Ep. #76: Pablo Palafox: CEO & Cofounder, HappyRobot

Before we get started today, I wanna give a quick shout out to our headline sponsor, Stoot Technologies. Every week on this show, I talk to operators who are crushing it on the service side, but still chasing down payments. That cash flow gap is brutal when you're trying to scale your business. That's why I wanted to tell you about Stoot. Stoot is the AI platform that does accounts receivable work instead of just assisting with it. Their CEO, Tarek, actually came from TQL, so he gets how logistics companies operate. I've known him for years and I truly trust what he's building. What Stoot does is simple. Their AI actually collects your receivables for you, not some software that just helps you chase the payments. It literally does the work. It's up and running in days, not months. As an example, Bishop Lifting is using Stoot for a 35% reduction in overdue invoices, millions more in working capital, and a reduction of manual work by 50%. If you're tired of your AR eating up time and cash, check out stoot.ai. That's stut.ai. Tell them you heard about it on the Freight Pod. They'll get you collecting in days, not months, so you can focus on actually running your business. Now let's get this episode started. Welcome back, welcome back, welcome back to the Freight Pod. It feels good to say that it's been It's been a minute. I think it's been 3 or 4 months since I've recorded and at least 3 months since I've released my last episode with Chris O'Brien, which was a great one. You should check it out if you didn't see it. But I got the itch again and some friends were bugging me like, you're just sitting there playing pickleball all day. Why don't you just at least put an episode out, couple episodes a month, keep us engaged? And I said, why not? And I got the opportunity to interview someone fascinating who's building a Really interesting, unique business in our space. And this is, I think, a great one to come back with. Mr. Pablo Palafox, did I say your last name right?

Yeah, that's perfect. Appreciate it, man. Very excited to be here.

Yeah, welcome to the show. I, you know, I know you've done some podcasts before and I know, I remember I watched some of the Freight Caviar one way back when you did that. And I think my goal going into this is a couple, I have a few goals here. One, I want to get deep with you. I want some depth. I want to really understand Pablo the person, and I think that's where we're going to start.

Awesome.

And then I certainly want to hear the story of Happy Robot, the origin of it, and how it's kind of developed over time. Certainly some of the accolades or some of the kind of milestones, as you guys have raised some serious cash and had some big moments in your I guess, relatively short history. And then I want to talk about what's really happening in this industry and how companies like yours are changing things, improving things, making an impact, and what the future might look like. So how does that sound in terms of a game plan for us today?

I love it. I love it. Let's get into it.

Awesome.

Very excited.

Okay. Well, let's start with Pablo the human then.

It's a hard one.

Before we get into Happy Robot, so kind of before Happy Robot, like, who were you before all of this? Like, talk to me about kind of the experience that shaped you and how you think and how you operate today.

I think if I have to go back to a couple people, actually, I would probably go back to my mom and my brother and co-founder, Javi. Obviously, my sister and my dad have played also a big part in me growing up, but My mom would basically sign me up for anything and everything, like sports, music. I would just go like 5 hours every day after school to the conservatory to learn how to play piano and stuff. And then my brother would actually, inside more of the sport piece, and actually sign me up for all soccer matches ever just to basically train a mini him. He would basically have a partner to play soccer with or football in Spain. I grew up in in Guadalajara, near Madrid. Guadalajara, there's a small city in Spain, 70K people right next to Madrid. So, we grew up there. I grew up basically multitasking. I'm like a— I'm a doer. I'm probably less of a strategist as my brother Javi is. I'm more of a default to action, no? Like, I just want to do things.

I'm the same way.

That defines my personality a lot. If there's someone proposing something, I'm like, I like that. Let's do it tomorrow. And like, wait, wait, wait, let's think about it. I'm like, no, what is to think about? I like the idea. But yeah, that definitely shaped me up growing up. A lot of multitasking, a lot of doing different things. Very curious, I would say. Very like teacher's pet, you know? Like I would just like be there like, yes, I know that answer. And then in university, you kind of get like the— you're in college, you get your You're facing reality, like holy shit! Like there's a lot of things that I don't know actually. And like I'm now getting like like Fs or like like barely passing exams. Like that's not that's not Pablo. But then like you learn how to like face reality and and face a failure if you will. Just failure. You're failing an exam. That's that's probably not not that big of a failure. But um but it it has shaped me up in that way. You know, like. Thinking that everything should be perfect in my life. No, like, I'm like the best student, like doing sports, blah, blah. Then like you go to college, I then end up doing my PhD and there's like a lot of facing reality again. Like, oh, I wrote this paper, I've put in like hours and hours of work into this research paper. This was when I was doing my PhD in Munich back in 2020 and it got rejected. I'm like, why? Why did they want to reject?

