Open Source Freight Tech: From TMS to AI Tools

Garrett Allen: These brokers are using things like, you know, grok and Chat, GPT and Claude to, to kind of hack around stuff on the weekends and like giving them the tools to say, Hey, you can go. Actually like use this TMS and like the code's right there, you can open up, feed it into Grok and say, Hey, add a page and it'll do it for you.

Andrew Verboncouer: Hey everyone. Welcome back to this side up where we dig deep into the minds of founders, builders, and operators, reshaping the future of logistics and supply chain. Today's guest is Garrett Allen, founder of Load Partner, a startup that's simplifying the chaotic. Of brokerage jobs. If you've ever worked in freight, you know how messy the back office can be, and Garrett saw it firsthand and he's decided to build a better way.

In this episode, we'll dive into the origin of load partner. We'll talk about automation with a human touch and why freight is just the beginning. If you're deep in the trenches of brokerage. You're just curious about building legacy industries. Garrett brings sharp insights and no BS perspectives on product sales.

What actually matters when building tools? For operators, let's get into it.

Garrett Allen: Been in freight for a little over 10 years. Uh, started out early career at a brokerage called Integrity Express Logistics back in 2013. Uh, I was a really early hire there, like hire number 50. Uh, their first kind of dedicated IT guy.

I got, I got brought on originally as a help desk guy. I installed PCs and um, you know, they were a small brokerage seeing a lot of, a lot of growth, kind of starting to scale. And, uh, they had this internal TMS system built that was, you know, they're having a lot of problems with their tech stack, you know, a small brokerage scaling quickly.

Uh, it's really important obviously to be able to turn, turn around requests quickly and have a really solid, uh, platform. And, you know, kind of long story short, I pitched 'em on, Hey, if you, if you gimme the resources, like I think we can build a TMS to, to suit you guys and scale up. And, uh, they ended up giving me the resource.

They gave me a shot, built an engineering team, and, uh, built the TMS they run on today. Uh, worked there for about eight years. Uh, when I left, they had scaled to over 500 employees and. Um, making a billion dollars a year on that TMS. So a lot of my early career was, um, you know, seeing a small brokerage working in the trenches in it for a brokerage like that, you know, scaling them up multiple locations, different modes of, uh, you know, logistics, things of that nature, and, and building this offer to support a brokerage like that.

Then from there, I went and, uh, worked at Transfix for a few years as a senior engineering manager. So there I oversaw the payments team, so a lot of accounting. And then also their shipper operations team back when they had, uh, a brokerage that was really like the shipper side of the brokerage and then also their integration.

So everything like EDI, things like that. And obviously a lot of overlap with payments that side of things. So worked there for a few years, overseeing those teams and then, uh, went and started load partner, uh, back in 2023. So I guess just on the load partner side of things, uh, we started out in 2023 launching our AI track and trace system.

It was called Lasso. We launched it in, uh, I guess it was, uh. What, April of 24, I believe we launched that. So that was one of the first like AI, voice and written communication systems. So it's really built to help, uh, track and trace night dispatch those kinds of teams with automating ai phone calls out to drivers, getting status updates, following up on like things like Bools and it worked across, uh, email, text, voice, everything.

Um, and would kind of like aggregate it all into like outstanding requests and things of that nature. And we did that for a few months and. Kind of realized, uh, you know, towards the end of 2024 that a lot of that tech was a bit of a race to the bottom. A lot of, a lot of VC money came in and, uh, started to like, you know, lower the prices.

They, people were charging these brokerages and we just couldn't keep up being kind of a bootstrapped company. Uh, we didn't go raise a bunch of VC money. That wasn't really what we wanted to do, so. Um, we, we looked at what we wanna do next and seeing like, hey, this is sort of a race the bottom. We don't think there's much of a future here, but this tech is really cool.

Yeah. And, uh, being that, uh, my background and my, my co-founder, um, his name is Nick. I didn't, I didn't mention him, but he also was back at IL with me back in the day, their first CTO. So he, he started at load partner with me. And so we looked at, you know, what do we wanna do? And, and, uh. We said, well, we've got a lot of experience building TMSs and people have always asked for an open source, TMS, uh, let's make it a reality.

