The Freight Pod
The Freight Pod is a deep dive into the journeys of the transportation and logistics industry’s brightest minds and innovators. The show is hosted by Andrew Silver, former founder and CEO of MoLo Solutions, one of the fastest-growing freight brokerages in the industry. His guests will be CEOs, founders, executives, and leaders from some of the most successful freight brokerages, trucking companies, manufacturers, and technology companies that support this great industry. Andrew will interview his guests with a focus on their life and how they got to where they are today, unlocking the key ingredients that helped them develop into the leaders they are now. He will also bring to light the fascinating stories that helped mold and shape his experiences.
The Freight Pod
Ep. #56: Paul Singer, CEO & Cofounder of FleetWorks
Andrew welcomes Paul Singer, CEO and cofounder of FleetWorks, a carrier-facing voice AI agent. It’s been a year and a half since Paul and his cofounder Quang Tran tested the idea for FleetWorks. They signed up for a load board, posted an attractive load, and to their surprise, three carriers responded and negotiated with their AI agent — a product Paul calls “janky” compared to what they have today. “Ever since those three calls, we’ve been off to the races.”
On this episode, Andrew and Paul cover:
- The recent DeepSeek announcement and why it’s a foundational moment for AI
- Paul’s biggest takeaways from his four years working in carrier product for Uber Freight
- What separates FleetWorks from other AI companies and how it's returning 10x the ROI in operating costs for its broker-customers
- The biggest challenges the FleetWorks team has faced so far
- Can a brokerage with a sophisticated tech team build AI on its own?
Follow The Freight Pod and host Andrew Silver on LinkedIn.
*** This episode is brought to you by Rapido Solutions Group. I had the pleasure of working with Danny Frisco and Roberto Icaza at Coyote, as well as being a client of theirs more recently at MoLo. Their team does a great job supplying nearshore talent to brokers, carriers, and technology providers to handle any role necessary, be it customer or carrier support, back office, or tech services. Visit gorapido.com to learn more. ***
A special thanks to our additional sponsors:
- Cargado – Cargado is the first platform that connects logistics companies and trucking companies that move freight into and out of Mexico. Visit cargado.com to learn more.
- Greenscreens.ai – Greenscreens.ai is the AI-powered pricing and market intelligence tool transforming how freight brokers price freight. Visit greenscreens.ai/freightpod today!
- Metafora – Metafora is a technology consulting firm that has delivered value for over a decade to brokers, shippers, carriers, private equity firms, and freight tech companies. Check them out at metafora.net. ***
Hey FreightPod listeners. Before we get started today, let's do a quick shout out to our sponsor, rapido Solutions Group. Rapido connects logistics and supply chain organizations in North America with the best near shore talent to scale efficiently and deliver superior customer service. Rapido works with businesses from all sides of the logistics industry. This includes brokers, carriers and logistics software companies. This includes brokers, carriers and logistics software companies. Rapido builds out teams with roles across customer and carrier sales and support, back office administration and technology services. The team at Rapido knows logistics and people. It's what sets them apart. Rapido is driven by an inside knowledge of how to recruit, hire and train within the industry and a passion to build better solutions for success. The team is led by CEO Danny Frisco and COO Roberto Lacazza, two guys I've worked with from my earliest days in the industry at Coyote. I have a long history with them and I trust them. I've even been a customer of theirs at Molo and let me tell you they made our business better. In the current market, where everyone's trying to do more with less and save money, solutions like Rapido are a great place to start To learn more. Check them out at gorapidocom. That's gorapidocom, all right.
Speaker 1:Welcome back to another episode of FreightPod. I'm your host, andrew Silver. I am joined today. We are in the deep dive of AI. I've got another tech startup in the space, this one founded by a gentleman who has some industry experience, some very interesting experience. My guest today is Mr Paul Singer, the CEO and co-founder of Fleetworks. Paul, welcome to the show. How are you doing? I'm doing well, man. Thank you for having me. Of course, you're in an interesting space. This voice AI stuff. It's seemingly every week we learn something new. That's I don't know, maybe game changing. Let's start with the deep seek stuff. As somebody who actually understands what's going on with all this, give me your kind of understanding of what the world just went through with DeepSeek, announcing that it had created what it did on such minimal funding. Whether that's true or false, I'm just curious your thoughts there before we get into your story.
Speaker 2:Yeah, I don't know if people have actually looked into how verified those claims are. It seems to be true. Yeah, I mean for context on folks who haven't seen this so a Chinese hedge fund that had sort of an AI research department basically created a really high-end what we call a reasoning model with far fewer resources than OpenAI Anthropic or any of the sort of big foundational AI companies. So typically it costs, you know, hundreds of millions of dollars to train these models. Deepseek claims to have done it with, I think they said, $6 million. So super impactful for NVIDIA, because every investor assumed that NVIDIA is just going to have infinite demand with sort of these AI companies training their models, with sort of these AI companies training their models. Turns out, you know, deepseat did it basically with just a few sort of lower-end chips and so the implications for kind of the whole infrastructure of AIs is huge. Obviously Implication for OpenAI Anthropic is really big.
Speaker 2:I have friends who work at all those companies and you know there's certainly a kind of a moment of reckoning for all these companies. They're like wow, wow, we aren't as protected as we thought. For companies like us, we are the application layer on top of AI. It's really exciting. Typically for us, we need to assess what is the right model to use for any given task, and so we'll use, you know, we'll use open source models that we kind of fine tune for our own use cases. We'll use sort of foundational models for some parts of our product. We haven't yet used these reasoning models because, number one, they're a little bit slow. They're designed to be a little bit slower, but they're also really expensive, and so DeepSeek basically takes that whole expensive thing out of the equation. I think the numbers I've seen is it's 97% cheaper than using OpenAI's reasoning model.
Speaker 1:Can you explain the difference between? You mentioned slowness, but the difference between what you get from a reasoning model versus what you're using.
Speaker 2:Yeah, absolutely so. A reasoning model basically has built in what we call or what's called chain of thought reasoning, and so basically when you submit a prompt or you submit a request for an AI and a lot of folks who are listening have already used ChatGPT so you see, chatgpt kind of thinking out loud how it works and so a reasoning model kind of just does that 10x. So it's going to really spend a lot of time going through every possible avenue before it outputs a result. So the results tend to be a lot of time going through every possible avenue before it outputs a result. So the results tend to be a lot more accurate. But the downside is that it can A cost a lot of time and since we do a lot of things on phone, email, text message, real-time communication, it's not a great application for what we do today. I think, given the way that this tech is advancing, maybe in a year or two we'll be able to use reasoning models in these sort of dynamic use cases.
Speaker 1:Got it Okay, and what do you see as the overarching impact? Is this a foundational moment? Is this deep seek announcement a foundational moment in AI? And I know you're going to be somewhat guessing, or at least it's an educated expectation or guess but is this a foundational moment long? What do you see as the long-term impact to the open ais of the world who have invested billions um and and maybe disrupted by a much cheaper alternative?
Speaker 2:yeah, it's a good question. So I kind of view this more like a covid moment for ai, where I think COVID really accelerated digital adoption for a lot of things. Every digital platform did really well in COVID, so the rate of growth really accelerated and then it sort of normalized back to what it was pre-COVID. But COVID kind of accelerated adoption of digital things, whether it be like a, you know, like a toast terminal at like a local restaurant that you go to, or like using Airbnb versus like calling a hotel to book a booking, for example so just accelerated digital adoption. I don't think this is anything new, though. Like frankly, what we do see every year in foundational AI models is we do see step change improvements. So I think this is more of a COVID moment, where the step change was quicker than people anticipated, but the way that we're all modeling these improvements is that it's just going to continue to really if not linearly, maybe even exponentially improve these models.
Speaker 1:And last question here before we're going to get into your background and then we'll go through the whole linear story how, as a CEO of a up-and-coming AI business, do you factor in these types of step change improvements that can even be exponential? How do you factor that into your own product development and kind of company evolution? How do you take that into account?
Speaker 2:yeah, it's a good question I we think about in a couple of ways. So, um, with these foundational models, I would kind of think about them on three different axes basically. So there's speed, there's power and there's cost and with all these models, and probably in perpetuity, there's some trade-off to be made between speed, power and cost. Like we talked about reasoning models being really powerful and more expensive and DeepSeek sort of changes, that dynamic but for sure they're slower. Right Versus you have these mini models that are cheaper, faster and way less powerful.