What were you writing about? What were you studying?

I was doing, at the time, deep learning. I was doing 3D computer vision, basically. So, you would have a depth camera looking at a room and you would reconstruct the people in that room moving, the geometry of it. So, it was deep, like we would say, neural networks. At the time, just deep learning. We would not even say AI at all. That's like the sacred word. You don't say AI in 2020, 2019, 2020. I ended up doing a research internship at Meta in 2021. That actually faced or shaped me up also in a way that I realized I didn't want to work in big tech. This was 2021, COVID time. I did an internship remotely. I should have been in Sausalito. Today I'm calling from San Francisco. I should have actually been here in 2021, but— COVID did not let that happen. And at that point I'm like, okay, I'm done. I don't wanna write more research papers. I've done enough. I'm gonna quit. So I drop out of the PhD in 2022 and then I take like a bit of a sabbatical with one of my best buddies, Luis, who's our CTO now. He comes to Munich, we jam for months, we kind of do the, let's think about ideas. So like we get some grants from Germany, We play ping pong every afternoon in a small village in Munich, near Munich in Neufahrn, tiny village next to Munich. Javi, I was at that time like still working as the CFO of this olive oil distributor leading North America, is like giving us stipends to like sustain. Like someone's like, let's try startups. Like you don't even know what a startup is. And we're out of Germany in Europe, which is not like a, It's a decent environment compared to other places, but not like SF in terms of VC money and all that stuff. Maybe I'll stop here. I don't know. I've probably gone too far.

No. Yes. No, you're not going. You're certainly not going too far. I do have questions. Let me ask first. How old are you today? How old is your older brother?

I'm 31. I just turned 31 a few weeks ago, and Javi is 10 years older.

41. Okay.

If you met Javi, you would know it. make that bet. I mean, he takes good care of himself.

Yeah, I mean, especially being a CFO of a North American olive oil company, I assumed he was a little bit older than you. And you certainly have younger vibes to yourself as well. So I'm curious, you seem to have been a very good student, a learner, but it didn't Not much of what you talked about screamed entrepreneur founder. And I'm curious, like, was there a point when you decided, like, I don't want to be a student anymore, I want to be a builder? Or was that always the path? Like, talk to me a little bit about when that mindset became what it was.

I actually have this question for candidates. I ask them, guide me through your light bulb moments in life. And I explain to them the one I'm going to I'm going to explain here. I ended up doing my master's, a double-degree master's in Munich for mechanical engineering. Although I had done a lot of robotics during my bachelor's in Spain, when I ended up in Munich in 2017 doing this double degree in mechanical engineering, I was not really sure what I wanted to do. I start seeing my colleagues, some of my friends do this thing called ML and deep learning. I'm like, what the hell is that? I've just been doing some very basic Python stuff. I didn't even know how to code properly. And I see these guys building neural networks. I didn't even know what that was back in 2017, detecting objects, predicting the geometry of things. I'm like, that's pretty cool. I want to learn about that. So I remember one night in my student room in Munich, I'm just going to drop everything I'm doing with these Mechanical engineering thing, I don't even want to do that. I'm just going to go to the informatics faculty tomorrow, see what I can sign up for. I won't get the credits, who cares? And I'll just figure it out later and just try to get into a PhD position there. I actually kind of realized at that point, I actually want to go deeper into learning all of these deep learning AI stuff at the time, if you will. And that kind of shaped me up a little bit. That was my light bulb moment. And that lightbulb moment, I think it's connected to the entrepreneurial journey in that I wanted to have something to do with the real world, no? Like, I wanted to see applications of it. I think that was starting to, like, instill in me a little bit of that entrepreneurial mindset. I wanna build things that are tangible. I bought a couple cameras, I set up, like, a rig, and then I could, like, generate the geometry of my room. I would just, like, record it and, like, generate the 3D geometry.

Wow.

