow perspective, still relies on email, right? I mean, email is pervasive. So is Excel, right? Why is that? Why isn't there software that's doing all of this execution of freight movement? Right? There's no shortage of technology out there. There's a lot of great TMS systems. And I did some research a couple months ago, and I think I can safely say that north of half a trillion dollars has been invested in technology to solve this problem. Yet here we are in email. And why is that? Yes, it's complex, but the real problem that the industry, this industry must deal with is that it's not a single company business process. You're dealing with a network of other companies, whether it be suppliers at origin, brokers, freight carriers of different modes, warehouses, et cetera, that all have their own systems, their own processes, their own data standards. So this whole idea of give me the digital information that I need, the way it's been attempted has been, and here's the way I need you to send it to me, I need you to follow my protocol or this standard. And it could be everything from how a document, a bill of lading is structured, whether it's an EDI file or a PDF, to what's the location code that we're going to use, or what's the time zone we're going to use.
If you multiply that across a supply chain of a large freight forwarder, one of the top 50 that are moving tens of thousands of containers or air freight cartons a month or a day even, it's a substantial problem. The issue is that everybody uses email. And the problem with email is that means humans have to go in there and they have to look through things, read it, and then interpret it, and then re key it into a different system. And the minute that happens, there are three things that go wrong. One, it's often missed because you're...I had a customer at Slync that were getting 40,000 emails a week. And of those 40,000 emails related to freight moves, probably 10% meant something. The other 90% were updates on the logo of the company or an event they're going to. And so they have to find that, understand what's important, what's not, and then rekey it. And of course, if you push an eight versus a nine, the machine, as far as the machine is concerned, it didn't happen. So this is a problem. And so you've got armies of people that are doing this very administrative, low value job, which is not a great job, by the way. So you look at a staff of an army of individuals that are just doing data entry all day long, and it's not that good. And if you have 80% of your data right, and 20% is wrong, you really can't run a supply chain that way. Right? This is how it's been going, and it's still going in the year 2024, which to me is kind of shocking because I've been a technologist for a long time, but it's also super hard and complex. So where AI can really begin to make a difference is just like it reads license plates or reads your face, that it can consume this information, look at a PDF of a bill of lading or a commercial invoice, digitize that data, and then begin to learn how to put that data to use. And that is a profound opportunity to solve a problem that still hasn't been solved. And when you look at the potential of doing that, the ROI gets way up there. And so I think that's the kind of thing that has the market very excited.
[00:09:18] Speaker A: Okay, Greg, thanks for that. So if I've understood correctly, so essentially AI, it's not some sort of passing fad, a conference topic like you mentioned blockchain before, this is far weightier, and it goes far beyond what the rest of us know about AI, which probably might be limited for some of us to the likes of chat, GBT and its ilk. But can you give us maybe some examples of where AI is already a key part of how supply chains are managed, how it could help, or how it has helped with some of the disruptions we've seen this year, like container ships around southern Africa, obviously something we've covered a lot on the Loadstar podcasts in 2024.
[00:09:56] Speaker B: Yeah. So I think, first off, I would say that you mentioned chat GPT and how it's captured the world's attention, right? And I think one of the things that happened because of that is it caught not just the world's attention, but the C suite's attention of large corporations? There isn't a company on earth that's not investing an enormous amount of their IT budget trying to figure out AI, right? So the investment fuels more innovation. And I think that wasn't happening until November of 2022, when suddenly people could write essays. But when it comes to ocean freight or air freight or any of these international complex logistics workflows, right, there's more to it. And I think, based on what I just said about the data challenge, the information challenge, you may have a Ferrari, but historically you're using bunker fuel to drive it, right? So the Ferrari does not do 180 miles an hour on the expressway. Right. It's chugging along and spewing smoke. AI has a chance to turn that into high octane fuel. And now your car is going to run really, really well. And I think that if you think about the human dimension that's being used today to get this information. So take the Red Sea incident as an example, where there's been a disruption, unplanned. You thought your vessel was going to arrive in Rotterdam on such and such a date, and now it's got to sail around Africa and it's going to add four days to the journey. At what point do you know about that? Because once that happens, there's a lot that has to happen downstream from that, right. You've got to change your plans, a destination you've got to rearrange, who's going to pick up the materials and take it to DC? And there's a whole realm of different things. So this idea that you can get that information sooner and not waiting for a bunch of army of people to read stuff, rekey it in and hopefully key it in properly so you have the right information. Like, just imagine getting that information several days earlier. That's actually a huge, huge improvement of what's been done in the past. And when you think of airfreight, where it's not days, its hours, or were living in an environment, right, where the data moves slower than the goods, where the goods have arrived, and suddenly youre getting an alert saying the plane landed, which was two days ago. Right. So this idea of information velocity and information accuracy is key to dealing with these disruptions which happen all the time. I mean, Red Sea is a, is a terrific example of a very high profile, big problem. But we had a rail strike in Canada that's causing all kinds of frustrations. Back in the days when I was at GT Nexis, we had this Japan earthquake which took Japan off the map for several months. We had a radiation plant meltdown. If you're a food shipper, you need to know if any of your food moved within 1000 miles of Fukushima. The ash cloud in Europe, which shut down western airspace for several weeks was another one. So these things are constant. And so I guess when you think of the disruptions and the kind of the unexpected that happens every day from micro events like a fire at a single port to a sea lane being changed, information is what's going to make you successful. And if you're relying on email and people, which is the way it's still being done, there's only one way to go and it's up, it's better.