Speaker 2:Evolving space. I would say we're pretty consistently evaluating what's the right trade-off to make at any given time, given the advances in those foundational technologies. Ultimately, I think we've found the right balance for now. But essentially what? Basically what's going to happen is we're going to keep upgrading our foundational models and they're going to be more powerful. For example, as you think about carrier negotiation, today we do a lot of guiding of the AI to make sure that's really successful, but in the future the AI might be able to do that more autonomously for the same or even cheaper cost today. So basically we just assume that all the foundational models will get cheaper, more powerful and faster. I would say every nine months is probably the right cadence.
Speaker 1:And is it easy to sorry, I'm going to stay on this for a second but is it easy to, let's say you are using, you're using Model X, and Model Y comes out it's a different company, whole different thing and you just look at it and say this is better than what we're using today? How easy is it to just pick up your product off of Model X and dump it on Model Y and then you're just off and running.
Speaker 2:On a high level. It's super easy. It's basically just changing an API key. I think this is getting a little bit more detailed about what we do, but customers trust Fleetworks to be their front of house. Freight brokers, when they use Fleetworks, are saying I'm going to trust Fleetworks to talk to my carriers, which is one of the most important things that we do as a brokerage. So it would be, I think, foolish of us if we were to just say we're going to swap from OpenAI to Anthropic or OpenAI to DeepSeek, because while switching is super easy, there might be some unintended consequences because all these models perform slightly differently. So if we were to do that, we have rigorous testing in place to make sure that when we make a model change, that it's passing all the criteria that we would expect, so we're not like creating bad experiences for carriers and brokers.
Speaker 1:What's an example of an unintended consequence that could come up.
Speaker 2:If you didn't do it right.
Speaker 1:I'm not saying this has happened to you. I'm just curious Totally.
Speaker 2:There's probably dozens. One really easy example would be, I think, around what we call hallucination rate, so the AI's propensity to make up answers to things. We've done a lot of things on our side to reduce and eliminate hallucination rate. But one great example would be Kara calls you about an available load. Let's say they're a dedicated carrier. They say, hey, I run this lane regularly for you. Can you just send this over? Do you still have that load for Monday? Dallas to Chattanooga, I'll do it for $1,000, like usual. And the AI says, instead of saying yes, I'll send over right now it says, oh, no, we need $900 on this one. That could be an unintended consequence where you want to honor that rate on file with the carrier, but because you swapped the model and because you don't know how it's going to perform, maybe it's going to do some weird things for you. So those are the kinds of things that we test for whenever we think about a change.
Speaker 1:And that's an example of a hallucination rate where the history and guidance provided in the model is for. It should know that Dallas to Chattanooga has been run by this carrier eight times for $1,000. It's an approved rate in our process and it will just hallucinate the idea that it should try to get cheaper and then go for 900.
Speaker 2:That's one great example. I think another one could be appointment times. So Kara says, hey, this load's supposed to pick up at noon. Any chance we could shift it over to 2 pm. Sometimes we know we can, sometimes we don't know. The last thing we want to do is tell the carrier yes, we can when we're not sure. So that's just another example of how we prevent hallucinations, so something that we're pretty careful about.
Speaker 1:Interesting. How do you catch the hallucinations happening? Because that's not a super crazy situation. If I look at a brokerage without Fleetworks or any tool like this, that type of conversation whether it's a mistake or the right answer is happening all day, where the carrier is asking for a rate or the broker is responding with an approval or not enough. And how do you even notice that those are hallucinations?
Speaker 2:Yeah, we could kind of get really deep into this, but I think AI totally changes how you build companies. So typically when you build technology, you have really rigid what's called test cases, like basically like a source of truth, Like if we expect if X happens, it should always end up as Y, and so the cool thing that AI enables you to do is actually have your test cases be a lot more fluid, which means you can create like you can create hundreds of thousands of test cases. So we're constantly on every single call that we're running and every single email, every single piece of communication, we run test cases against those, and so we have basically AI evaluating AI. So we'll run the call or run the email through another AI and says this was good for one, two, three reasons, or this was bad for X, Y, Z reasons. Therefore, you should go and change the original model. That's how we think about kind of the AI self-learning and the AI self-evaluating.
Speaker 1:But even in those examples, is there not a person who has to tell the AI this is good or the AI is capable of? At some point? There's a person who's got to be giving the input of yes or no, good or bad, right.
Speaker 2:Yeah, we build the test cases. So we build the test cases and then we let the AI run at it.
Speaker 1:Okay, Got it Understood. All right, Now let's go back. So I appreciate you kind of. I don't know this conversation could go deep into AI for a number of reasons.
Speaker 2:I'm down. When did you first? When I left Uber I think it was GPT-3 had just come out, or basically ChatGPT had just launched, and the technology was super simple, or at least the outcomes were simple. You could give it an email and you would tell it hey, make this more professional or make this funnier, write me a joke. There was none of the infrastructure built to be able to build what we have today.
Speaker 2:I started playing around with it. I would be lying if I said that we had this idea at the time. We were just playing around with AI. At the time, I was building a carrier-facing product. We thought that there could be a better driver workforce management product for carriers, because drivers are churning all the time. They don't have good visibility on what their pay looks like. We were building in that space. It was a total flop. It was a really bad idea. That space it was a total flop. Uh, it was like a really bad idea. Um, frankly, like after talking with probably 300 drivers over the course of three months, we're like this is, drivers don't actually care and they're fine with with how things are. Um, so we went back to the drawing board and I would say it was about a year and a half ago, year and a half ago where you could build a viable voice AI agent Barely viable, I would say.
Speaker 2:My co-founder and I, we were playing around with various ideas. I remembered back in my time at Uber that, despite our app we had built this app, carriers were using it, we were getting a ton of bookings through it, but still about a quarter of our bookings were done off of Collins, gatt Collins, and so I was like, hey, like, if we could build a voice AI agent to at least like, screen the carrier and share the details of the load, that could be pretty helpful for a lot of brokerages. So we built a pretty scrappy prototype. We found a load board that let anybody sign up and post loads, and we signed up for that load board and posted just one load that we knew would generate a bunch of interest. And this was straight up one of the worst I think one of the worst products we'd ever built. It was janky, the lag was bad, the voice quality was bad, it was nothing like what we have today, and I remember we got three carrier calls and on two of them the carrier knew they were talking to an AI and they still went through the whole call and on one of them the carrier didn't even know they're talking to an AI and negotiated. And after that, those three calls, we kind of looked at each other. We're like this is real With this product that we built, that is bad on every dimension. It's working Because we were first in the space. It was such a big arc to convince people that, hey, like this thing actually works, like carriers are are down for this, um and and I remember we we had a conversation a while back like hey, do you care? Like what do carriers even think about this? Um, and I mean I think we wouldn't even have a company if carriers didn't mind the process. I think ever since those three calls, we've really just been off to the races, and what's been cool, as we've kind of talked about, is the underlying technology has just gotten better faster than anyone thought, and so what we can do for our customers has just continued to expand over the last year and change since we launched the product. Who's we? We is an ever-changing definition.
Speaker 2:My co-founder and I met through mutual friends in New York. His name is Kwong. He used to lead the experimental engineering team at Airbnb. It's been a great partnership. I feel very blessed, first of all, just to be working with him, and then I think it's just a well-known secret that good talent attracts good talent, and so we've got eight more people in San Francisco, a few more people remote. Every technical person on the team is either a current founder or a former founder, so everyone on the team is a former CTO. You've got myself, you have a couple of other markets.
Speaker 1:Everyone on the team is a former CTO of a business. Yes, that's wild, of 10 people, 10 people are all former CTOs, essentially.
Speaker 2:Yeah, we have seven technical people and seven CTOs.
Speaker 1:Wow, very cool, yeah. And so what was the timeline of when you built this product and made three calls on a load board?
Speaker 2:How long did it take us to do those first three? Well, when was that?
Speaker 1:This was when was this?
Speaker 2:This was, would have been summer 2023, so call it like, uh, like a year and a half ago okay, yeah, and interesting anecdote.
Speaker 1:You mentioned that. So you went, you worked at uber for five years. We should talk a little bit about that and maybe let's do that now to get that out of the way. Um, how did you get to Uber? What was the? You were trying to work in freight or you just saw the cool Uber name? What was the story there?