Yeah, I could maybe build a product out of this. I don't know. And then during the PhD, while I was getting frustrated a little bit of just simply writing papers, although I was building really cool stuff that actually worked, I really wanted to bring that to reality. And I was very frustrated that even during the internship at Meta, I was not able to truly see that in practice and come to fruition. It was just like a research project that Probably someone took over. Fun fact is that I actually waited my entire internship for a certain dataset of faces. I wanted to do face recognition and stuff. Never came. I had to do other type of geometry, like full body scans or whatever. So that was like, okay, I know I want to build stuff. I know I don't want to build it in big tech because it takes ages. So I guess my only alternative is to build something myself. And then I dig deeper into all these MBAs. I do find this online MBA, the Power MBA, some Spanish guys that built an online MBA. Okay, I mean, I'll do that. How bad can it be for $500? And it was pretty decent actually. And that stirred me up, like, there's something here, I want to do more. So, at the time, I was also doing, in parallel to the PhD, For fun, just looking into into buying property in Spain. In Spain, I mean, you can find a place for like forty k euros or whatever, rented for five hundred five hundred dollars or five hundred euros a month. So like super super good ROI, especially in certain cities around Madrid and especially not in Madrid. No, like so in my in Guadalajara where I grew up, I'm like okay, I'm gonna buy five houses. I'm going to ask for loans, like zero interest rate or whatever. At the time was like super good interest rates. And I'll just like get rich doing that.

Yeah.

And then I ping my buddy Luis and Javi and I'm like, guys, let's do this. Let's like do this together. And Luis is like, Luis, our CTO, he's like, dude, no, let's build a software to help us find the best houses, and then we sell the software. I'm like, yeah.

That's thinking.

That is, that's my, see, like, you get like the default to action, no? Like, I just wanna like do stuff. And then like, Javi and Luis were more strategic, right? So that was like a little bit of the trigger. We applied to some accelerator in Spain, in Barcelona. I mean, this is quite funny. Like, we had one of the advisors, He was like, guys, this is super interesting. Let's build it together. I'll get 20% and I'll advise you. I'm like, that sounds like a lot, but I've never done a startup, so I guess it's right.

I'm— yeah, I understand that. It's, you know, you're the first time around, you're just kind of poking and figuring stuff out as best you can. And, uh, especially when you're looking for, like, you know, ideally your advisors are the ones who would then tell you that's a bad deal. But when you're finding the first advisor, that's maybe the one that could rake you over the coals a little bit. English is your second language, is that right?

Yeah.

Spanish is my first. When did you start speaking English?

Maybe around age 5, 6.

Oh, very early. Okay.

Relatively early. Yeah. My mom found this supposedly bilingual school. And she was like, okay, I'm kicking you out of the current school. I was like 5. I'm like, okay, I guess I'm changing schools now. Uh, and she put me in this bilingual one. We had like a few teachers that were like either British or American. So I, my accent is all over the place. I've, after living in Germany for a while, my accent actually got like very Germanized, which was funny. My brother Javi would make fun of me.

Is that, is that common in Spain that, that kids would go through that and be learning English at such a young age? Like, I mean, certainly it sets you up. Your mother never could have known that. 20 years later, you'd be building a business in the US. But if you hadn't learned English 25 years ago, I don't know that it'd be as easy for you to be sitting on a podcast right now and having an English conversation as fluently as you are, or really just building a business. So I'm curious, is that more common where you were, or did your mother have some foresight that others didn't?

She did have a lot of foresight there. My generation and my brother's generation, it's like 10 years older, definitely did not have that exposure to these opportunities. This was like a bit of a one-off in our city, actually, and probably like one of the very few in Spain that were part of this British Council. Like, this British Council would actually find like a few schools and they would give them like access to teachers that they could send over. So, very lucky, very lucky to have had that. I think now it's more common, especially because you have exposure to TV shows. Like, my nephews, they speak almost like native speakers, and they're like 5, 10, and 12. And I'm like, guys, this is crazy. Like, when I was your age, I would just say, hello, my name is Pablo. I'm here to work with this in this park.

Yeah.

So, I mean, And now it's actually, it comes really natural to just, actually many times it comes more naturally to speak, to explain something in English than in Spanish. I did a podcast in Spain and it was terrible. It was terrible. Like, I would just like not find the words to explain. We're building an AI workforce for the enterprise. How do you say an AI workforce? It was so bad.

Really interesting. I mean, how often do you speak Spanish now? Is it just very rare?

I mean, with my wife at home, for sure. And then with Javi and Luis, when we get more into fun stuff or like just talking about anything, we just talk in Spanish. But anything work-related, it's harder to explain in Spanish for sure.

Do you have any customers that are over there or it's all, everything's US-based basically?