So spent the first half, kind of the end of 24, prepping in the beginning of 2025, uh, actually building an open source TMS completely in the open live streaming the development every single week. Uh, just, you know, showing people pulling back the curtain and showing people how freight tech gets built and what it really looks like.

Uh, you know, basically mid 2025, we actually launched the first open source TMS. It's out there today. You go on GitHub and look at it, go sign up for our, uh, we have a startup plan for brokerages that, for like solo brokers that are small and then we have really reasonable pricing. And, um, so yeah, we, we did that, which was, uh, awesome.

It's been really cool launching that. It's kind of recent, only been a couple months. It's been live, but we're working with, uh, our partners on getting different integrations in place on that. And then, uh, in the, on that journey, I also, uh, I'm a serial builder, so I'm always building multiple things at any given time, pretty much.

And, um, we, we were building our integration to the F-M-C-S-A for KE details, right? Um, that's, you know, how many times have I done that over my career? Too many. Yeah, it dozen. And, um, and so, uh, I, I built it again. I was like, Hey, you know what, um, what if I just throw a UI on this and. Put it out there for people to use it.

And so I did. And um, people really liked it. It got a bunch of traction. Uh, I, I basically, you know, I spent a little time making the UI actually look dice. 'cause a lot of those systems, they look like they were built like in the early two thousands. Pretty dated. And, um, so I called it search carriers, search carriers.com.

And. Just kept adding to it, you know, adding, Hey, you know what, let's let, so people can track, uh, get emails when there's a new inspection for a company and like, let's add, uh, some really fast search. So I, I spent a bunch of time on this really robust search engine that returns anything on under 500 milliseconds.

So you can, you know, search for like a company owner's name and it'll pull their company up instantly. And just keep adding on top of it. Just things that I think are useful. And, uh, basically over the last few months that's really taken off and seen a lot of, a lot of traction on that front. So, um, been spending a lot of time there.

I just recently announced a, a partnership with, uh, freight Wa, freight Waves Playbook Program to help, uh, owner ops with some of their processes. It's gonna be adding a way for them to like challenge inspections and things like that, to kind of get back to the carrier side of things and got a bunch of things cooking for the brokerage side for, um, you know, things across vetting, onboarding, things of that nature.

So, yeah. It's

Andrew Verboncouer: kinda where we're at today. Yeah. Awesome. Thanks for that rundown. Obviously you've been in, in freight tech for a long time, right? Building the, um, broker TMS, is that, is that what IEL is? Yeah. Yeah. And the cradle to grave specifically. So not quite the silo model you hear about. Yeah. Yeah. So I mean, really, really comprehensive end to end.

And then, you know, you, you were at Transfix for a period of time and I guess, you know, coming out of, you mentioned a little bit like an open source, TMS, we talked about this previously, like we see the need for that. We've been in, you know, not in the space as long as you have, but 2018, I think it's when we first started doing stuff and in, uh, logistics and transportation and, um, yeah, you don't see it a lot like one, you don't see a lot of open source stuff that people can build.

A lot of it is like. Proprietary, you know, stuff that's ultimately using other web frameworks that are already out there or um, you don't see particularly people building in the open. Right. And that kind of caught my eye. I think we met, I dunno, maybe last fall at some point and I was like, I think that's the guy from these live streams on, on LinkedIn and uh, I dunno where else are your live streaming?

But, you know, I've seen him and tuned into him, so it's always good to watch as a builder. But what was the. I guess one, how has the, how has attraction been for the open source TMS, um, at Load partner and then also like, are you seeing more people subscribing or kind of like forking it and just, you know, rolling their own with kind of the.

The foundation that you've built?

Garrett Allen: Yeah, it's been really interesting because, um, you know, obviously, like you said, the industry's always been really closed door, so it's been, uh, it, it's funny you have to actually almost like convince brokerages or like customers why it makes sense to be open source 'cause they're so used to the closed door.

