AI & Logistics: Will a ‘Job Apocalypse’ Hit the Supply Chain?

Speaker 1 [0:00] Definitely.

Speaker 2 [0:00] We have Jonah McIntyre. He's the chief platform officer at Trimble. Jonah, how are you today, sir?

Speaker 3 [0:06] I'm good. This is fun to listen to you guys' conversation.

Speaker 2 [0:08] Yeah. Well, I mean, it'su2014 there's never a shortage of news in freight, as you know, Jonah. Trimble is a massive organization that really has a very wide breadth of offerings and technology platforms. What exactly As Chief Platform Officer, what exactly do you do? What does that role mean?

Speaker 3 [0:31] All the people who work in product or engineering report to me, basically. It's almost half of the folks of the headcount that are in the transportation segment. So you're making products, whether that's designing them, conceptualizing them, strategizing them, or building them in a coding sense, or maintaining them, then You're in my area.

Speaker 2 [0:54] So AI is obviously taking the world by storm, we'll say. And this is an areau2014 I mean, Trimble's got an enormous amount of data, probably and arguably one of the largest, sitting on one of the largest datasets in the entire industry. I believe the number is something like 65% of all OTR capacity has a Trimble product, everything from your Trimble Maps products to your TMSs to your, you know, well, formerly telematics devices, your ELD devices are now part of the Platform Science merger. What exactly what does that mean for Tremble in terms of this mountain of data and systems that you guys are sitting on as you look at an AI agentic world?

Speaker 3 [1:32] Yeah, so if AI can sort ofu2014 it's like a genie coming out of a bottle. If you have a coherent idea, you can have software created that, that does it. AI lowers really rapidly the cost and the timeline to get software made, but software still is encumbered by needing to have data, like accurate data about the world. And, uh, this is something you can't make up. Uh, you also need to insert it into real operations. So there's these kind of choke points of you can have great software, but it needs to be deployed into cabs, or it needs to be employed in the back office. And then network, you know, think about something like LinkedIn or Facebook or, or FreightWaves for that matter. You can make a really slick website pretty quickly But getting a bunch of people to use it and all your business partners in particular to use it, that's the network problem. So those are the advantages we have coming into the AI age. It still is incumbent on us to make proper, attractive, desirable AI products that take advantage of those aspects.

Speaker 2 [2:41] Now, Jonah, at times, Trimble, a very acquisitive company, it's one of the primary growth engines of the Trimble model. Is through acquisition. That is both great because you get scale, but it also creates a lot of integration issues. Is that becoming easier in an AI world where you're able to bring in some of these new platforms and create better systems transfers between them and make sure the systems are talking to one another?

Speaker 3 [3:07] Yeah, it changes the equation quite a bit. So the post-acquisition integration can, can indeed be faster, but it also raises the bar pretty substantially on what's what are you actually buying? What's attractive about what you're buying? So if you're buying a clever piece of software or clever piece of even hardware, like clever piece of kit, that's just worth less now than it was in the past. But if you're buying a network, if you're buying a trove of data, if you're buying kind of like a distributed system that's out there in the world and you can now inject like new intelligence into it, those are Those are the things that have risen in value rather than declined in value.

Speaker 2 [3:52] Now, one of the things that Wall Street has beenu2014 we'll call it brutal in some sense, or at least has really punished a lot of the traditional software stocks in the last couple of years. Trimble is certainly not an exception to that. I think the concern, or at least Wall Street's perspective, is that AI is going to replace systems of record. Really impact it. That seems a little disingenuous to, at least in the freight market, because these systems are so critical. And I can't see, I mean, we still have AS/400s that probably run something like 80% of the freight that is in the market is run on a traditional green screen system that was 20 years old on an IBM mainframe. What is your perspective on that as someone that's dealing with both the public company element, and I know you can't talk about sort of forward-looking statements, but you can talk about your perspective as it looks in terms of your role at Trumbull, in terms of the perception of Wall Street and how they're looking at these systems of records.