Speaker 2:I knew literally nothing about freight Seven years ago. I was like, well, kind of introduced me, so I started my career in management consulting, which was great training ground, but generally I would not recommend it. I think, if you're someone that wants to start in management consulting, if was great training ground, but generally I would not recommend it. I think if you're someone that wants to start a management consulting, if you're someone who's like, let's say, 20, 21, 22, listening to this podcast, I would recommend that. If you're someone who thinks you're a hustler and you want to be analytical, go learn a little bit of SQL, go learn a little bit of data analysis. A little bit of data analysis. And cold email 30 startup CEOs and say, hey, I will do anything for you and I bet you like 20 out of 30 are going to respond and you'll probably get an interview with at least half of them. So I'd really recommend that career path as an aside. So I did consulting for a couple years, met some really smart people, learned a lot of stuff, but did two things that kind of elevated me to Uber Freight. So first was I did a consulting case on industrial metal doors. This was back in 2017.
Speaker 2:Turns out, the industrial metal doors industry is fascinating. There's like three companies. Two of them are private equity owned, one of them every PE company wants to own but the guy refuses to sell. So it's like a funny, like a little little dynamic in this, in this niche space. And, unsurprisingly, the biggest part of the industrial doors market is warehousing. Right, like every warehouse needs whatever dozens of doors.
Speaker 2:And so I was like this is kind of kind of this is interesting. Like I I grew up in suburban connecticut like I've never, I've never, I don't hadn't seen a warehouse at that point. Um, and then a couple months, so I was like okay, like so trucking is big, warehousing is big, that's cool. I didn't really think that much about it. Um, then a couple months later, the partners of lek came to me and said hey, we're writing a white paper on Uber Freight. Uber Freight had just launched, maybe six months earlier. Do you want to help us do research for the white paper? I was sort of between projects at the time. We call that being on the beach and you kind of just help the partners with research and development.
Speaker 1:I thought it was called being on the bench. It's being on the beach.
Speaker 2:Being on the beach, being on the beach yeah, I thought it was on the bench no, it's being on the beach.
Speaker 2:It's kind of a reward for, uh, a lot of hard work, like once you're like on a project for like four months. They're like you get to go to the beach for like a week. Right, um, got it. So. So I was beached and I was doing this research and I was like this is crazy. Uber rides was at the time, still exploding, growing double digit percentage year over year, off of tens of billions of dollars, and Uber Freight was starting from nothing. At that point they were maybe 100 loads a day, 200 loads a day, and our thesis was that Uber Freight would not grow as quickly as Uber rides, which we got right. But the reason why we got wrong? So we thought that because the industry is slower to adopt technology, which I think was wrong. The industry's actually pretty quick at adopting technology if it makes sense.
Speaker 1:If they want to, yeah Right, yeah if the ROI is there.
Speaker 2:They were like oh, we thought that because the industry's slow at adopting tech, uber rides or Uber Freight's not going to grow as fast as Uber rides, so it did not grow as fast. But the reason is because Uber Freight ultimately was providing a somewhat commoditized service to shippers and no shipper is incentivized to give any single brokerage all or most of their wallet share. So Uber Freight's thesis was hey, we can procure carriers for less, therefore we're going to win more freight from shippers. I don't think that thesis proved to be true and I think Uber buying TransPlace was a great move on their part, because ultimately, that is kind of how you tap into that shipper wallet share in a different way.
Speaker 2:Ultimately, my experience at Uber Freight was just amazing. I mean super smart people, really passionate. I think Uber did a great job of bringing in talent who otherwise might not have worked in logistics, and you've seen now so many startups come out of Uber Freight, us being one of them. You've got Loop, you've got TruckSmarter, you have a bunch of others as well. So really, really great group of people and I wouldn't trade that experience for anything.
Speaker 1:What would you say have been the best takeaways you've had from that experience, or learnings from the four to five years you spent at that business?
Speaker 2:I think there's some personal stuff and there's some professional stuff. So on the personal side of things, I very quickly was like holy shit, this industry is crazy. I was just blown away by the amount of problems. So my first project was the ELD mandate had kind of just gone into effect. That was my first project and I was supposed to come up with a strategy for how we were going to engage with the ELD providers and the ELD aggregators. So we built that strategy and we executed on it. We ended up partnering with Samsara, Keep Trucking, P44, MacroPoint, like all the big players.
Speaker 2:I really loved building and working with small carriers. Like for me, that was. That was that's what kept me motivated and that's what kept a lot of people at Uber Freight really motivated. Um, I had, I had the personal phone numbers of probably a dozen carriers. We would talk regularly every single week. Um, and it was really because of that experience that, that experience that I got the chance to go from operations to strategy to product and ultimately lead a big chunk of the carrier product team. I just love building products for carriers.
Speaker 2:I think empowering those guys who, as a tangent fraud, is a really hot topic right now. Everyone is really concerned about fraud, as a tangent, fraud is a really hot topic right now. Everyone is really concerned about fraud, as they should be. I really personally think it's still important to remember that, hey, 99% of carriers out there are pretty honest, hardworking people who just want to provide good service. And so, for me, what gets me excited, even today at Fleetworks, is how do we help those people make more money, how do we empower those people, how do we continue to be a carrier-first company? I think that's what Uber Freight instilled really well. Personally, for me, it just felt like, you know, Uber gets a lot of knocks, but internally, I will say that every decision we made was through the lens of how can we create a great carrier experience, Because that was our differentiator. That's how we got so many carriers was carriers liked the Uber Freight app Like they really did.
Speaker 1:Yeah, that's great. What would you say on the flip side were, like, the biggest misses or biggest challenges that you endured or noticed while there, or why you think it didn't maybe reach a level that it might have wanted to yeah, um, we spent a lot of time thinking about that and that's kind of some of the foundation for, I think, why flea works should really exist.
Speaker 2:To be honest, um, I think it's really hard to digitize freight through digital channels, and what I mean by that is, like when you have an application, when you have like a website or when you have a mobile app, you can very quickly get a lot of demand for that, as Uber does. Right, like I mean, at this point, tens of thousands, if not hundreds of thousands, of drivers are sessioning in the Uber app every single month. The problem is those guys aren't very sticky. They tend to treat Uber Freight just like they might any other broker or just like they might DAT. Right, it's just a transactional place to get freight.
Speaker 2:I think where Uber kind of missed the ball is sort of two pieces. Number one is the carrier relationship. Like how do I understand at a really deep level who this carrier is, where they like to run, like what makes that carrier tick and like how can I make that carrier successful? I think a great carrier rep, a great individual rep, can do that, but they're sort of limited by just the amount of neurons in their brain and the amount of time that they have. In a day Like me when I started on the carrier floor and I was only there very briefly you can maybe keep 10 or 20 relationships in your mind or in a spreadsheet. It's pretty hard to scale beyond that. So I think that was one miss is Uber thought they could scale that relationship and they never really invested in it. And then number two was that this is a little counterintuitive but it's obvious if you've been at brokerage.
Speaker 2:Uber did really well on short haul freight your intrastate Texas, intrastate California, regional Southeast.
Speaker 2:Uber crushed those lanes and Uber outper regional Southeast.
Speaker 2:Uber crushed those lanes and Uber outperformed a lot of brokers on those lanes.
Speaker 2:But, as you know, those are not good paying loads.
Speaker 2:You might get 200 bucks on that load and Uber might beat out some mom and pop brokerage beat them by 10 bucks, but Uber's making $30 of margin on these $200 loads and suddenly, when you add any operations into that gross margin equation, uber's just losing money on that freight right.
Speaker 2:And so where a lot of brokers tend to go is they build density with short haul but then they invest a lot in this long haul freight where it's high margin, high top line but a little bit higher touch, higher service, really hard to cover through what I'll call digital channels, and so they rely on phone and email and text message and WhatsApp and Google chat and all these ways to procure capacity that Uber just wasn't going to invest in. And so Uber was really bad at long haul freight and really really good at short haul freight, so they kind of started to do a lot more short haul freight and it was sort of this weird negative selection problem where they weren't able to really optimize that long haul freight that's going to drive most of the margin for the business.
Speaker 1:So just an interesting thought. I had listening to that. There was a point when Convoy went under that I was, I guess, arguing with my old co-founder, matt Bogrich, about Convoy and what I was saying that I had heard was that Convoy had secured like exceptional automation and higher than industry average margin percent on some of these dedicated lanes. Some of these dedicated lanes and the higher than industry average margin percent was throwing me for a loop and throwing both of us for a loop where he was just like I don't believe it, I don't, I think that's just a lie and I was like I think it's actually true. But the only way it makes sense is if they were really good and really heavy on short haul lanes that were like 200 a load and you could be at 30 or 40, $40 a load in margin, which does not cover your cost. It's not enough money, but on a margin percent it looks really good 15%, 17%, 18%, whatever it may be.