We have quite some customer base, quite a customer base there. We have 4 offices right now. We have San Francisco, Chicago, And then Madrid and Barcelona. Barcelona is actually more of a, for now, more of our machine learning hub. We have 5 ML engineers there. Don't ask me how we ended up having that there, but we just kept finding good guys. Barcelona has attracted traditionally a lot of good talent from all around Europe. Actually, one of the students that I supervised during my PhD time, we poached him before he started the PhD, we poached him and brought him back to Barcelona where he's from. He's actually originally Chinese though. But so we have that. And then Madrid right now, we kind of serve it or use it as our main hub for Europe, although we're thinking about other locations. And we probably have maybe like 20% of our customer base in Europe. The thing is, a lot of the freight forwarding and international logistics, if you will, that we're doing today. And it's 24/7. We have deployments in Brazil, in Australia, Singapore, starting some in United Arab Emirates and Middle East, basically, France and Denmark. So it's all over the place in terms of geography. So we have to cover 24/7, basically.

Got it. Hold up, wait a minute. Let's give a quick shout out to our sponsor, CloneOps.ai. Logistics moves fast. There are calls to make, updates to send, documents to track. Your customers are always waiting. CloneOps.ai helps teams stay ahead with a platform that can be white-labeled and a marketplace of AI agents built for real logistics operations. Inside the CloneOps marketplace, you'll find agents for tracking, Documents, after-hours support, finance workflow, fraud prevention, and more. These agents plug into your workflows and existing systems to handle the repetitive tasks, whether it's making calls, sending emails, replying to texts, gathering info, or writing updates. Your team stays in control with real-time visibility into every interaction. And because the AI is always turned on, operations keeps moving even when your team can't. A standout is carrier screening plus voice ID. Which ties a voice to a DOT number and alerts your team when that same voice appears across different or multiple DOT numbers, helping catch fraudsters before they cause damage. No one else has this. And for companies building a long-term AI strategy, CloneOps.ai gives you the ability to white label, build, own, and control every AI use case across your business. And this team is run by David Bell, founder and leader who I've known for years, a good friend of mine, and someone who Frankly, anything he touches in this industry turns to gold. I trust him with what he's building here. If you're ready for automation that supports your people, visit cloneops.ai. Now let's get back to the show. I have questions about that, but I'm not going to ask them yet because we'll get stuck on the business and I'm not ready to go all the way into the business yet. We will get there though. So just one more, I guess, culturally, I'm curious because I don't know enough about Spain. So I'm going to take the opportunity for myself and I'm sure plenty of listeners will appreciate this too. Um, you know, growing up in Spain, what's kind of— as, as, as an American, I don't know much. So like, what's your favorite— what was your favorite thing culturally about, um, growing up where you did that's very different from kind of American culture?

Well, I've only been here for 2 years now. I think the biggest difference, and this probably depends on the city you're comparing But across the board, I would say that Spain, from an early age, since you're like, I'm gonna say 9, you probably just walk alone to school. You just hang out with your friends after school. You go to the park together. It's very, it's very walkable. Like all cities are very walkable. And I would say that's pretty much applicable to all of Europe. Now in Spain, since you have like pretty good weather generally, winters are pretty harsh though, but like in, in general you have pretty good weather.

Well, harsh is relative because it's like 5 degrees in Chicago right now.

Harsh is relative. I agree. I agree. That's a good point. And I've lived through that in Munich. It was also like pretty painful. But yeah, no, growing up in Spain, you get to enjoy the city more, I would say. I think that's why we like San Francisco a lot, my wife and I. Because it's very walkable in certain parts. I mean, obviously, maybe you don't want to go to certain places, but in general, it's very walkable, very neighborhood-y. Like, you can go to a cafe and like just meet your friends there, go play some padel, learning some pickleball as well now.

Very cool.

But yeah, like, you would do a lot of things out there, like outside with your friends, family. Weekends would be like, you know, going to the going to the park with friends, going to the church with your family. And then, like, you go and we have the post— like, the pre-lunch, the pre-lunch, like, maybe 11 to 2 PM, 11 AM to 2 PM. You would have, like, that time range to basically just, like, have a beer, grab some bites. Then at 2, you would have lunch. That's actually a very interesting piece of part of Spain, like, you have lunch very late, like, pretty late, like 2 to 4 is like, like lunchtime. Um, maybe if you do work in Madrid or something, like, you maybe do like 1 to 3, but, uh, lunch is like a very big part of the day. And then another interesting thing in Spain is how, how late we, uh, we have dinner, which actually shapes a lot of the things we do during the day. Like, um, if you walk around Madrid, At 9 p.m., 10 p.m., it's so lively. Like there's so much going on. The city never sleeps on it, really. It's it's very lively. It's very nice. You feel you feel a good vibe. Generally, I liked it a lot.