[00:13:09] Speaker A: Greg, there are questions that people tend to ask about AI, and I'll just quick fire them at you. I don't know, maybe you think they're myths, but put me right. So one of those questions is, will AI take my job and make my boss richer by doing so?
[00:13:24] Speaker B: And the answer is possibly, you know, look, I've talked a lot about these kind of low value, heavily administrative jobs. I mean, that you're in a gigantic room keying out all day. So I mean, really, who wants to do that, right? I think, look, there's always going to be peripheral side cases that need a human brain to go and solve it. But what our customers are saying are it is relieving people of doing the parts of their jobs that nobody wanted to do anyway. So in other words, yes, you may not be having to enter 10,000 documents a month into your TMS system to keep your customers freight running, but those same people now can be redeployed to provide better services and have a more satisfying job. One of our customers I was talking to the other day and he said, look, I'm hiring people for their brains, not their fingertips, for their typing ability.
I'd say lastly, this is a very volatile industry. We were cruising during COVID and then the bottom fell out. And what is happening? Well, people are getting hired and fired constantly because of that, because volume ebbs and flows unpredictably. So that's the other problem is you never really catch up if you're a forwarder because you don't have any agility in human capital. And I think it's like, take the people you have, provide the services that your customers expect and let the AI do all the mundane stuff. And that's the kind of stuff that's happening today.
You asked me earlier about, well, what's really happening. There is a lot of that going on. There's more to be done. But I think in terms of the information dimension, it's getting better and better and smarter and smarter all the time, sort of like a Moore's law applied to freight. It's not perfect yet, but we're on the way. And just the way AI and machine learning goes, it's not a matter of if, it's a matter of when. This really becomes lights out and it becomes the norm, not the exception. You know, it's a lot like cloud. You know, back in my GT Nexus case, cloud was this new thing. Everyone was freaked out about putting their systems on the Internet, who now uses any business software without a browser. It's common. We overcame that, and that will happen here.
[00:15:29] Speaker A: Another one for you, Greg. How does AI make supply chain companies more productive, and how does this help the key audience here? Shippers?
[00:15:37] Speaker B: Yeah. Well, to me, it's basically, you've got more human capital to apply to delivering better services. I think customers expect a level of performance that may be unrealistic, to be honest with you, just in light of what we've been talking about. It's very hard and challenging and complex and in a constant state of change. But look, I mean, you can take your workforce and redeploy them to work those corner cases to provide that extra added value so that you're a steward of information, which means you're a better steward of your customers' freight. Right. The two are very interrelated. And I think that, again, if you're living in email and data centers on the other side of the earth, it's probably not going so well. There's probably a lot of hard conversations that take place unnecessarily all the time. So I think this world of being on your heels and firefighting, which is sort of the life of a logistician, it can go away. And maybe for a chance to go on offense, imagine what that would be like. Because the other thing, of course, is better data also means better insight and better analytics in real time, which normally doesn't come until three months after the product has been delivered. And I think that we're just getting started there.
[00:16:46] Speaker A: If we're saying that AI will harness unstructured data in ways that has not really been possible before, how exactly does that work? How do you get AI to do what you want?
[00:16:57] Speaker B: So this is where it gets a little more nuanced, because I think that there is a misnomer of this magic black box floating around out there because we've all tried chat GPT, right? You're a journalist, you know, I'm sure you've tried 'write me an article for 1000 words on how great Raft is as a company' or something, right?
[00:17:16] Speaker A: You I think I came up with these questions by myself, Greg?