Speaker 2:It sounds like Uber in some respects was in a similar situation Exactly, and it makes total sense when you think about Uber rides. If you go to New York City and you call an Uber, it's a four-minute wait time. The marketplace is really dense, it's really good and things tend to be really efficient. Let's say you go to I don't know Birmingham, Alabama, small city. You call an Uber, it could be 20 minutes away. Like, the marketplace just gets worse and worse the less dense it is, and so it totally makes sense why a mobile app-based brokerage would suffer similar problems as Uber rides.
Speaker 1:Yeah, I made this analogy, I think, talking to another AI business and just the idea that the marketplace freight matching concept. It doesn't work if you think about it like a Tinder, in that Tinder's job would be to match all of the people with all of the people. Tinder and Hinge or whatever they're good at getting some people matched and then those people are gone. But if your job, like your job is as a brokerage, is to execute every order I've taken from a customer, which means that in the Tinder analogy, tinder's only successful if they match everybody. And the reality is some people don't have a lot of matches and whatever the reasons are, it doesn't matter. But there are reasons like loads. If a load is really ugly, it's not getting matched easily. No one's showing up and wanting to take it unless you start offering to pay asinine amounts of money. So I don't know. I just think that's interesting. We don't have to go much further on that Other than, I guess, to talk a little bit about kind of your takeaway in starting Fleetworks. You know, one of the comments you made earlier was that Uber still had something like up to a quarter of its loads being covered by Collins from boards like the DAT, which validates a point that I've made before, which is nearly every brokerage still has to use load boards, and shippers don't like to hear that, but it's a reality.
Speaker 1:And just because you use a load board doesn't mean that's a bad thing. I think that it's all about what you do with the carriers that call you, and it's what is your strategy. If your strategy is to simply post and pray and rely entirely on the load board and hope you find good carriers, that's not a good strategy. If you use it as one place where you are getting visibility of your freight out to the masses and then, once the masses call into you or email you or you talk to them, however, you get in touch with them once you, if you then vet them appropriately and develop the right relationships with them and set the right expectations with them, you can have a very successful partnership that allows you to service your customers. But I just think it's interesting like the company that's invested hundreds of millions of dollars into creating a carrier marketplace so they could not have to use load boards still has to use load boards, which segues, I think, nicely into the idea of Fleetworks, right.
Speaker 2:I agree. Yeah, I think ultimately, freight is so interesting because so far there is no single market clearing marketplace. Dat is probably the closest that we have, but I don't think it is. It's obviously not the single marketplace, right? There's a lot of freight not booked on DAT. It is a good segue into us and why we should exist. So if folks haven't heard about us, so Fleetworks yeah, go ahead, go ahead and explain what I was going to ask and I just didn't.
Speaker 1:I wanted to take a drink, but it's all good, tell us what is Fleetworks, yeah.
Speaker 2:So we're a lot of things. Fleetworks uses AI to connect carriers and brokers. What that means is when a carrier calls you off of DAT, fleetworks can answer that call. We can vet the carrier through our partnerships, whether that be Highway RMIS, my Carrier Packet or my Carrier Portal. We can describe the load to the carrier, including all the way down to the special instructions and the expectation setting of the load, and then we can negotiate on price and, depending on what our customer wants, we can actually go all the way to booking the carrier on the load. That's our core product today. That's where all of our customers are finding no joke 10X ROI just from a operating cost standpoint. So there's 10X ROI with Fleetworks just on that product.
Speaker 1:Because we've built this Can you clarify something? Yeah, sorry, you mentioned at the start of that answer, for what Fleetworks is you mentioned when a carrier calls you in off DAT. I'm pretty sure, and just want to confirm, you are not connected to DAT, in that it has to be a call in off DAT, correct?
Speaker 1:It can just be connected to DAT, in that it has to be a call-in off DAT. It can just be connected to your company call line. Anyone who calls into the company could be answered by a Fleetworks AI agent, correct? That's exactly right, yeah.
Speaker 2:So when a carrier calls you, whether it be DAT, another load board or even just your general line, we can answer that call for you. Customers also use us to make outbound calls to carriers, whether it be to offer available loads, like, hey, I've got this carrier in my routing guide, let's go offer it to those carriers first, before we put it out publicly on the load boards. How do we encourage more relationship building? And customers, of course, are going to be using us for track and trace. Carriers, for various reasons, don't use GPS services or might turn them off, or there's some issues and you want to call the carrier and you want to know is this load on time? Have you arrived? Are you being loaded? What's going on? We take all that information and we automatically update our like.
Speaker 1:Your TMS is operating autonomously through the use of Fleetworks voice agents, our email agents and our other sort of communication agents like text and WhatsApp. What has been the biggest challenge in building the product itself?
Speaker 2:Yeah, On the product itself.
Speaker 2:Yeah, on the technology side, I think one thing that we've done well which has been hard has been being very diligent with our customers about offering them the customization whether self-serve or whether we do it for them the customization to build AI agents in a way that really fits their business.
Speaker 2:I think every brokerage has their own bit of secret sauce and their own way of doing things, and so it was really hard upfront to build a product that could support different types of negotiation, different phrasing, different ways of doing business. But I think we've gotten that and I think we're continuing to invest there. But that's a really big thing for us is not forcing brokers to fit to our process, but rather bringing a technology that lets brokers adapt it to how they want to do business, which is kind of a new way of technology. I think that's the other thing that AI sort of unlocks is you don't have to build your team to fit what that tech provider tells you to do. We share best practices with our customers, but we let them kind of customize things to how they need to run the business. And is that?
Speaker 1:customization work done by your team of seven technical people, or is their team that can go into the system and change things however they see fit?
Speaker 2:It's both. It's both yeah. It really depends, and also, it depends on their level of sophistication. Some of our customers are really sophisticated. Their developers are directly connected to our developers on Slack and we're talking every single day. Sometimes we offer that as a service to our customers too.
Speaker 1:And why wouldn't a sophisticated brokerage or a brokerage with a sophisticated technology team? Why wouldn't they just build this themselves?
Speaker 2:They could certainly try, they could try, could certainly try, they could try. What they will find is that it's really easy to stand up a proof of concept. They could build a voice AI agent in. I don't know, maybe if they were really dedicated on it let's say two months, I think what they'll find is that the amount of UI that operators need to understand what the AI is doing, but then also actually all of the detailed nuances of how all these AI agents work together to actually complete a call it's actually really, really hard to pull off, really, really hard to pull off. When you think about negotiation, when you think about sharing details of the load, when you communicate appointment time flexibility, all that infrastructure we've built. So could a CH Robinson build this? Absolutely, it would probably take them a couple of years to catch up to where we are, and at that point I'm like, why not just get ROI today?
Speaker 1:Yeah, well, help me understand what you mean by the AI.
Speaker 2:agents work together yeah, it's a great question. This is like getting really nuanced into how we run our product without giving out too much of our secrets.
Speaker 1:Sorry, I don't want you to give up your secrets, but I just I want to understand kind of what you mean, because this is more me as a curious, non-technical person and I'm sure my audience is curious about this stuff, certainly not trying to give you a trade secrets, but keep that in mind as you answer.
Speaker 2:I'll share what I can, so. So, basically, in order to do the things that we need done well, like whether it be a track and trace call, making decisions around what types of questions to ask in that call, making decisions around how to route calls, depending on the type of negotiation, the type of load it's not just one AI, it actually could be dozens of AI, which, from a carrier perspective, it doesn't sound like that Carrier perspective, it's just they're talking to one AI the whole time. On our end, we're actually handing off between lots of different agents who are evaluating the outcomes of prior agents and making decisions on how to move forward. So that's not really like totally secret sauce. Our value is how we stitch those agents together. Um, our value is then how do we?
Speaker 2:Also is actually really important for operators at brokerages to really understand what the ai is doing on their behalf and where they need to lean in as a, as an operator. Right, because the ai can't do everything for you, but it can partner with you on negotiation, on vetting, and you can provide input to it. You can say, for example, a great, great example of how our customers partner with the AI live. So we're negotiating with the carrier. Let's say, a customer marks a load as a hot load. Right, you're like I'm willing to take things above my max rate because this load's really important to me, or this shipper's really important to me.
Speaker 1:We'll send a Slack message or a Teams message to our customer, to the rep on file, on that load, and we'll say hey, Joey, when you say we you're talking about an AI agent is sending a Slack message, Yep, okay, keep going.
Speaker 2:If it was we man, I we'd have to hire a lot more people to to send slack messages just trying to understand.
Speaker 1:Yeah, just trying to understand um, that's good.