Yeah. Um, but we've had a lot of good interest, especially from like small and mid-size brokerages where these brokers are using things like, you know, GR and chat GPT and Claude to, to kind of hack around stuff on the weekends and like giving them the tools. Say, Hey, you can go. Actually like use this TMS and like the code's right there, you can open up, feed it into Grok and say, Hey, add a page and it'll do it for you.

Uh, and that's been really cool to see. And you know, there's a lot of companies out there that are kind of like, they're not TMS providers, but they're providing services that are like. Adjacent to it. And they're like, yeah, we, we have people that are using Google Drive to Google Sheets, to track loads and stuff.

We're using, you know, HubSpot or something goofy and we'd love to provide them ATM S. And it's like, well, we've got a TMS that's open source that, you know, you can take that and like, tweak it for your use case and sort of like have ATM S, you can now offer your, your customers, uh, without spending a million plus dollars to build your own.

Um, so it's been really cool to see just like all the different use cases and ways people are, are able to leverage that tech now that it's out there and actually like open.

Andrew Verboncouer: Yeah. And is search carriers, you mentioned a little bit, is that, um, is that open source?

Garrett Allen: So Search Caries is an open source, but there is an API that you can sign up and use.

So, um, I'm pretty, it's not open code, but I'm very open with basically everything you can do on the front end. There's an API that you can just go sign up for a Pro plus plan and get straight access to it. So, yeah. Um, trying to really open the data side of that more. Uh, the code itself is like honestly not as valuable as like a TMS because people aren't really trying to go into search carriers and like.

Make a custom version. Yeah. They're like, I want the data to be clean and easy to find. Like that's the use case. Yeah. And so I'm trying to really focus on taking all that data. There's millions and millions of data, uh, data points in there. Uh, and just making it really easy to like read and understand and access so that people can build on top of that API, uh, and I can just kind of handle that dirty work.

Andrew Verboncouer: Yeah. Yeah, it's really aggregating some of that. You mentioned F-M-C-S-A, is it bringing in other data from other sources as well? I'm assuming

Garrett Allen: so, yeah. There's a couple other data sources. The primary data source is the F-M-C-S-A. Um, but we're also working with some other providers to bring in just different things around like VIN details for example.

So if you open up like search carriers and you go to inspections, you can see all the detail around equipment that's on there and actually click in and see like what's the make and model of that truck and like when it was manufactured and all that detail. Then also I recently launched the image search on there, um, which I'm working on ways to basically aggregate that data back so people will be able to um, you know, actually validate the trucks of the truck that they think it's supposed to be.

Uh, going back to, you know, some of the vetting tooling that it's in the works right now. So, um, you know, finding other ways of just other pieces of data that are useful to people. And again, just making as easy as possible to like connect all the dots. '

Andrew Verboncouer: cause there is a lot. Yeah, I, I did see a lot of activity around that post, the announcement for the image search on Twitter or X specifically.

Yeah. Like what, what's your take on that? Some people love it, some people are like, who's, who are you trying to, you know, exploit here? What's going on? Like, what's, what's your take as the creator?

Garrett Allen: Yeah, so I mean, honestly, the reason I created that is I, I built this tech, uh, last year actually for a, a software called Truck Verify that I haven't, I, I, we saw, we sort of soft launched it last year, uh, probably gonna relaunch it later this year, which is basically a system to let, um, a shipper take a picture of a truck at pickup.

Verify that it's the, the right truck to make sure the MCs right, the VIN number matches all that. So just another step for vetting fraud prevention, right? And, um, I took that model that I had sitting around that was already trained to like read truck details and identify trucks. And, uh, I saw all these people using search carriers and posting on X and saying like, oh, here's a truck with like the mc taped on the side.

Like, who is this? Let's look 'em up and see what their inspection history is. And I said, Hey. I'll just make this super easy. Like you just literally take that picture and search carriers. You don't have to type anything in, it'll just pull it right up. Yeah. And so, you know, these people were already taking the pictures and everything already, so I just figured let's cut a step out and make it so they can actually start just jumping straight into it.