[00:17:19] Speaker B: Oh, there you go, there you go. Right. But so what's the brain for that? Where is AI learning about that? Like, I like to say, write me a blog about the history of the ocean container. If you do a Google search on that, there are more than 38 million pages that come to you when you ask Google that question. Okay. So there's an expectation that this magic of chat GPT and AI is just, you know, you plug it into this black box, garbage in, magic comes out the other side. And the reality is, while that may work for an essay or draw me a picture of a cow jumping over the moon, or today even drive me from point A to point B, it's largely becoming commoditized crazily when you get into this world where it's not only a highly complex process, but it's handled differently by every company. The Internet is only marginally valuable as a brain, as a source of knowledge for the AI to do its thing. So you've got to have a way to train the models to work, to take the data, to read the information, feed it into your TMS, to operate in the way you've configured things, because everybody's got a different use and a different way. They're going to take a bill of lading. A bill of lading is a somewhat standard document in terms of what it is, but first off, it comes in many different forms and formats, sometimes in multiple languages, but then the consumption of it is different, right? So there could be two freight forwarders that are almost identical in size using the Raft platform, for example. And forwarder A is going to take the data elements of that bill of lading and apply it completely differently than forwarder B. So you've got to have this ability to train the models so they work for you, not against you, and you can't use the Internet for that. So for that reason, what's got to happen is that you've got to have workflow applications as part of any AI solution, meaning that users actually use screens and tools to orchestrate and move freight. And by doing that, those actions are being validated and trained by humans to make the models get better and better. So it gradually takes more and more of the work and it's going to be different for every company. So while there are standard things that you can use, batch level, Internet level stuff like location codes and time zones, and currency conversions, and things that everybody needs, how the address is used in a bill of lading, or how the equipment type is used in a bill of lading is going to be different, you could ask ten forwarders and there could be like seven different answers to that. Right? So it requires that automation, workflow capacity, and it also requires a level of scale and transactional volume to get the models to where they're stable and mature. And that's what's going to come next. That's what's got to happen in any business, not just logistics, but any business, because the black box where you put a bunch of stuff in and it comes out in a spreadsheet is fine, but it's only like 4% of what has to be done.
[00:20:08] Speaker A: In the end, there's lots of market entrants claiming they can bring AI to the forward in logistics industry, that this is a manual industry, that supply chains broken, et cetera, et cetera. But then analysts say, really all they're doing is automating documentation, not really automating the supply chain, nor helping drive cross partner and cross-departmental processes or customs compliance, or they're not automating shipments. Where does Raft fit in in terms of these challenges? And where do you see low hanging fruit for the sector as a whole?
[00:20:42] Speaker B: Yeah, well, I think what I was just describing is what Raft does and what you just described is the fundamental challenge that that approach is designed to solve. Because it's not just knowing what data to take out of the dock and put it in what way into a system. It's also a dimension of who to alert, what department, what partner do we need to alert, that won't be the same all the time. So this again, all goes into that building your own version of chat, GPT, that's built just for you, mister Freight forwarder.
I think that this is the piece where I think a lot of people get confused, because there are, as you said, a lot of companies, startups coming out, because data parsing reading is becoming rapidly commoditized. It's a race to the bottom there, and it's cool, don't get me wrong. I mean, the ability to feed a bunch of unstructured data and documents and come out with a nicely organized, structured thing is valuable without a doubt. But I think it probably only solves a small part of the problem, because again, it's the workflow orchestration dimension that has to be taken into account, and it can't just be patched onto a system like a TMS. A lot of TMSs are 30, 40 years old. It was code that was written in the nineties, maybe the two thousands. Right. It never was designed with AI in mind. So while you can feed data into it, it's not ingrained and no one's going to recode it. These are big, complicated systems of record that do a lot of very important stuff, by the way. And I do think that any AI solution in logistics has to know how to work with those existing systems because the investments there are off the charts and those systems aren't going anywhere. But it's how do you augment those investments and make them better with modern tech without having to rip and replace them? So I think sitting on top of a TMS or a warehouse management system or an ERP, we've got a lot of customers where they want us to interact with their CRM because a lot of the data is important to what they need to communicate to customers. Salespeople need to see it. I do think that as part of any big technology inflection point like this, youre going to have a lot of hype and noise and startups and investment that are coming in, and you're going to see a lot of these companies have great websites. I saw a company the other day that just raised $50 million. I don't know how many customers they have. Their website's killer. They hired a great design agency to build a great website. I'm a marketing guy. I was impressed as heck by it. But not a single customer press release, 150 LinkedIn followers. I mean, you can see it right there. So cool. But let's see it do it 100 times a second versus a demo that's canned on an Internet site.