Speaker 3 [4:53] Yeah, I think that there is some truth to the fact that a certain kind of software business is of less value now. If your primary moat was, well, we have software that took us a long time to build, and, and that's why nobody can compete with us. So something like SAP, it's like, took us a long time to build, it's very huge, and no one's going to compete with us because they can't, you know, build it, or it wouldn't make economic sense tou2014 for them to try to catch up. I was like, well, now you can catch up, it's notu2014 that's not that big of a deal. If your primary moat is, uh, we're already in a bunch of places and it's hard to change systems, I mean, that, that that part hasn't gotten better. To your reference earlier, there's a lot of people hold on to old technologies simply because it's painful to change them. So those elements, the data elements, brand also, like you can produce some pretty whizzy technology with a small team, but you got to wonder like, is that supplier actually going to be there in a couple of years? So those are elements I think haven't diminished in value or maybe have even gone up in value.. And it's tough for markets to figure out what's the blend of that. With somebody like, uh, Trimble, you know, surely we do have software that falls in the first category, and we also have a lot of stuff that falls in the, you know, beneficial category. And so, uh, it's hard to sort those all out. I also think it's interesting to see at this moment there's also a kind of like a high perfection of, uh, software businesses. That's why you can see startups that take the playbook of how to make a high profit, high-growth, you know, software business. And but they don't need to hire as many people and they can build the software super fast. So you've got these hypergrowth businesses. I just suspect that they're, they're kind of like a sugar high. They're, they're, there's not a lot to prevent them from also being then disrupted by the same forces that let them grow really fast.

Speaker 2 [6:53] Yeah, I mean, it's a great point. I mean, AI wrappers, you know, 2 years ago, some of the voice technologies and trucking, these technologies that we're using some of the AI underlying technology sort of showed up and had the advantage because it was still a new technology. It was hard for people to work on. They weren't familiar with it. With Claude and other AI platforms, it's incredibly easy where, I wouldn't say anybody could do it, but it's not hard. The gap is incredibly easy and accessible. It puts a lot of pressure. But one thing I know from running a data business, I think one of the most empowering things is we're sitting on a mine just just absolute vast amount of data here at Sonar. And we're able to do visualization layers, present stories to the data, present new ways to visualize it, new ways to make it prescriptive that we've never been able to do, that used to take months to ship, and now you're shipping potentially within hours. It's not changing the data. The data is the value proposition, but the visual layers are incredibly easy to ship. I imagine for a company like Tremble, who's just sitting on an absolute mine of data, this is an empowering technology for you guys.

Speaker 3 [8:09] Oh, 100%. But I mean, plus, you're probably too discreet to raise this, but I think that particularly in the US, there's a lot of reputation around our TMSs are kind of legacy, they're old TMSs. And that's true. I mean, we have, for example, a product called Innovative, which is arguably the first TMS you know, commercially available TMS in the world. Uh, this thing's from the '60s, right? Pre-deregulation in the US. It would, it would be easier to start a fire department today than it was to start a trucking company, like a, a commercial trucking company the way we would talk about it in the '60s. You had to prove that youu2014 the community needed you in the same way that, you know, you need a fire department or something. It was, it was very much non-commercial. And, um, and thatu2014 you look at that TMS, the fact that it survived all these decades, uh, is, is incredible. Like, so many of the companies that it was up against competition, they're, they're no longer here. But you still, as you said, you look at it, it's a mainframe-based system, it'su2014 it looks old. And now we have theu2014 with AI, we have this ability to rejuvenate and like reinvigorate our our, you know, kind of legacy applications for our customers and produce these totally modern experiences. They look like anything a startup would produce in 2026, but they don't require any change really to your base operation. You don't have to migrate data or anything. So that's a huge benefit. It's to your point, it's like the data is the value for the customer, that their customer's data is still there, it's still in the same place, but you modernize everything around it. In situ without having to lift them up to a new system. And that's just a huge benefit for us.