Do you feel like you'll want to go back, or are you you know U.S. for life now?

Man, I wish. I mean, things are not looking good there in terms of the economy and. They see, I mean, I'm positive for the future. I think for now, definitely not. Although we do have, we're very bullish on serving European customers. And I think it's a super interesting market. But personally, like me going there, it would be pretty tough for some. For many reasons, but the culture there also is very slow, which is good if you're maybe retired. But if you're trying to build something, unless you're in very specific hubs in Barcelona, Madrid, it's almost consuming. Almost people judge you for doing the work you do. Why do you work so much? You're putting too much on yourself versus here, they actually see you as Oh, like you're trying to do something cool and like building something that adds value to society. That's so awesome. Like kudos to you. There we have a little bit of a stigma. There's a lot of— we don't have to get into politics, but there's a lot of parties, if you will, that would judge you for like building a company just because you're like an entrepreneur or you have a small business. Small businesses, man. They treat them as garbage. Like, they— there's laws in Spain where you have to pay taxes even if you have not generated revenue that year. There's like a pre-tax on like entrepreneur, like business owners, basically.

Very anti-business, basically.

Super anti-business. It's crazy. You compare that to maybe like a UK where it's pretty simple to get started with your own like side hustle. And you get depressed. So that's basically where I sit.

Yeah, got it. Understood. So Happy Robot's not going anywhere. Staying in— you're staying in, uh, in San Francisco?

Yeah.

All right, well, let's, let's, let's transition into the business itself now. Um, I do appreciate the color on, on your own kind of origin story. Um, how did all this come to be? You know, your brother was CFO at this olive oil company. You and your Bestie Luis, we're jamming. You know, how did we, how did we come up with Happy Robot? What was the origin of that business?

2023, we get accepted into YC with some random computer vision platform we had built based on my supposed experience on computer vision. So Pablo knows how to do computer vision. Let's capitalize on that. So we did that. We get into YC, we go through the batch. We don't find more than one more customer than what we had before we entered. We probably had like $70K in ARR when we got in and we came out with like, I don't know, like $80K, I don't know, like something super tiny. And we're like, okay, this is not working, guys. And on demo day, there's this demo day, this presentation to investors, the last day of the batch. We woke up that morning already knowing, and actually we Slacked our partner, Knowing that we would pivot, that we would just like not do that, uh, computer vision platform anymore.

Did you know you're gonna pivot to?

Nope.

You just, you just were supposed to present, you're like, hey, we— this isn't gonna work.

Exactly. We just, we just knew. We just felt it in at our core.

Okay, so then Pivot hell.

Pivot hell for 3 months, maybe 2 months or so. That was September when we decided to fire all of our customers. And then we go into this search mode, to this life quest to understand what we want to do.

Mm-hmm.

And we kind of kept trying to find what experiences we had gathered ourselves.

So wait, is at this point, is it just you, Javi, and Luis?

Yeah.

And has Javi quit his CFO job?

Yeah. Yeah, he quit.

Oh, wow. Okay.

Before he actually quit, before we knew we got into YC, like around May that year.

Okay.

So he had already quit. Um, and then literally the day he was— this is funny— the day he was arriving in San Francisco, Because we had decided to move to San Francisco regardless of the outcome of YC. We got the call from our partner, Diana Hu, that we got in. So, he was literally unboxing and he got the call that we got in.

That's awesome.

That was great. That was one of the best moments of our lives.

So, we go— Hold on. Was YC worth it? Was it a meaningfully beneficial I mean, you went in with one business and you left with that same business, but it didn't work. Uh, so what did you really get out of it?

A really solid business now.

Like, uh, this came from that, even though, you know, at the end of it you still were— you, you know, it feels like if I understood you correctly, you get to the end of YC to Demo Day and you're like, this isn't gonna work, we're done with this whole business that we spent our whole YC term doing. And then you spent 3 months trying to find the idea for Happy Robot. So like, I'm trying to understand like where the value in YC was. Please help me kind of gauge that.

In knowing that we could aim for bigger.

Okay.