Speaker 2:I was like, man, do we need to change our hiring plans? Um, no, so yeah. So rai will send a slack message to the operator, say hey, uh, you know. Hey, joey, we've got this carrier on the line. Here's the rate, here's why we're sending you a message. The load is hot. Joey can then respond and say take that rate. Or hey, try to negotiate them down. Or hey, ask them when and where they're empty. We can collect all that information. So that's kind of how we let individual operators at our companies kind of scale their impact a little bit more.
Speaker 1:Got it? And when you're out selling to new potential prospects or customers, what are you selling as the differentiator for why Fleetworks is the answer to them?
Speaker 2:Yeah, so I think there's maybe two sales. There's like, why should you just use Fleetworks in general? And then maybe why should you use us over any other AI agent company? Right, so why Fleetworks in general? I think, first of all, the carrier experience when talking to a brokerage can be amazing, but it can also without AI. It can be amazing, but it can also without without AI. It can be amazing, but it can also be really bad. I've, I've called some of our customers, their general carrier line, saying hey, press one for it to talk to the care sales team, and I've been on hold for like three minutes, right, me as a carrier. Like that's not a good experience, um, so why? Why should you?
Speaker 1:why should you use Fleetworks? Well, number one, one that's a wildly that's a wildly low number to use as an example, because there's definitely examples where it's 30 minutes and no one ever picks up the phone but keep going.
Speaker 2:You're right, maybe I'm being too generous. I appreciate the modesty in that answer.
Speaker 1:I mean I've talked on the show about abandonment rate and how brokers have in some cases an abandonment rate of up to 20%.
Speaker 2:That's super low In those cases 20% is good.
Speaker 1:Well, 20% is good. 20%, I mean, if we're talking about abandonment rate as the percentage of phone calls where a carrier hangs up because no one ever answers, I would think three minutes is pretty good relative to 20, 30, 40% of the calls just never being answered.
Speaker 2:Yes, we've seen as high as 75% calls being missed. I'd say average wait time for a carrier, based on our data, is two minutes of wait time before the carrier just hangs up and just calls the next broker down the line.
Speaker 2:Got it Got it. So we've certainly seen crazy wait times. Why should you use Fleetworks? Number one we solve the wait time problem for you. Number two we actually solve a massive data problem for you. When you're calling out to carriers or when you're taking inbound calls from carriers, your reps are too busy to take that call and in the same breath log into their TMS and log that offer or lack of offer in the system. Right, like, everyone wants the reps to do that, but the reps just like don't have time. The reps are trying to do a lot with their day and so you're losing huge amounts of data.
Speaker 2:So we help our customers increase data captured per lane anywhere from 2x to 10x on some of your hottest lanes. So we help our customers increase data captured per lane anywhere from 2x to 10x on some of your hottest lanes. We have customers who in a given week, we might give them 100 offers on a lane and that's actually a signal for them. We say, hey, go to a customer and get more freight there. And also, you might want to think about what your carrier pay actually looks like on that lane, because there's clearly a lot of demand that you're just like not tapping into. So that's so. We kind of help the carrier experience, we help the data piece, but the other thing is like we also just help our customers become infinite in a way that like was never really possible. So what could you do as a brokerage if you could call every carrier and understand their preferences? What could you do as a brokerage if you were not rate limited by track and trace calls? What could you do as a brokerage if documents were automatically parsed? What could you do if every sop that you dreamed of was executed on time and the data was inputted back into the tms? Um, and I think it's really that level of possibility.
Speaker 2:That has been like that's been the most rewarding part for me is like I, when I was in chicago two weeks ago, I was mostly talking with with a lot of our current customers and I think the common theme was like Fleetworks does carrier sales, fleetworks does track and trace for us. How else can we deploy the technologies? I think we're in just I feel like a very rare and, frankly, a very humbling position where you know this, when your customers are coming to you and asking how else can I use your product, like that's a really, I think, great place to be because the team works hard, right, and it's gratifying that when we work hard, customers come to us and say you've earned our trust and let's go deeper. But it's also a great opportunity for the business, right, because as we do more for our customers and the relationship really up levels. But it's also a great opportunity for the business Because as we do more for our customers, the relationship really up-levels.
Speaker 2:The ROI from Fleaworks is so obvious that I don't even have to necessarily go out and do another business case. For me it's just a no-brainer of. I want to use Fleaworks as much as I can.
Speaker 1:Basically, what does the customer base look like today? How many brokerages are you in with? How? How big is it?
Speaker 2:yeah, um, we're we're millions of dollars in revenue. Um, we have a few dozen brokerages, some big logos that I cannot share right now. Um, but I would say that basically every brokerage of meaningful scale is either using us or considering using us, and so it's been awesome, it's been really fun.
Speaker 1:That's impressive. I mean millions is a big number for the time that you all have been in business and certainly must feel validating.
Speaker 2:For sure.
Speaker 1:Where do you see the product is fully baked and where do you see there needs to be? There's more room to develop.
Speaker 2:Yeah, the product is not fully baked anywhere, which is a weird answer because we're we're selling this product. I think the product works really well, but I there is it's. It's an insane problem to have, as like as a former product person, um, the prioritization that that we have to do is is really interesting and exciting because, like I just think there's improvements everywhere. For example, outbound calling, how to get just more accurate on outbound calling, or looping in the operators, or capturing appointment time flexibility. I just think what a broker can do or what a brokerage does is so complex that I just still feel like we're we're in.
Speaker 2:I feel like we're in the second inning. You know, I think the first inning was does this thing work? I think the second inning was can you get more nuance? I think we're in the bottom of the second. We're maybe getting to the third inning where I just think there's vectors of improvement, um, on every possible channel. Um, and I'm just really excited particularly about helping our customers get smarter about their networks. Like that's really I think the best brokerages are great at network management and we want to help all of our customers get there.
Speaker 1:So I definitely see. I mean the idea that your commentary around help you be infinite. I don't know why I started like envisioning Tony Stark and like Marvel as you've said that.
Speaker 1:Yeah, or Thanos, I don't know, but I see what you're talking about and I see how much data you can collect using this and how much more reach you can have using tools like this. And I guess what I'm curious is I'm curious of a number of things. Let's start with the competitive landscape. So you know, I've asked pretty much every AI company on here a question of this nature, so I'm sure you're not surprised to hear it. But how do you think about the competitive landscape? You've got at least one major competitor in Happy Robot that just raised, I think, 15 million bucks or something like that and an $85 million valuation, which is huge. Congrats to them.
Speaker 1:You've got I just interviewed Dave Bell has clone ops. I think that might've been what inspired you to message me like yo, can I get on the show? Like, and you know, in that example that's a brand new company hasn't launched their product. I think they're launching their product soon and it's, you know, a guy with like 30 years of industry experience and a lot of success, and then I'm sure there are more coming that I'm not aware of. So how do you think about the competitive landscape? And like, how does, how does the competitive landscape impact your decision-making for how you approach the business and how you approach growth, fundraising, hiring and such You're absolutely right.
Speaker 2:I think any opportunity that is worth chasing is going to bring, I think, great competition. I wish I could share more on our fundraising, but we're not public about that just yet, but more to come soon. The yeah, I think overall intense focus on the carrier relationship and how that broker can build, nurture and grow that relationship with carriers has been critical for our business. So I mentioned this earlier. But everything we do and everything we did at Uber Freight and everything we do here at Fleetworks is really carrier first and carrier centric. That's kind of why customers tend to choose us is because we're not just a voice AI company. We don't believe in spamming carriers. We believe in helping our customers actually grow meaningful relationships and grow meaningful networks. So that's really our main goal.
Speaker 2:There's a bunch of AI companies I think there's only, to your point, I think, two that actually have real working products publicly, and I think our speed of execution just continues to be just absolutely killer. I mentioned this earlier. We have seven technical folks. They're all former founders, which means, yes, we have seven people, but technically those are actually like 20 or 30 people worth of work, and so our ability to attract great talent continues to be like my number one priority as a CEO. The reason we can attract that talent is because we started from a really high base and then we can pay well above market because we only hire top 1% talent. So instead of hiring two people who are good, we're just going to pay a top person like 50 to 75% more, and so our salary ranges are going to be competing with, like the big tech companies. Because we can afford to do that.
Speaker 2:Folks are really attracted to us because we're already profitable. You know, in a really short amount of time we've built up enough customer trust and I think ultimately, the thing that for me that really sticks out is again, all of our customers come to us and say how else can I use Fleetworks? It's that proactive outreach that every single day and for sure the competition is intense but every single day the fact that our current customers are saying how can I use Fleetworks more? Why is this working the way it does? And basically just saying, hey, we want to spend more with you, we want to grow with you. I feel really good about that direction.