Speaker 2 [9:57] Yeah, but in fairness to Innovative, I mean, they were running Swift's entire fleet of 20,000 trucks on this old school mainframe that Ernie Bettencourt had barely spent any time developing before he sold it to Trimble. I don't know if Swift, I don't know if Knight-Swift is still operating on the old Innovative system, but you know, when I was at US Express, we had something like 300 engineers that were working on some technology that we built in-house. And you had Swift that had like 2 people that were running on this like old school. And I think one of the interesting things that I remember when I was in trucking is that the disruption, you know, US Press was constantly deploying new code. It was really Schneider and US Press were sort of the 2 big asset carriers that were building their own technology in-house. The problem was anytime you did a system of record change, you changed out a fuel card, you changed out an ELD device, it was disruptive to the operation. Ultimately, trucking companies just want to book freight.

Speaker 3 [10:55] Yeah, yeah, that's right. Yeah, it's actually one of the things that always is in my mind that the sort of the peak, the apex of great products in our space is not that they're flashy and they sort of draw your attention and you're wowed and wooed by them. It's, it's, it's a bit more like clean drinking water. It, like, it's a miracle that everywhere we go in our environment we have clean drinking water just right there, and it's so cheap, it's free, and you never think about it, and you don't have to, like, carry water with you everywhere or a water sanitation kit or something. Like, that's a miracle of civilization, but its apex is that we forget about it because it just works. Same thing with, like, trash collectionu2014 huge logistics problem. But it works so well and so cheap, we just don't think about it. That's really how these systems ought to be. They, they ought to be so good, so reliable, so smooth, and so cheap that they're like the water, you know, network or trash system where they, they just get out of your way and let you focus on, to your point, like operations and moving freight, hiring drivers and keeping them happy, and all those other elements.

Speaker 1 [12:03] I love that analogy, Jonah. I want to ask, uh, there's this, uh, I guess quote from you, or you know, something that I heard about a job apocalypse. You said that AI could potentially create a job apocalypse when it comes to logistics. Do you still believe that? And what does this look like, I guess, in let's just say 5 years?

Speaker 3 [12:25] Yeah, I love that. I said this in passu2014 well, not in passing, I said it in answer to a question. It became like the hallmark next to my name for a while. And what, November, October, November, something like that. Then it quieted down. Then now everybody's like referring to it because it's coming to pass. Yeah, uh, like if you do a jobu2014 you think about this for a secondu2014 like you do a job that's mediated through a screen, like all, all the data input and output goes through a screen, like a lot of back office operations and logistics are mediated through a screen. I, I don't really see a lot of reasons why those can't be, displaced or eliminated by, by AI. And what's left is often a lower-paying coordinator role, like au2014 so it's, it's some jobs get eliminated, some get, you know, devalued. There's a maybe a thinner level at the very, very top where, where you have leaders who are kind of AI bosses over a large number of, of AI. But that theu2014 you do the math and it's not like there's some you know, law of physics that says that the amount of money distributed to the staff has to remain constant or something. Like, it, it, it will just be cheaper to have a back office because you'll need less people, and you'll haveu2014 and the people youu2014 the remaining people may get paid a little more, they may get paid less, doesn't matter. But it's not like you redistributed all the salaries. So yeah, I, I stand behind this. Um, I should say that, like, I don't think that means all these people are just gonna go sit on the curb and, like, ask for handouts and be jobless. There's a concurrent rise in things like solo entrepreneurs andu2014 or just normal entrepreneurs, or people joining, moving to different sectors of the economy that are more dynamic. Like, this is the creative destruction of capitalism. It just is the case. Like, think of all the typists and all the other jobs in the past that got eliminated. It's not like all those people just are outside waiting for, you know, food handouts or something. Theyu2014 the economy moves on.

Speaker 1 [14:32] No, it definitely does move on. Now, I guess a follow-up to that is you kind of mentioned it, but the human aspectu2014 where do you see the human being most important? Just for my personal curiosity, because this is really fascinating to me for sure. Yeah.