Um, and yeah, I know a lot of people on there talking about like. You know, is it, what's the privacy look like? Yada, yada. But at the end of the day, like if you're on the highway, like it's a public place. Yeah. Like people could take pictures of your truck. Um, that is, it is what it is. I'm definitely not advocating for people going around and, you know, taking a picture of somebody's truck in front of their house or anything like that.

And there's clear disclaimers on the app that that's not allowed. And on top of that, all the pictures that are taken right now are not shown publicly. Uh, now in the future they, there will be a system to basically vet and approve those pictures to allow them to be shown against profiles. But for now, that's all completely behind the scenes.

It's not shown because if I'm going to show those publicly and start surfacing that data. We're gonna spend a lot of time making sure that, hey, these aren't pictures of people. It's not a picture of somebody's house that they're, that they're, you know, privacy conscious. So, you know, I think at the end of the day, we're really trying to, to fight the bad guys.

And I think pictures in public places are, are totally within the realm of reason of trying to, to find the bad guys out there stealing freight and, um. I mean, you've got other companies out there already doing a lot of that stuff too, like, like shout out to gen logs. Like they're, they do some really cool stuff over there and you know, they have thousands and thousands, maybe millions.

I'm not sure what their latest is of pictures of trucks all day, every day across the United States. So, um,

Andrew Verboncouer: yep. Yeah, yeah. I met, I met a couple of the, the folks from that team, one of the, um, they're backed by one of the VCs here in Green Bay, Titletown Tech. They do a lot in logistics, supply chain. And so I met those guys at a, at another, uh, a conference.

And, you know, I think mixed signals too, in the industry, it's like, Hey, we're trying to fight fraud. Like the, they're in the public. Um, they found a lot of fraud, right? If you, if you have a ca, a case or a claim and you say, Hey, I need to find this, like. They're gonna help you find it. I've seen so many case studies of them where it's like within an hour, like they've located the truck somewhere and it's, you know, reclaimable, you can almost think of like search cures as like, I could see an evolution of it where it's like gen logs at rest.

Like you have your own camera pointed at your lot. It's just automatically verifying people that come in and, you know what I mean? Like you can Yeah. Um, rest assured that this is the right truck, whether that's a pickup or a drop off or what, what have you. Um. So, yeah. Really cool.

Garrett Allen: Yeah. And, and to add to that too, um, one of the things I, I had mentioned on, on X too, I think it was last week, um, is that that image model, that that's actually like doing all that work behind the scenes on search carriers is going to be available through the API, uh, and then your future.

So it's going to be something where if you do have a yard and you've got a camera and you're like, Hey, we wanna like automate. Pictures of trucks rolling in, that's gonna be available through the search carriers API within the next like month. So yeah, you know, that's stuff that we can just support you building that yourself, or if you want us to help build it, we can do that too.

Um, but, you know, really just trying to, again, right, going back to it's, it's not open source, but we're trying to just like, give you the tools and like to structured data to be able to build what you wanna build as easy as possible.

Andrew Verboncouer: Yeah, for

Garrett Allen: sure. So is, is search key is part of Load partner or is that

Andrew Verboncouer: separate?

Garrett Allen: So search carriers is, is separate from Load Partner. So Load Partners focused primarily on like the TMS side of things. Our open source TMS. Yeah. And then search carriers is uh, is a separate company, um, building for, you know, basically carrier vetting tools, carrier sourcing. I think the official title on the, on the website is a carrier research and discovery tool.

Um, 'cause that's most of what people have been using it for. That's what it's been optimized for,

Andrew Verboncouer: for sure. Yeah. It's super interesting to see, you know, 'cause you're kind of. The Indie hacker of freight, right? Like are you familiar with like Peter levels and all these people that are like really into like micro SaaS?

You know, you're starting to like build these small, very point, very problem specific tools, which is cool to see. Is that your vision or are you thinking like these roll up into something, you know, into one platform? It's a suite of tools, I guess. How have you thought about building out this tech, uh, free tech empire?