[00:23:19] Speaker A: The obvious question I've got to probably follow up there with Greg is how then does Raft's products, how do they function with existing TMS systems? A lot of these products are already saying they've got AI integration. Then is that from Raft?
[00:23:34] Speaker B: Yeah.
Every customer we have, we're integrated to their TMS. Every single one. And it's such a write, it's a read write, it's a bi directional integration. But every one of the processes we're talking about has this unstructured data problem that limits the full value recognition of the TMS. So I'll give you a very specific workflow that we deal with that's probably one of our most mature products. Take AP invoices. A big forwarder is getting thousands of AP invoices from the freight they're booking on behalf of their customers a day sometimes. And so, you know, you've got shipments that have been created in a TMS where accruals have been put in when the freight was booked. Then these invoices come in and somebody has to look at them, or something has to look at them. So put AI on that to really extract and understand what the data is, and then begin to put that into the TMS and really validate against the accruals. Was the price what we thought it was? And if so, let's update it. And that was all done by people before, but the AI is increasingly taking more and more of that, where then you can automate approvals and alert parties when things are being changed, and then suddenly it's going in through payment and payments being processed faster, the money is moving better, the suppliers are happy, and of course, your people are again dealing with these corner cases. What we're also being told by some customers is, look, because of this, we're seeing spots in our process that are problems that we didn't know about before, right? Because everybody was on their heels just trying to keep up with the volume. So the same thing happens with customs clearance, the same thing happens with bookings and pre alerts and all these different things. And again, it's a large, large hairball. And so we are just somewhat focused on the freight forwarding zone of it. But think about supply chain visibility for a minute. Hey, I helped create that market at GT Nexus, right? And there are a number of big unicorn-level companies doing supply chain visibility. Yet when you really zero in on the international piece of that puzzle, their data quality is in the 70% range. You can't run a supply chain of three out of ten times. It's wrong. You can't do that, right? So maybe you can take that up to 99 and 98% and suddenly Red Sea disruptions are handled a radically different way than they are today. And I think that is where we're headed.
[00:25:53] Speaker A: Very interesting. Greg, may I take a flight of fancy then? What will AI in our industry look like in two or maybe five or maybe even ten years? Or to reference the Tom Cruise movie, that helped embed a huge range of new technologies, far faster than any experts thought possible at the time it came out. Can you give me your minority report version of the future for logistics and how AI technology will be used? I'll put it another way, Greg, what do your wildest log tech dreams look like?
[00:26:32] Speaker B: Yeah, yeah, no kidding, right? Take the gloves off here. It's funny, I wrote a blog the other day about this, and I used maps, right, MapQuest, which came out in the mid nineties and became a billion dollar company. It was acquired by AOL, remember them? Right around the.com bubble days. Now, you fast forward 25 years, and we have self driving cars that pick us up. And in between there, there's things like, you've got Siri maps, and mapping has become standard in almost every website, et cetera. And so the question becomes, do we expect the same thing here? And I think the answer is yes. I mean, the difference with AI, of course, is AI can code. You don't need people to write software code anymore. It can self-correct, it can learn. And I think that if you look at what we've been talking about, where I think phase one is just to get the information strategy in order so that the data that you're getting is no longer PDF attachments to emails, but it's fully digital and with a high level of accuracy so that the AI can now begin to start making decisions. There's this notion of AI agents, and I think it's, I can't pronounce it. It's like a i g e n t s. Like this idea that, yeah, there's workers that are purely AI. And if you think about that example I talked about earlier, where there's been a disruption in the Red Sea, for example, do we really need people to replan that at destination what has to happen? Because the reality is that if there's an alert that comes through and says, look, the ship has had to go around Africa, it's going to be four days later than expected. And coming to a different terminal at the port of Rotterdam. Okay, who's going to handle all the downs? What happens next? Well, right now there's an army of people that feed that data into a planning system. They have to rebook trucks, they got to rebook warehouses, alert yard guards, et cetera, to make sure this thing doesn't sit at the port for two weeks. Okay, and why couldn't AI take on all of that? I mean, once you know what's happening or what's changed, the AI should be able to literally orchestrate that entire move. So I think early on, it'll be hard for humans to let go of that. But I also think if you really think about how it's being done, is it really that great? Are the humans really crushing it? No, I don't think anybody would say, oh, my gosh, it's wonderful how we do it. I mean, I think that it might lean out these organizations over time because of the complexity is just crushing. But I do think that's one. Another one is just how you replan and how the planning systems, you know, TMS planning systems, ERP systems are starving for what I call executional data. They're starving for it, right? And oftentimes theyre not getting it for months after the product was delivered. And then after that they realize, oh my God, we lost money on that because all of this happened that we couldn't foresee, like the Red Sea or rail strikes. Imagine if that was all happening in real time and AI could be replanning constantly. And I know if you were talking to a planning software executive right now, he would, or she would probably be saying that exact thing to you. So that'll be your next podcast. So I think that's first. Certainly autonomous. You read about this Star Wars minority report. I think we're actually seeing it with robotics. Robotics and warehouse are pretty standard now, right? We're hearing about automated port machines that are driving around. Can we really handle robotic trucks carrying freight on the roads? Based on what Tesla's going through with self driving cars, every time there's a Tesla crash, it's the headline story on CNN. A big rig is an entirely different thing, right? Could we get there? I actually think it's possible. Again, I look at what's going on with personal transport. I mean, I don't know if you've ridden a Waymo before, but it's kind of spooky, right? This thing shows up, there's no driver. You get in this thing and it gets you there. It drives like a grandmother. I mean, do we really need cars like our own cars anymore? I mean, so I think that is what lies ahead. I think the question is, how long will it take? Will it take 30 years like Mapquest did? I don't think so. And because I do think that the AI is a snowball, right? It's just gonna get smarter and smarter and faster with time. And I think that it probably is in our lifetimes, we'll see a lot of this stuff.