Speaker 1:But it's early days, yeah, I mean I. Yeah, good about that direction, but it's early days, yeah, I mean I. Yeah. Yeah, no, I mean I commend any any. I'm a big believer in just following the customer and if, if the customer is giving you feedback that your team's doing a great job and that they want to see more from you, then you figure out how to give them more, and it's it's part of why I'm curious.
Speaker 1:I have two questions here, but one is around. It's interesting that you talked about the hiring process and how you're able to pay more than the average company would be, how you know there's certainty that more is coming. Is there not a temptation to go raise a lot of capital and turn seven really strong technical people into 25 really strong technical people and then you can go faster and bigger and more expansive and get your arms around the industry faster? So it's harder. You have maybe a deeper moat at that point, like how do you think about the idea of you's harder? And you have a maybe a deeper moat at that point, like how do you think about the idea of you know what you're doing versus something more aggressive?
Speaker 2:yeah, um more to share on fundraising soon okay that's the whole answer. That's the whole answer, okay, um.
Speaker 1:Okay, all right, then let me ask you a different question then. How do you contemplate navigating pricing in an environment where more competition is coming? But whatever dollars per call or per minute you're charging, or cents per minute or cents per call, someone may come in and offer a cheaper number. How do you navigate that kind of idea as you build out a product like this?
Speaker 2:That is a great question. We spend a lot of time thinking about that as a company. So, first of all, we've already heard that some folks are trying to give away this product for free I'm not exactly sure who. To me, that doesn't really make sense. I think the folks that we tend to partner with they understand that they're putting the trust of their business in our hands right Like we are their front of house and that inherently has value to it.
Speaker 2:And our customers, like the way that we think about our revenue, is our customers investing in us, like they're giving us the resources to go out and hire more people and support their business and build, like really deep products that support the business. Um, as we kind of think about moats, um, there are a few things, um and I'm thinking about what I could, what I want to share publicly versus not but, um, overall I I think there's a bunch of moats in this business. No one has switched away from us and switched to our competitors. Now it's possible, come year-end renewal time, customers might evaluate what we do with our customers is. Our product is so deeply intertwined with how they do business and how they cover their freight and how they think about managing their freight, that to pull us out, um, it's technically possible, like, like, people switch their TMSs. Um, sometimes I don't think anyone has ever pulled out their TMS and said, wow, that was like a great experience and incredibly easy and not painful, like pulling Fleetworks out is painful because, like, we're doing the job of some of your people and so you know, getting rid of us would be like pulling out, you know, dozens of carrier reps, like, or, if you're a large brokerage, hundreds of carrier reps. And then saying, okay, I'm going to bring on a new provider or I'm going to go hire people to fill that gap.
Speaker 2:And as we find that our customers build processes around the data and the execution that we help them with, we find that customers tend to just be really happy. And I think my promise always to my customers is hey, we're always going to price fairly. We're going to price fairly. If we actually feel like there's an opportunity to lower prices for you, we may do that. And ultimately the ROI that we demonstrate for customers is kind of a no-brainer. So if someone comes in for one penny cheaper a minute or 10 cents per load cheaper, our competition cannot provide a 10x better product. So that's kind of how we think about it. Yeah, that's fair.
Speaker 1:I'm curious if the product is able to do the job of dozens or in some cases it's scaled brokerages hundreds of carrier reps. What does the shift look like within the company when they deploy Fleetworks? If I'm a carrier rep, am I worried for my job when Fleetworks gets installed, or is my job just fundamentally changing into something completely different? Help me understand what that looks like.
Speaker 2:Yeah, it's a good question and I think it's really good timing given where the market is. So, no, secret market's terrible. Like everyone, I think every brokerage had to make or a lot of brokerages had to make pretty deep cuts. A year or two ago Feels like consensus is we kind of hit the rock bottom. Things are coming back up. I think this is a great time to invest in a platform like ours, because you bring us on and suddenly your hiring plans change.
Speaker 2:It's not really about firing people. It tends to be. I mean, we have seen people change workforces, but I think those are companies who just felt like they were really overstaffed anyway. For most of our customers, it's more like, hey, we like our people and we've trained them up really well and we don't want to go through the pain of training from scratch more people. So let's use our existing people. Let's shift them to actually doing more meaningful work. Instead of my carrier reps spending 25% of their day doing track and trace, 25% handling call-ins off of postings, 25% of their day shooting the shit and 25% of their day doing outbound calls and outbound emails and building relationships, let's go more 25% of the day shooting the shit, 75% of the day building relationships right, and so that's the calculus that you know a lot of our customers are making.
Speaker 1:And how does, like you mentioned that you want to be very carrier-centric and help develop or evolve the relationship between the broker and the carrier. It just naturally feels like a product like this detracts from the relationship Right from the relationship right. It theoretically removes a lot of the human conversation that has to happen in a load booking process. So help me understand how the relationship can actually be improved through a product like this.
Speaker 2:I love this question because it's so foundational to what we do. I think when we started, like I said, there were doubts like are carriers even okay with this process right? And I think when we started, like I said, there were doubts like, are carriers even okay with this process right? And I think, fortunately, we proved that they are right and if we didn't, we wouldn't even have a business. When I built relationships with carriers, it's a two-part, three-part process. I think number one is the data that you collect about the carrier where they like to run, what their equipment type is. Number two is the execution on that data. And then number three and what I mean by execution on the data is like you, you you actually tender the load to the carrier when you have to give them business.
Speaker 1:Yeah, you have to give them freight. You understand. You understand their needs and wants. Yeah, that's step one. Yeah, step two of the relationship is actually like executing an order together, leveraging the information they've given you.
Speaker 2:I'm with you, keep going and part three is what's your dog's name, what's your wife's name, what's your kid's name, what's your favorite sports team? Like? I think ai can do number three. My general approach is it, it, I think it's, I think it's funny, but I don't really want to invest in it. Um, my approach is ai can definitely do number two really well, so we can definitely do number two and we can definitely do number two really well. So we can definitely do number two and we can definitely do number one. So we can help you collect data, or we can be a repository for your data and we can consistently be that execution layer that helps you execute on the data that you collect.
Speaker 2:And so, andrew, if you're going back to the brokerage floor and you're working at Ally Logistics or Capsule Logistics or any of our customers, I think your role changes. I think your job is to develop carriers and put that data into Fleetworks to help you make more money, more money. We're going to help you make more money as a carrier rep because we're going to help you book a lot more consistently with owned carriers. And then what you tell the carrier is hey, man, I'm going to have an AI reach out to you, but you're in the routing guide. You're number one in the routing guide, so you're always going to get this load before we put it out on the open market, and carriers appreciate that.
Speaker 2:Me, as a carrier rep, I would sometimes forget to tender that load to the carrier if it was like an informal agreement, right, because I had just a lot of things going on, and so we consistently help our carriers do that. But if we free up the carrier reps time to say hey, joey, like if you ever need me, I will always be here and my time is more available now because I'm using AI to help me book freight. So I'm using AI to help me book freight, so I'm always there for you, but AI is going to be my assistant and you can always tell the AI connect me to Andrew, and we always will.
Speaker 1:So let me just say, hearing that kind of explanation the three parts I am 100% sold that, if the technology can do as it's intended to, this is a home run solution to evolve a carrier rep's time and by evolve I mean make it more efficient and allow you to do more. I've spent years as a carrier rep and I've had some really strong relationships with owner-operators and I still, to this day, can call a number of them and I have their numbers memorized and they would do what I asked of them if I needed them and I would do what they asked of me and, and that was a product of just relationships built over time. But that was like a grit relationship. That was me calling them all the time, understanding their lanes, memorizing it, a million note cards and I had as good a tms as anything at bazooka, at coyote, and and and I could have logged it, but my process it was just easier to do the notes, as I think about this one the ability to get seemingly infinite information from carriers on where their trucks are going to be, what lanes they just got from new customers that they need backhauls for, whatever. I mean. There's clear value in being able to call more carriers and get more information, housing that information as a repository and then using it to make more decisions. I mean that's another big win. And the ability to again infinitize it. That's not a word, but like to make it infinite, to make it so you'd never miss the calls. There is so much data that is missed from brokers because, like you said, the abandonment rate can be up to 70%. I've seen 20, 25, 30, whatever it is. It's a big number. It's a meaningful number that if you can replace a missed call with an answered call, information provided and taken in and then saved and used to make better decisions, that's another home run.