Speaker 3 [14:47] So I think where we're at with AI is that it finishes the sentence, but it doesn't really know where to start it. So you think about it as like this incredible autocomplete. You know, mentioned Cloud Code a moment ago and Cursor and Codex and all these options. You know, someone who isn't technical but has an idea, and as long as the idea is, you know, consistent and complete, they can make some software. And same thing could be said for they can make a piece of content, they can make a website, they can make imagery. I mean, there's a lot of stuff you can make. Uh, but what itu2014 AI hasn't been able to do so far, you know, satisfactory, is, um, to have the initial idea to begin with, right? So, you know, the, the kind of the king or the queen needs to know what they want before their, their people can go, can go make it. And the people now are largely becoming AI, but someone's got to haveu2014 got to pick up like the initial spark, the moment where they have an idea, good idea about what they want. And that, that's at the level of entrepreneurship. It's all the way down to individual task, like what, what should AI be working on in the next moment? That's a, that's still a human-led decision for practical reasons, like AI is not good enough for it yet, but also for almost like moral level questions. Like we want, we want humans at the steering wheel of our companies and our, you know, our industry. We don't want that to be a black box right now.

Speaker 1 [16:18] No, we do not.

Speaker 2 [16:19] So, General, one of the sort of, you know, industry issues that have been brought up when you mentioned Trimble is a big business. So you have a wide variety of businesses you guys have acquired, different software platforms, a lot of them just trying to get them to talk to one another. And at times there was a lot of like customer frustration. You mentioned Innovatev, you know, you could go to Maddux, you go to IDSC. You guys have bought a lot of business. In fact, co-founder at Sonar and FreightWays was actually Ben Murphy that sold a business to Trimble, or TMW, that is now Trimble. One of the sort of issues was with all these acquisitions, you guys have this vast amount of different service requirements and different platforms. As someone that's designing the future of Trimble's platform, how is AI enabling you guys to serve your customers better potentially on those legacy platforms and move their businesses forward?

Speaker 3 [17:14] Yeah, besides modernizing the point solutions themselves, another one of our initiatives is to establish a common platform. So my thinking on this is you have to walk a line, you have to be careful. What we want is maximum interoperability with minimum interdependence. Like, I don't want people to have to buy, you know, their entire tech stack from us if they want to buy anything from us. I don't want to create this sort ofu2014 we point to something like the Apple ecosystem where everything just works together. Well, that's when it works, right? That's great when it works. But I think there are a lot of businesses that they would rather make those decisions one by one. They want the interoperability without being required to go all in. And, uh, so what the way we think about this is we have a minimum common platform. There's elements there that I feel like every customer would just expect if they have one product and they try to add a second one. Those elements remain. So things like your master data, setting up your business, um, your connections to the outside world, you know, those sort of things should be common. And then everything else doesn't have to be common. So one of the things I let I let be different is like look and feel. I have the sense of, you know, these point solutions mostly have their own users. They're not hopping around from one product to another product all day long. If the look and feel is a little different, like, that's okay as long as it's optimal for that point solution. I don't really see why the back office, you know, invoice to cash team needs to have the same look and feel as the workflow solution that's given out to drivers in cab. Just doesn't seem to be a priority to me. But the fact that they can talk to each other, like, yeah, that's pretty important.

Speaker 2 [19:10] Yeah, AI can do that really well. In fact, you know, one of the frustrations that I think customers of any, you know, Tom McLeod, who's a competitor to Trimble, for sure, the worst thing you could ever ask him for was a custom screen. Like, you didn't do that. And I think there was some frustration probably at Trimble in the old days You know, you went and asked, innovated. Now this is the box, this is what you buy. This is what we have. But AI enables you to build these custom experiences. And really, one of the problems of buying third-party software, and Salesforce is sort of the classic example of this, is you buy this massively big CRM that has all these reports, and you can't find what you want to find for that specific function because it's, you know, there's 100 other reports that have been built by the clients. One of the great things about AI is building custom-built or providing custom-built experiences for the user.