Garrett Allen: I mean, I'm definitely a little bit inspired, uh, by Peter levels. Yeah. I mean, uh, I, I love building stuff. Um, I, when, when I see a problem, I love to just like. Get in there, hack around it, get something valuable into the pe, into people's hands. Like even back at, you know, integrity Express and Transfix. I was always the guy that was like, Hey, we have a new project, it easy to go zero to one.

Like, Garrett's the guy that's gonna lead that. Uh, and so like, that's always been my strong suit and so I really leaned into it. So, um, yeah, you're totally right. Like I mentioned that Truck Verify product, that's probably gonna be another separate product I've been back and forth on. If that's gonna be part of search carriers or its own standalone thing.

Yeah, probably it's gonna be its own standalone thing that's just like integrated or something. So, you know, maybe they all roll up into something in the future. I've thought about that as well, building like a. Broker tools package or something where it's like, Hey, you pay a flat fee and you get access to all these different tools that have been built.

Um, but I know one thing is for sure, which is that I'm just gonna, I'm gonna keep building software that I see and there's, hey, there's a problem space that we can, we can assist with. Let's build something to help people out. That's just gonna keep happening. So, uh, yeah, just be on the lookout for more tools.

It's, it's, it's gonna keep going.

Intermode Ad: Today's episode of this side up. Is brought to you by Inter Mode a software consultancy. Doesn't it feel like everyone in logistics is trying to figure out how to design and build great software just to stay competitive? But how do you make it happen if you don't have the right talent on your current team?

Well, that's where inter mode comes in. We're the folks who help both startups and industry giants bring their tech dreams to life. Within logistics and supply chain, we don't just give you a strategy to follow. We execute with you by designing beautiful interfaces, building scalable design systems, and bringing Silicon Valley level engineers to build it right the first time.

Whether you're a new player with a bold idea or an established company looking to modernize your systems, we've got your back. We've actually worked with some names you might recognize Ford, Proctor and Gamble, Trimble breakthrough US Venture and more helping them tackle their biggest tech challenges.

Here's the thing, no matter how big or small your project is. We're all about turning your vision into reality. Wanna see what we can do for your business? Just head over to inter mode is inter mode making ship happen.

Andrew Verboncouer: Yeah. So you mentioned like obviously, uh, dart or Lasso is doing more in, you know, using AI for communication, things like that.

Like what, what do you see obviously coming up in, you know. Coming up in IEL and into transfix and into some of those to like, now you really understand the problem space of brokerages, of carriers. Like what role do you see AI playing now and into the future, and how are you thinking about, you know, continuing to leverage that?

Garrett Allen: Yeah, so it is a good question. I mean, I think AI is definitely gonna cut some of the, the crot out, especially of like high volume, uh, conversations, you know, things that people are doing a hundred times a day. AI is just gonna get better and better. Now, obviously we had RP apo, robotic Process Automation where, you know, yeah, we could already automate a lot of that, but with ai, you know, the big thing and like what we launched load partner with originally is like we can go to where the problem is now.

And, um, you know, where you have emails with like all this, you know, uh, gray area before of like, how do you decipher what somebody's saying in an email and match that to a problem, and like, who does that go to? And we don't have, we can solve that now. So I think, you know, moving forward, AI's gonna continue to do really good at those like high volume tasks.

I think as far as fully automating a brokerage and, um, getting into that, hey, can we, can we actually end to end automate this with AI and, and lower the margins by a bunch? I, my take is honestly, we've seen this tribe before. Um. You know, with, uh, there, there's a bunch of companies that have attempted to fully automate a brokerage and say like, Hey, we're just gonna connect the shippers and the carriers, and like, the carriers are gonna be right here, the shippers be right here.

We'll just automate that conversation, that flow. The tracking, the alerts, AI was already not really needed there because they had full control of the platform and we still saw a lot of problems. And so I do feel like we're kind of going into probably another bubble of people saying like, Hey, it's the end of brokerage as we know it.

AI is gonna come in and run freight now. Yeah. Um, we've seen, we've already seen this once, like this, this has already happened before. Um, is it, are we gonna like move the needle and get closer? Probably. I mean, definitely I'd say. But I don't think we're really to the point of it being fully automated yet.