[00:30:34] Speaker A: Well, I was gonna say your guess is as good as mine, but I think your guess is definitely way better than mine. I'm sure a lot of what you said will come true. And you've got the career to back this up. We've mentioned GT Nexus, but you did play a big part in the success of that, which was that was formed back in 1998 as Tradiant, renamed GT Nexus in 2001 before expanding to become a major logistics management technology company. We mentioned before it was sold for a lot of money. That was to Infor back in 2015.
Are there any similarities for you in that journey of GT Nexus and what it was trying to do for supply chains back then and what Raft is aspiring to achieve now in terms of transformation of the application of technology to the supply chain?
[00:31:21] Speaker B: Yeah, a lot, a lot of similarities. I mean, on one hand, this idea of there's a new market, this new way, this new kind of technology. So back then it was the Internet, right? It was cloud, what we struggled for a lot of years, and what to call it, it was on demand software as a service SaaS hosted. And then, like in the late, like around 0708, this idea of the cloud kind of morphed and everyone could kind of grok that. And so that eventually became a thing, you know, this idea that you could use an Internet browser on the web to use business software, because no one had done that yet. There was e-commerce and travel websites and things in Craigslist, but no one was using it for an enterprise. And of course, the way they would get it was they would buy a disk, set up a server, load the disk, and had all the risk was on the buyers, and the sellers made all their money upfront. And then the paradigm shift was you rent the technology, you pay as you go, so the seller as the delivery burden, and if it doesn't work, the buyer isn't out anything because they just stop renting it and go to something else. The other factor was visibility. This whole thing we were talking about, what GT Nexus eventually became really was a global supply chain visibility platform. And really that was the land of what they call track and trace at the time, and you were getting like 5% of the picture. And back then, and this is where it's similar to today, the challenge was data. It was data quality, it was data we could not get the information needed to deliver those beautiful maps that everybody was clamoring for, because the data was coming in as PDF's and different formats of EDI. And so the value proposition that GT Nexus we had was rather than send a file across a network, which is kind of like the telephone game, every time that the information moves to a next party, the story changes and it gets more different. And by the end it's a totally different story versus post the information in the Internet. And we compared it to LinkedIn. And if you think about the days when we had address books, you may be too young for this. There was a time when we had rolodexes.
[00:33:26] Speaker A: I wish I was Greg.
[00:33:27] Speaker B: And when you got a new job, what did you do when you got a new job? You had to let everybody email at least or tell everybody in your network, hey, I got a new job. My new email is
[email protected] and you were hoping that everybody would go into their own email box and Outlook and update that information. Now what do you do today? You go to LinkedIn, you change your information, you hit save and a picture with balloons goes out to everybody that you're connected to. Imagine that same information model being applied to the status of a shipment. That was how we sold that product. So fast forward 15 years later today, that did solve a lot of the problems, but the data information challenge still persists. And I think that stuff that we've been talking about with AI, to really, really kind of solve that once and for all and put an end to email and excel and PDFs and really provide that high octane fuel that is so desperately needed, it's another version of the same thing. But I think AI really is profound enough where I think it is going to change it. And I think that is the justification for all the hype and excitement that we're all being blasted by.
[00:34:36] Speaker A: Greg Kefer, Chief marketing officer at Raft, thanks for joining me today on this sponsored Loadstar podcast.
[00:34:40] Speaker B: Been great, thanks for having me.