Speaker 3 [20:00] 100%. That's why we actually just released a new TMS for the small and medium-sized carrier market, particularly in the US, and it's predicated on a build-to-order model. So, you know, the old world would be like, code's expensive to write. Man, once we write code, we really want to sell that code as many times as possible because it doesn't cost us anything more to sell it again, so we make high margin. And you would kind of try to convince customers or prospective customers that they don't really need to do work the way that they think they need to do it, that actually our other customers accomplished the same thing this other way, which is another way of saying, like, please don't make us write any more code, just buy the code we already wrote. And the new world is like the unit economics flip. Uh, it's probably cheaper for, and like more effective for us to just make what you want. So if you, if you, if you're like, hey, we move around, uh, like chocolate and it's really important that every time, you know, the truck comes to a stop, uh, there's like a check on the refrigerator unit and it, you know, it's humming away properly. And we'd be like, okay, well, like, you want it, we'll make it, we'll just make it for you. And, um, you know, at a, at a very attractive price point, you know, frankly lower than what you would have had a normal TMS for a couple years ago, we can just make what people want. It's a beautifulu2014 like, I, I really love this moment where we're much more in tune with actually what customers want because we're not trying to convince them to reuse what we already built. And, uh, it kind of goesu2014 it brings us back to a, to a phase where we're instead of like enamored by the repeatability of SaaS, we're more like engaged with the specificity of customer need.

Speaker 2 [21:54] So Jonah, do you see a worldu2014 I mean, Trimble's historically been an enterprise. You certainly have some downmarket products.

Speaker 3 [22:00] Yeah.

Speaker 2 [22:00] But really enterprise business is really the bread and butter of Trimble. Does AI enable you guys to go downmarket to much smaller customers because the customer acquisition cost are not as intense and customer servicing costs are not as intense?

Speaker 3 [22:17] Yeah, I think it does. That's, well, at least that's our premise with releasing this new TMS called Trimble TMS. The concept here is that this build-to-order model in particular caters to someone who they either have a TMS, but for whatever reason they've outgrown it, they really need something new, but more often than not, these people are kind of net new. This is the first time they're buying a TMS. They may have strong ideas about what they want. They also, you know, want something that was very price attractive and very malleable because their business is growing. So they want to be able to adjust the product, adjust the TMS rather than having to adjust to the TMS. I think these are, you know, these are great It's a great direction for us to take because it combines our benefits of scale of being able to have a dedicated applied AI group, which is like significant. Our applied AI group is very large as a resource pool, but then it's applied to a segment of the market that would not usually be able to get that kind of scale.

Speaker 2 [23:26] Yeah, I think it's a really important thing, and we've seen it at Sonar. I mean, we used to focus only on enterprise, We released Blue to Blue, which is our downu2014 our consumer mobile app. We've released our Market Monitor, which is again, just a summary, executive summary. And then we have the new QuickRates, all enabled because AI has enabled us to go to market, build it and go to market much more efficient. It's a really remarkable time. And I think it's going to be interesting to watch. Startups have lived and really made their play on serving the smallest clients because that was an area, you know, it's the old innovator's dilemma. It's an area that Trimble was focused so just solving your enterprise customers' problems. It's gonna be really interesting to watch this sort of down long tail business.

Speaker 1 [24:12] Oh, no doubt.

Speaker 2 [24:12] Where companies like Trimble are like, hey, let's go after that. And now you've got the best of the best in terms of application. You've got the best of the best in terms of resources and go-to-market expertise. It's going to be a real challenge for some of these startups.

Speaker 1 [24:25] It absolutely is.

Speaker 2 [24:26] But they're going to learn a lot. I mean, the great things about startups is they're constantly doing things. Trimble can watch it too. So this is au2014

Speaker 3 [24:32] it's a really dynamic system. Exactly. The dynamicism is theu2014 well, it's good for everybody, basically. It's good for the industry to have that.

Speaker 2 [24:41] Well, I think there's going to be plenty of jobs because all these companies are going to be constantly creating new product offerings and going in new directions. John, I really appreciate you. Joining us, we'd love to have you back. And thanks. We'll be, we'll be back right after these messages. Thank you.