Speaker 1:The third piece is thinking about the actual carrier rep. And that's where I'm putting myself back in the shoes that I sat in 15 years ago and recognizing that there's so much time I spent gathering the information. And if I just had the information in front of me and I could spend my whole day calling the right people who I already knew what their information was, and my job was solely to buoy them up or become their friend whether it was a man or woman, it didn't matter or become their friend whether it was a man or woman, it didn't matter and you know there's so much value to that and I see how there are certain things Like, for example, you know I remember one of my carriers, peter Achukwu, and his wife Catherine. They used to do a lot of local Southeast Coca-Cola stuff for me. They were based in Marietta, georgia, and they would have to go to the Coca-Cola Atlanta facility and I've talked about this facility, for he hated the woman who ran the facility. He said she was an asshole and was a real pain in the butt and we used to always put them on loads that the appointment had already been missed.
Speaker 1:Like AI can get me to the point of matching him to the load and making me, letting me know that I should call him and convince him to take the load, because that's what it was at that point was like I knew he didn't like the load. I had to lean on our friendship to make that happen and I don't want ai talking to him about like, oh, the lady at the facility, like he and I can have that that's the personal element to the conversation is around the nuance of the load. That's not exactly. I mean you could call it a data point, but it's.
Speaker 1:It's. You know, it's something that feels a little bit different. That AI wouldn't be involved in is like the idea that there's a person there that's a pain in the ass or that you're going to be waiting a long time. That's where I want to talk to a person who has compassion, so I can have compassion for the fact that Peter would have to wait six hours to get loaded and our conversation is going to be a lot more productive if I'm understanding what he needs for that versus an AI. I think I just don't think that's where that plays. So, yeah, this is my I guess my stamp of approval.
Speaker 2:I don't know.
Speaker 1:I was trying to think of the word yeah, whatever of the idea, because my brother and I at one point were arguing about this on Twitter or something, where he made the age old argument of anytime I call an airline or the bank and you know, a robot picks up. I'm yelling agent agent wanting to talk to a person, and I've been there, I've been that person. I've frustratingly said agent over and over again until they put an actual person on the phone. But I think there's proof in the pudding. I mean, how many phone calls on a daily basis? Or just give me a number of phone calls that have successfully been executed between Fleetworks and carriers, even if it's not exact. Give me a number of phone calls that have successfully been executed between Fleetworks and carriers, even if it's not exact, it's a number that represents that, I believe, the market is open to this.
Speaker 2:I mean millions.
Speaker 1:Exactly, it's millions. So if millions of calls have taken place where a driver was okay with the voice AI on the other end, that's a large enough sample size for us to say that carriers in general are going to be open to this and therefore, with the type of return it has, it's coming. It's coming to every brokerage near you. So I guess I'll ask a question now, given I just kind of went on a little rant there what do you see as the long-term impact? So let's say I'm Joe's Trucking, I've got 10 trucks, I'm based in Nampa, idaho, and every September my phone starts ringing off the hook from anybody who's taking onion or potato loads, potato loads out of the Northwest and they're seeing if I'll come help them with my 10 reefers. But once every broker has either Fleetworks or your competitors and they all have set up this outbound data collection process, am I not getting a phone call 400 times a day by people wanting to talk to me?
Speaker 2:It's a good question. I think there is danger of that 100% Couple thoughts there, I do think. Well, first of all, actually, I want to go back to something he said before. It sucks when I call because I actually like doing things over the phone. I like calling support on the phone, and it's so frustrating as someone who has built a pretty successful voice AI company that I still run into old school like press one. If you want to talk to bookings, I'm just like God. Can you guys just use this technology so that I can get my job done? It sucks because I was like another startup needs to come in and sell to Comcast and United and all those guys. So that's my rant. But to get to your question, I think your phone could be ringing off the hook.
Speaker 2:This is our job at Fleaworks and why I'm really emphasizing being carrier. First, it's our job to understand number one how does that carrier like to be contacted? So they might not want a phone call, they might want an email, they might want a text message. So we want to be wherever that carrier wants to be. Beyond that, I do think we might want to be wherever that carrier wants to be. Um, beyond that, I do think, um, we do. We might move to a world of carriers being more platform, more like I don't even know what the term is like platformized, um, like, I think, like convoy has obviously built like their platform, uber freight's got their platform. Um, I don't want to open that can of worms. I think you had the LinkedIn post of the century last year when Uber came out with that. But there's all these carrier platforms and I think because we help our customers book so much, freight carriers may come to us and we may help them book freight easier without a phone call.
Speaker 1:So you're saying there's a world where your own business drives away the phone calls that are happening in the business. I mean, I don't mean to say it like that, but there's a chance that you have so many calls that you're creating for customers that you now need to create a new solution to help carriers get away from the calls. Quite possible, yeah. Or you have to create the same agent to take the calls on behalf of the carriers.
Speaker 2:Totally yeah. The problem is that carriers are just, they don't have like small carriers don't really have a platform today, and that's why the phone calls are so powerful is because that's the only way to reach the carrier. So I think more to come on that. I also think that, frankly, the government does need to catch up to this in a way. The laws that govern AI phone calls were written in the 90s when robocalling became a problem. I think those laws are pretty antiquated, obviously, because those calls were literally spam calls, but you could have AI spam. But I don't think we're spamming, I think we're trying to do business and I think the government needs to create either industry frameworks or universal frameworks to regulate this, because otherwise I think you could certainly have bad actors who are not carrier first, um, and who who are kind of unscrupulous and who are going to just be uh you know, literally downloading the fmcsa database and just clicking call on 300 000 active carriers, right yeah, I mean I like why wouldn't someone do that?
Speaker 1:actually, you could create an interesting business if you just created what you've created with your voice AI agent and then you just go collect all the data from the carriers. I mean it'd be messy, but that would be interestingly valuable data and I'll use that as a segue to a question. Is there any avenue for you to take the data that you're collecting and leverage?
Speaker 2:it in a way that is so.
Speaker 1:I interviewed. Wait, let me make sure this is public. Yeah, it is. It is Part of the green screens strategy is kind of helping all the small brokers that they partner with aggregate their data so that they get a. They have a small piece of a big pie, but the pie gets bigger every time there's a new provider who signs up and therefore the data gets better because they're all putting data in and they're all being able to leverage that data to get better pricing.
Speaker 1:Do you get what I'm saying? Is there an avenue for if I'm Ally Logistics and I have call it a thousand calls a day, that Fleetworks is doing for me, I'm only as powerful as the data that comes from a thousand calls. But if I'm Ally and I'm part of a network of 30 Fleetworks customers, that a thousand calls becomes 100,000 calls a day. Is there an avenue for all of that data to be leveraged by all the customers?
Speaker 2:I think there are opportunities. Yeah, how do you think about it? Because I think data sharing in this industry is both common practice but also maybe a little taboo. So I'm curious how do you think about that, and what data should be shared versus should not be shared.
Speaker 1:Yeah, I think it's an interesting question, which is kind of why I asked it, but I definitely think that there are two trains of thought, or maybe there's more than two trains of thought. I think that a lot of brokers think their carrier capacity is their differentiator and the data on their carrier sourcing is a differentiator. And I get why your network should be a differentiator. There's not a ton of other options for differentiation. There are a few, but not a ton.
Speaker 1:How you run your business, the rules you use to govern, how you execute for customers things like that, how you engage your employees but that's not the point of the conversation. So I could see why a broker would say absolutely not, don't share my data with anyone else. That's not the point of the conversation. So I could see why a broker would say absolutely not, don't share my data with anyone else. That's mine and I want to use it for myself. But alternatively, especially on the smaller broker side, you want to leverage whatever piece of information you can get and you can't compete with the CH Robinsons on the data front. So I could see why they'd be more open to it.
Speaker 2:Yeah, I think it's on us to come up with the right rules of data sharing. Certainly, we've got some mega customers and they've expressed concerns I think very valid, about hey, I don't want my data being shared with a brokerage who does 50 loads a day right, because they're getting more value from the data than we're getting from their data. On the flip side, I think it's a good point. On the flip side, uber Freight, to your point, has spent hundreds of millions and, if we count TransPlace in the equation, billions of dollars building this business. They still pay DAT a lot of money for access to their data, a lot of money. And so, clearly, even Uber Freight, who, despite all their investment, has gotten to 1.5% of the brokerage market roughly, despite all their investment, still not a market mover Could they benefit from 40 other brokerages lumped together under one tool?