Knowing that we could go bigger. I think that was the biggest value. Realizing that there's very ambitious people in this country and in particular in this part of the of the world trying to build very ambitious businesses just because they know that even though it's going to be hard, it's possible. So one of the things that one of our investors, Ade from Base10, so we have our main 2 investors are Andreessen Horowitz and Base10. We'll talk about that in a minute, I guess. But Ade, he's a Spaniard as well. He co-founded the Spanish Facebook, Tuenti, at the time. He sold to Telefónica. One of the telecoms in Spain, and he tells us that after he sold his business to Telefónica and came to San Francisco to do his MBA and whatnot, and then built another company that he sold to Workday, he was always thinking too small. And he urges us every time we talk to to think big and to be ambitious and to to know that we can do it. No, I think we we have a little bit of a stigma in Spain that you know like we're not we're not like that smart or we're not that. good. Like, you know, like the Germans are so organized and so well polished and the Americans are so well spoken and they sell so well. And then we have a little bit of that stigma thing in Spain where we don't value ourselves enough. So I think YC helped us rewire ourselves a little bit and think bigger. So, the cliff was there. We just saw how the cliff of having to fire your customers was scary, but it helped us cross the line, cross the finish line, if you will, or cross that bridge to know that we had that backup or that backing, if you will, that backing from YC as a little bit of a brand, and knowing that other people were also in the same boat. That helps tremendously knowing that maybe 50% of your batchmates in YC are going through the same or worse. So that helped a lot. I think that's why we exist today, honestly.

What do you feel like outside of dream bigger? Like, I guess actually just take me to the point where you realized what Happy Robot should be. Like, how did you get from You know, this, this original business isn't gonna work. We're firing our $80K of ARR. You're all gone. Take me from there to we know what we're gonna do.

Mm-hmm. So after pivoting, we go into introspection and we say, guys, what have you— what have we all done? And then we look at Javi because he's the one that he's actually worked at all in.

What have we done? And you're like, nothing. Played soccer.

It's vague. Does that count? I worked at Meta for 6 months. I mean, Luis had been working at HPE as a cloud DevOps person. I mean, Luis is extremely smart. He's like the hacker in the team. He builds whatever you ask him to build in no time. He's also super creative. Very, very, very nice guy. Extremely nice guy, I should say. He basically wanted to do stuff for a while on voice, basically some form of English trainer or some form of psychologist, the typical app where you can just go B2C and sell an app that people can use and make use of it. So voice at the time was growing bigger, but then we say, what are we going to do there? We don't even know. We're not a psychiatrist. We're not a teacher. We're not teachers. We don't know how to teach English. So then we killed those like B2C companies. Like, okay, Luis, forget about that. Like, let's again, look at Javi. Let's look at Javi. Javi, enlighten us. What do you have? And well, Javi had been working for a while in this company called Diolio, which is the biggest olive oil distributor in the world actually. And he was leading North America as CFO. So he had gone through a lot of pains in managing his supply chain. In particular, he was working with one of the top 10 freight brokers, and I'm not saying who, they're great guys, but he was working with a certain broker. And at the time, they didn't have enough bandwidth to basically just track all his loads. So he was messing, getting all his scorecards from Walmart and Kroger with basically gave them like bad scores and like some fines for late deliveries. And he ended up hiring interns to call drivers, like essentially bypassing his broker. They were not happy about that, but they figured it out. But yeah, that was how we started digging deeper into, okay, there's a problem there. There's— you had a pain point, which is you're getting late deliveries to Walmart and Kroger. And they're getting fines, and he had to hire people to call and email drivers and carriers, whatever the carrier was, right? So we started digging deeper there and said, okay, there's actually something here. What if we build a voice agent? Because I wanted to do LLM stuff. I hadn't really done any LLM stuff. We had done, remember, all this computer vision. So it was a mix of, hey, there's this cool technology that is unlocking this problem that Javi might have. We don't know if this is a big enough problem. That's the issue. We don't know if it's big enough. It might be that Javi's situation is unique when he was working at this company. Maybe not. We'll explore. So what we do is ask this company of his and we dig deeper into the whole broker space. So we go and I think sign up for FreightWaves, November '23. We send Javi there and he demos. a terrible voice agent at the time. I mean, it was decent at the time, but now it would be so terrible. Basically, the latency was super high, and it was a track and trace call, and he demos that to folks. And we get some more leads and more leads. And then I ended up in the Tech Bro Tuesday with Reid and these guys, actually thanks to our friend, Adrian, from GateGo and Money Transfer. He connected us to Reid and I present in one of these TechPro Tuesdays in January. This is already January. And I do a demo and they're like, wow, this is so good, man. And it happens to be that one of the— there was a couple of VP, like one VP from one of the top 15 brokers and a data science guy from one of the top 40. The top 40 was like, both were emailing me like, we should talk. The top 41 was like, we should talk and I want to deploy these tomorrow. It happened to be one of our first, actually our first broker customer, Circle Logistics. So Andrew Smith, who's a great friend of ours, he got this demo from his data science guy and he connected us to his CFO. He gets approval for us to work together and turn fast forward to April 2024. We're deploying an inbound carrier sales agent. I didn't even know at the time. I mean, we were just pitching track and trace, so I didn't even know what that was. I mean, coming you didn't even know what carrier sales was. Yes. Yeah. I mean, now I think you could put me in a floor and I would be like top, like number one broker.