Um, there's just too much nuance. And at the end of the day, starting your own brokerage isn't that hard. So even when you get to that point of automating tons and tons of freight, you have to compete with that guy who's willing to pay a hundred dollars less to book a carrier to take that load from you.

And that little brokerage, that solo guy is gonna beat out the AI 99% of the time. And, um, I think this, that piece of it is just gonna be why. I don't foresee huge brokerages necessarily using AI to kind of like, take over the industry basically. So, I mean, I guess all that's to say, I, I do think AI is going to push the industry forward.

It's gonna be an improvement. We're gonna see, um, a lot more, a lot less overhead on certain things, but the people that think it's gonna be like full, we're, we're on the path to full automation and the platforms are happening. I think we're just repeating the cycle a bit of, uh, yeah, trying to do better, trying to automate it.

And we're gonna come out the other side closer, but not there yet.

Andrew Verboncouer: Yeah, when you see like dat, like did they just bought Convoy, right? Or completed? I don't know. Started the purchase completed? I don't, I don't know what the right verbiage is, but um, yeah, it's interesting to see the movement in the space between, you know, what they were doing.

Obviously like Tech first to, was this the third owner of Convoy? I don't know what's, I think so, yeah. Um, what the path is there, you know, are they becoming a full brokerage? I don't know.

Garrett Allen: Yeah, I know people have been back and forth on that. I think Craig Fuller made a great post, um, the other day about why DAT probably won't be a, a full brokerage.

Like that was my initial reaction. I'll, I'll be completely honest, like I saw this and I was like, oh my gosh. Like anybody using DATs gotta be freaking out, right? Because now they're competing. Um, but they're not really doing that. I think they are gonna just be, they're gonna be giving those tools to the brokers and in the attempt of.

Like, Hey, we still just wanna be the, the provider of data and tools for brokerages to do better, but we don't wanna be a brokerage. And Craig Fuller made a great point of, you know, hey, look at DATs valuation and then look at like the top 100 brokerage valuations like DAT is worth. A ton, even though their revenue is not anywhere in the ballpark of these brokerages, but yet that's worth more because of how they play being like that provider of tools.

Like why would they not, why would they like change that? Like they're doing great. Yeah, yeah. They don't want to get into being a brokerage. And I think that's a great point. Why I Getting into a lower

Andrew Verboncouer: margin business?

Garrett Allen: Yeah.

Andrew Verboncouer: Yeah, exactly. Yeah. I, yeah, I think it like probably commonality too with like Uber Freight launching, you know, broker access and some of their other stuff is like.

Tooling to help them leverage the data and systems and, you know, focus on the SaaS and all of that.

Garrett Allen: Yeah, yeah. And I know, uh, it's, it's interesting 'cause there's just so many things going on right now in the space with, you know, load boards are currently like the big frontier I think. And so people are trying to do things to, I.

You know, just that marketplace side of things is really like heating up, I think more than, more than it's been in a few years of, I think Highway just launched their, their, was it the trusted, uh, freight exchange thing last week on Friday? Yeah. So, you know, you've got like a load board through Highway Now and uh, I think Truck Stop just launched like their private load board now.

So you can actually put like, loads up that only are like approved through RMIS carriers can look at them. And so, you know, I think it's good. It's a, it's all a step in the path of like fighting fraud 'cause. Fraud in the industry is just completely outta hand. It's a mess out there. And so, uh, I don't know that any of those are, and I don't think any of those are like the silver bullet, but they're definitely gonna move the needle to make things better, which is great to see.

Andrew Verboncouer: Yeah, I mean, uh, I think when you come at it, a little bit of improvement at each angle, you lower the risk quite a bit. What do you think is, or what recommendations do you have for folks after building a few tools with ai? Um, you know, I'm sure you were, uh, privy to some of it before when you were at Transfix, like not.

Full ai, but also at at IEL. Um, what advice do you have for brokerages and, you know, just teams in tech starting to leverage AI into their operations? Like, how have you thought about, um, utilizing it as a tool rather than, you know, thinking it's gonna be the holy grail and it's gonna solve all the problems?