Speaker 2:I think, maybe. I think maybe We'll have to prove it with them that when they access our data, they can price more accurately and better. That's kind of how I think about it. I certainly think there's a lot of openness from I would call the more small, mid-size, like the $100 to $300 million brokerage. They understand that, hey, our secret sauce is kind of like our sales team and how we service our customers. And, at the end of the day, our carrier network is tiny compared to the overall market, so if they can use Fleetworks to tap into a broader data pool, I see that as like a win-win.
Speaker 1:Is that an offering? Today, though? Are they able to tap into a larger pool? Not yet.
Speaker 2:Not yet Okay.
Speaker 1:Yeah, still looking at it. Okay, and when you think about the product today, everything is centered on voice, right? I mean, are you thinking about doing some of the other stuff to compete with, like the you know, just to speak to my former guests like the Vumas and the drum kits of the world?
Speaker 2:outside of voice. We're already executing on our vision of meeting carriers where they are. Email is going to be a big channel, text, whatsapp those are all really common channels for carrier communication. Voice is obviously just like a slam dunk and it's going really well and that's still why a lot of customers come to us. But all those products are alive and well.
Speaker 1:But in terms of that's a good point. But okay, so you mentioned all carrier-facing, though Is some of the stuff that the Vuma and Drumkit guys are doing are focused more on the customer side of the house? Are you dabbling in the customer side of the house or everything is still carrier-centric and planning to stay carrier centric? Or tbd, maybe fundraising, who knows?
Speaker 2:sure I love fundraising, as you know, as my as my mysterious answer look, I mean I think I think here's the thing that that I will say about it is like first of all, we see our core mission as connecting carriers and brokers. If talking to customers or engaging with customers enables better outcomes for coverage and carriers and more economic opportunity for carriers, we'll absolutely go that direction. Given how nimble and sort of well-executed our team is so far, I could see us going into that path. I guess what we're calling more agentic workflows and building operational processes or rebuilding processes using AI.
Speaker 2:Absolutely, like I said, I think we're in the bottom of the second in terms of how do brokers connect with carriers, and so I would say, if a customer wants us to do email quoting automation for them in February 2025, I would say you should 100% go with Vuma, but if you want the best product that enables you to cover your freight cheaper than a human, both from an efficiency standpoint and a gross margin standpoint, you should go with us. That's kind of a no-brainer Is it an or or? It could be an and it can for sure be an, and we have many customers that use Vuma for some of their core products and they use us for our core products.
Speaker 1:Okay, got it. What's been the most challenging element to selling technology into freight brokers?
Speaker 2:probably the number one right now is just how tough the market is. I think you have a lot of brokers who are on tech pause, as they're like we need to see when this market comes back and, to your point, it's not a question of if, but rather a question of when. So I think it's a timing thing, but look, I'm still really happy about the team's process and progress, despite the fact we're in the worst freight recession. Coming out of the worst freight recession in recent memory. That's a tough part. I think we're also very fortunate that we're coming in.
Speaker 2:If we were selling this product 10 years ago, I think a lot of the core technologies, like particularly the TMSs, were not originally built to support integrations of this magnitude and of this speed. So we're very fortunate that the TMSs all have the capability now to support a product like ours, and so I'll actually, as a piggyback on that, I think people have been really burned by bad technology in this industry. We come to people and they're like how do I know what you're saying is true? How do I know I'm going to get the ROI? I'm like, look, man, just try it, we could integrate.
Speaker 2:I love the story of LGI and, if you know Brandon Bay there he is a true homie. Brandon was very public that we from signing to when we handled their first call from them, so we did the whole integration and operational process and launch in 22 hours, and so our goal is to really elevate the bar on what a tech vendor can do in this space. So that's the hard part is actually we're selling against, I think, the missteps of a lot of companies who came before us. But I think that just comes down to having a great team building customer trust and continuing to just rebuild customer trust over and over.
Speaker 1:Yeah, great answer. All right last two questions for you, and they're kind of go hand in hand, both sides of the same coin. What would you say is the number one thing a happy customer is saying about Fleetworks? What are they raving about? What do they love?
Speaker 2:Number one they're like God damn, this thing works. That's the first thing. Number two I're like goddamn, this thing works. That's the first thing. Number two, I think, is going to be just communication and support. We set up Slack channels or team channels with all of our customers. We're just continuing to win on how we communicate with our customers, how quickly we can prioritize their work and how we can build around their process so the product works. The support is great. You have access to the whole team whenever you join with us. And then, number three, I think, the speed of new product ads.
Speaker 2:I think a lot of people going back to prior vendors a vendor does one thing for you, right, and their product doesn't evolve. To be honest, I'm still catching up on trying to. I'm communicating to our customers all the new things that we're bringing to the table for them, and so I think a lot of customers are like God damn, whether we onboarded with you three months ago or we had our first conversation three months ago. I need to tell them hey, I need to run a new demo with you, because what I'm going to show you now is different than when you first talked to me back in September, october. So, yeah, it's going to be the product works, the customer support and just the speed of new build.
Speaker 1:Smart way to ask for one thing give me three, because it gives you a chance to brag about the business a little bit. Yeah, because now you got to answer the hard question, which is what's the number one thing that customers are complaining about or frustrated about?
Speaker 2:What is the number one thing?
Speaker 1:And you don't have to give me three here, you can just pick one.
Speaker 2:I mean, thank you for limiting it. I could probably give you a few, but I think we could do a better job at exposing to customers the rules that they have built in our system or that we're building in our system for them. So they might ask, hey, why is the AI doing this? The team jumps in and we're like, oh, yeah, it's because you asked us to do that two months ago. And they're like, oh, that makes sense. Or they're like, oh, let's change that. We don't want that rule anymore. So I think we need to continue to expose more of that whether it be customization or whether it be rulemaking to our customers. But it's great that we're now, frankly, at the point where I would say, 95% of customer outreach is the product is working as expected. We just didn't do a great job at exposing that and continuously communicating it to you.
Speaker 1:It's a funny answer, but I actually understand it wholeheartedly as someone who's run a brokerage and understands how quickly we change our minds about things and how fast-paced things are and how easy it is to forget. You made that rule a month ago and now you want to change it because some new element has shown up that's costing you money or time or energy, whatever it may be. So I actually understand that.
Speaker 2:Yeah, we go super deep with these guys right.
Speaker 1:So their process is our process Awesome. Well, I'll give you one more. You've been a founder now, or CEO, of this business for a year and a half, which can be a long time in freight CEO world. It certainly can feel like a long time For other budding entrepreneurs wanting to get into the freight tech space.
Speaker 2:What advice do you have? Getting into freight tech? Number one advice come work with us. I don't think you're going to work with smarter, kinder, more hardworking people than the people at Fleetworks. If you're not going to work with us, then you should go work for another freight tech company, one that's really high growth, one that's small.
Speaker 2:I think there's so many people that come into the business and they think they understand it and then they build a product. They maybe get to a few hundred K or a million ARR and they just dead hit a wall because they don't understand how complex this space is, how nuanced it is. So I think, if you've never been in the space, joining a company in the space for a year or two to actually really learn it is powerful. I think if you have been in the space, the number one thing that you should do is talk to your prospective customers or talk to the stakeholders that your business is going to affect.
Speaker 2:I think if you take that month to do respectful, tailored outreach on LinkedIn like if you DM someone on LinkedIn, not like a blast, but you say, hey, man, I see, you do this, I'm doing this, I would really love to talk with you for 15 minutes, or can I just shoot you an email with two or three questions? I think something like that can go a long way. That's why we, going back to us, we don't do cold outreach, we do very tailored outreach, we do very personalized approach, because people in the space are blown up by technology and if you're going to build technology in the space, you have to be very relationship oriented and be very respectful. Or join Fleetworks and we'll just put you in front of 20 customers within the first three weeks All right, all right, I'll take that answer.
Speaker 1:Any final thoughts before we call it? I appreciate how honest and transparent you've been about the business. It's a cool story. It's a cool business and um seems like one you've you've got a lot of traction with no, thank you.
Speaker 2:Thank you for having me. I mean, look I um, it's an exciting space. I think it's cool to see people it's it's cool to really like have been part of this journey from the very beginning not just this company, but also how this technology is impacting the space, because we're seeing their narrative really evolve over the last two years. I really appreciate the opportunity and looking forward to seeing you at some show at some point.
Speaker 1:Yep, absolutely. Thank you, man To our listeners. That's all we got. Have a nice week. Thank you, man, cool man, to our listeners. That's all we got. Have a nice week. Thank you everyone you.