You think so?

I think maybe not like cradle to grave. Like, I don't know how it would do with customers, but like just—

Have you sold a load yet?

Yeah. Yeah, yeah, yeah. Because I had like the bot messing up at the beginning. So I would actually call the, like, this is April '24. Like, we were still refining stuff and maybe the bot would say no to like a decent carrier. I'm like, wait, wait, I'm going to call that carrier. So I would call them back, hey, this is Pablo. And they're like, what do you mean? I just talked to you. No, no, like the real Pablo.

This is the non-AI Pablo. My AI screwed up.

My AI told you not to book this flight, but yes, I would've. So I've actually booked flights and I've actually, yeah, I've sat down enough with folks to understand how draining that can be, but how empowering it can be. Like if you manage to like create those relationships with your carrier base, That was fascinating. And that's also why we're now pushing a lot for like our workers to gather details and insights, our AI workers, if you will, our AI workforce for the brokerage space, to gather a lot of insights for the carrier base. We actually have like what we call Bridge. It's one of our brokerage products, which has a load view, like a track and trace view, kind of alerting and exceptions, like exception management. Then it has a carrier CRM kind of view.

Mm-hmm.

where you can keep track of, hey, Pete from ABC Trucking, he typically likes running these Dallas to Kansas City loads, and he has a daughter that plays soccer, and he likes calling us to update us on loads, and then he starts talking about soccer. So those type of insights are really interesting, especially now that we're doing a lot of Outbound carrier sourcing via email and phone. It's super interesting how we can connect the personal touch and even say that it's a yeah. We we now say both in tracking trace and and and carrier sales. We we typically say, especially if it's outbound, like hey, I'm I mean this is for the customer to decide. Like we we're very like if you will very enterprisey. So like our customers decide what they do with their workforce. We're just providing the the workforce for them. So they decided they want to say hey, I'm Pablo from XYZ Logistics, the AI assistant to Andrew, he would love to call you, but now he's busy. I'm just calling on his behalf to see if you guys would have any capacity for us tomorrow. So, I started seeing by sitting next to these folks how important the relationships were, but also how draining the non-relationship aspects of things where as well, a lot of this is just rinse and repeat. So we wanted to position the humans as the guardians of the exception. And that is what really triggered the whole going from just solving for a certain use case with track and trace or carrier sales to actually saying, wow, this is much bigger. We can actually be reimagining how work gets done in supply chain and physical operations. And we want to build a workforce for these enterprises to solve for that pain, to basically position the human as that guardian of exception that manages the 1% or 10% exceptions and edge cases and have a bit of that workforce of agents for sales, for customer sales, for carrier sales, for tracking trades, for accounting and back office, just working for the team. And initially it was super interesting to see all of these reps We're fearing the agent. The next year, basically this year, and we have some customers that do these sort of analysis, they ask their teams, how do you value these tools and what do you think about this tool or that tool? Last year we were feared. This year we're loved. It's their assistant. It's almost like they have personal EAs. They can talk to them on Teams. They can say, hey, They give it names. They call them names, whatever name they want to call it. Hey, Maria, or Tracy for track and trace. Hey, Tracy, check on this load. And they text it on Teams. They trigger Tracy to do track and trace for them via Teams or Slack. Hey, Casey, for carrier sales, can you trigger an outbound campaign to try to see if we get some capacity for these can-get load? We still haven't won from the customer, but Let's, let's see if we can get some, some market data, market data in, in the wild. So that was, that was really interesting to see that transition.