Garrett Allen: Yeah, I mean, I think you just have to be really intentional about it. Um, at the end of the day, like I said, it's really good at. Pulling like structured data out of unstructured like conversations and things of that nature and, and figuring out like where are places that you couldn't automate before, uh, that AI could.

Basically take you a step closer to automating, if not automating. Yeah. Um, a lot of people when like the, when you read about people, they're like, oh, AI's not gonna do it. It's not, well, it's not doing that great, blah, blah, blah. The use cases usually see are like, yeah, we put AI in charge of like booking the load and also doing track and traces and like nobody was there looking at it.

We just expected the AI to flag it to us. Those use cases where you're just like, Hey, AI is just gonna do all of it, and like, hopefully it lets us know if something goes wrong. That's where you run into a lot of issues. I think the right approach is really looking at it and saying like, Hey, um, somebody's still gonna be in here reviewing it every day, but instead of having to look through like the full details of a load, like the AI is just gonna have a, like a quick summary for them to look at and like the risk factors and like, or the priority of things they're looking at, and like really used as a tool to assist a human to perform at a higher level.

And maybe that's the, maybe that's the key is really about, you know, helping a person perform their job better or more efficiently rather than completely just replacing them. Um, I think that's, that's definitely a big piece of it. So, um, yeah.

Andrew Verboncouer: Yeah. I mean, I think that's how you build trust in, in what, what it's doing as well.

Right. Like optimization solutions. And I was talking with Shaman from Optum about it is like if you just give someone a black box and an answer like. They don't know whether to trust it or not, right? Because humans are used to being logical and saying, well, I would've arrived at this. Why? Why am I, you know, being told this is the answer?

What's the insight? What's the reason? What are some of those edges? And I feel like the true human in the loop, you know, that you're used to with AI is like really the stepping stone to that. I mean, obviously there's a path here where. You know, warehouses are robotic drivers, you know, trucks are driverless.

Who knows if that's gonna be reality with everything with Nicola and, you know, Tesla trucks and all the, all the stuff. But there is a future, you know, far off be it that could be automated, that systems could just give an answer. And, but I think we're a ways from that. And so, you know, how do you make that experience more human and more understandable?

I think that's, that's, uh, that's a good insight. Um, oh, go ahead.

Garrett Allen: I was gonna say like the, uh, I, I think one of the other things too is like just being able to like control the environment a little bit too, right? So like when you look at things like the self-driving trucks, like. For example, if you're just automating the trailer in the yard and like moving the trailer around, things like that, you have a lot more control over the environment and like the whole scenario.

Yeah. So those are like a lot easier to automate. And I think that same stuff applies to processes you have internally at like a brokerage. Like what are, what are scenarios where like. Everyone that touches this piece end to end is like an internal employee that you can control how they normally do things versus like, what's a scenario where like a customer might email in a request or a carrier might ask for something that you don't know what that possibly could be like.

Those differences are, are gonna be where, uh, if you just let AI go and you have a carrier emails in and ask for like a reschedule on a certain thing, like who knows how that AI is gonna react because you don't have the full definition. So, um, yeah. Yeah.

Andrew Verboncouer: The one, the one product that was, that really impressed me last year, we went to, I think it was F three.

Um, well, happy Robot gave a demo, like a live demo. I was like, this is a, I thought it was a lot farther off from any workflow, but like the way that they had trained and architected that to say like, oh, if somebody requests a lower price, I'm gonna send a Slack message to this person who approves it and get back and then have that feel like a seamless experience.

Like if I got that call, I wouldn't be able to tell. I mean, maybe. Maybe if I was doing it all the time, but, um, I dunno. Are there, are there other tools that you've seen in this space that you think are kind of pushing the envelope of what's possible with AI and like getting people more and more comfortable with this notion?

Garrett Allen: Yeah. I mean, yeah, that demo's a great example of like that human and loop like we're talking about, right? And like it being able to like reach out and like have that human, like, I think that's, that's fantastic. I mean, other, other tools in the space as far as like AI specific, I don't know, it feels like there's, it's, it's really diluted, I would say.

Um, there's just, there's a lot of people like coming out and like trying to offer new things that it's hard to know like what's real and what's not. A little bit, I think, yeah, because of like the demos you see that are demos. Um, so. Yeah, I don't know. I, I guess I don't really have any specific callouts, honestly.

I think we're sort of in like that. We, we went through like the, the intro phase where everybody's like leaps and bounds, right? Like the F three demos. You're like, oh my gosh, that's crazy. Yeah. I think we've plateaued a little bit waiting for that next sort of breakthrough. Um, I don't know what that's gonna be, honestly.

Um, I think having the platforms in place where, um, people like that are starting to try to control the whole thing, that might open some doors. We'll see.

Andrew Verboncouer: Yeah. Well, I mean, it's like, it's just a tool, right? It, it depends on how you use it. It's the difference between, you know, somebody having a sharp knife and someone being a, you know, five star Michelin chef, right?

Like, they're probably using the same knife, but the outcome's different, right? So it's like, how do you leverage AI and use it in a way? And I, I think a lot of it is just the nuance of, you know, the problem you're solving, understanding the edge cases, being able to feed it the right data and context that we've talked about, um, without that like.

It's just another feature, you know, it gives me a capability. It doesn't mean it's bene more beneficial or not.

Garrett Allen: Yeah, I think a lot of the tools that I like that I've seen are ones that they have a lot of that like no code kind of built in. Um, where you're seeing like some of the brokerages, like Ally Logistics, they're a great one.

Uh, where you see those, the everybody over there is trying to like build these different workflows. You see Dan Manin posting all the time about. Um, you know, the people on their team building entire workflows themselves that aren't like coders necessarily. Yeah. But they just have these tools with AI baked in where they can point it where they need it to and start automating workflows without it being some big custom build.

I think that's kind of the future of giving, you know, just like the, the load partner TMS, right. Giving tools to people that, Hey, you wanna go build that thing and you're the person with that problem. Like, we want you to be able to go build that solution. 'cause you're gonna know how to solve it the best.

Andrew Verboncouer: Yeah, I mean it's like Zapier for work, right? Like, uh, and I don't necessarily like Zapier's like, you know, step based approach, but that workflow software where, so you can give someone the tools to go ahead and, I mean even now they're getting really good with like, um, I'm trying to think of the example.

A Paragon is another one that's like an aggregator where you can like describe a problem or describe a workflow and it'll create the first version for you. I mean, if you can do that with your own business data, like now, those tools really start to become valuable.

Garrett Allen: Yeah, totally. Uh, I definitely agree and I, I, it's gonna be cool to see.

I mean, I, like I said, I think, I think the next, we're gonna have a new breakthrough soon. I think. Like, it just feels like something's coming. We're like, we, we've plateaued for like six months or so, I feel like now. So the, the AI voice and the text are getting so good. Like you said, like you can barely even tell on a good AI voice call.

It's like you wouldn't even know. Um, and so like where do we go from there? Um, I don't know.

Andrew Verboncouer: Yeah. I don't know either. Awesome. Well, thanks so much, Kiera, for being on the show. I, I think, uh, to kind of wrap us up, where can people stay in touch with you and find out more about what you're building at Load partner and search carriers and all the other micro SAEs that that may be coming down the pike?

Garrett Allen: Yeah, yeah. So you can find me obviously on like LinkedIn, uh, just Garrett, Allen, um, and then also on X at Garrett makes, um, you can visit, uh, garrett makes.com. I have a website up with some of links to different projects. And then for, if you need tms, go to load partner.io. If you're trying to find carriers, go to search carriers.com.

And, uh, yeah, that's, that's where you can find me.

Andrew Verboncouer: Thanks for joining us on this side up. Don't forget to subscribe, rate and review our podcast to help others find us. If you have any questions or topics you'd like us to cover, feel free to reach out and if you want to be a guest, let's connect. Until next time, keep your shipments safe, your logistics smooth, and your curiosity on the move.

‍