Okay, let's take a quick timeout and give a shout out to one of our sponsors, Rapido Solutions Group. Rapido connects logistics and supply chain organizations in North America with the best nearshore talent to scale efficiently and deliver superior customer service. Rapido works with businesses from all sides of the logistics industry, which 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 Icaza, 2 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 gorapido.com. That's gorapido.com. Now let's get back to the show. So it's interesting because the path— I can let you talk and you could fill up the whole podcast. I wouldn't have to even ask a question. I could ask you one question, you could go for an hour. I love that about you.

It's my Achilles heel.

No, it's not a bad thing at all. It's a challenge as a host because you have to decide, like, when do I jump in? Because I've got a question here, and then you keep going, I get a new question and then a new question. And I was like, where do I even go from here? But what I'm hearing is Day 1, this was a track and trace communication tool. Day 2, it became a carrier sales communication tool. Day 3, and the days aren't important, it's more timeline, but eventually it becomes a data analytics tool to some extent, or data capture and then presentation. And I'm curious, does the evolution from communication to data, does it eventually become like an ecosystem? Does it replace a TMS or does it live in a TMS? You know what I'm getting at?

Yeah, we've thought about that a lot.

And I'm curious, where do you see— answer that when you get a chance and then morph into In terms of the maturation of the business, are you a baseball or you're a soccer guy? You're a soccer guy. Are we at the 10-minute mark of the first half or are we at halftime? Are we in extra time? If the business's maturation is a soccer match, where are we in the timing of the game? Of the match?

Sorry, I want to stick with it.

Yeah. In terms of What this business will be?

Super interesting question. For the first part, the way we started was by building the execution layer. And we see the world in 3 layers. We see the execution layer, we see the data layer, and we see the intelligence layer. dashboarding that you talked about, that data piece that you talked about. So the execution layer is basically agents or workers getting things done. It's that track and trace agent. It's that carrier sales inbound or outbound, either via email or phone. It's that POD collection use case. That's execution. That's where we started. And we built, I would argue, and we don't do it We don't do it enough service or we don't do enough in putting these out there that we have probably one of the best orchestration platforms from a workflow builder perspective to build agents. I mean, we had to build our own orchestration platform to build agents. Have you played with any Zapier-looking tool?

No, explain this to me immediately.

Imagine basically if this, then that. Then send an email, then make an agent. So what we've built to scratch our own itch was the ability for our forward-deployed engineers to basically build agents really, really quickly for our customers. So we built our own platform to build agents. And any one of our customers will probably be able to attest to that, that we have a super powerful tool to build agents. maybe name exactly the customer, but it's, I'm going to say, one of the top 2 or 3 fleets in the US. They had these couple people, almost like PMs, data people, but they knew about the business as well, obviously, just building agents on our platform. So that speaks to the fact that we're not just building on top of some other agent builder provider. We actually are the agent builder provider for these enterprises. And we always knew we had to be that because a CH, an RXO, a WorldWide Express, a JB Hunt would probably not want to have anything that is not a developer tool of sorts, if that makes sense. They would not want to have a canned solution. They would not want to see a black box. They would want to see a developer framework, an orchestration platform for them to collaborate with their different agents because they will maybe build agents elsewhere. They will maybe build systems elsewhere and they have systems, their own TMS, whatever Salesforce instance they have. They have a plethora of systems that we had to connect to. So a black box would not probably make it a service. That's why we had to build a developer infrastructure, a developer tool, or if you want to phrase it in a different way, an AI operating system for these companies to build on top of. Because remember, we're targeting the enterprises. We today are not serving a pre-canned agent for track and trace necessarily. We have that and we can offer that to customers. We have implemented so many carrier sales and track and trace in POD. agents. So we, we have those prebuilt agents that we, that we have templatized over the, over the months, uh, years now. But we're more than that. So this covers the first, the first layer, the execution layer.

Before, before you go to the next layer, what might be different than, let's say my, I have a brokerage, XRO Brokerage, um, just to make some letters around. And I just take your canned template, like your track and trace, your carrier sales, your POD collection, like your 3 functions, and use those. What might be different from my utilization of those 3 versus, versus, uh, JB Hunt wanting to use the developer tools? What would they build that would be different than— The 3 functions that you— the canned templated ones.

This is super interesting, and it actually connects to that data piece because every business will have their own systems of record. They'll have, especially at the enterprise level, they'll have their own API connectivity. They'll have their own way of doing things, their own SOPs. This is where if you can connect this execution layer of agents doing work, they can speak, they can write, they can read emails, they can process documents, they can do all of the basic agentic stuff. If you can have them do work that then they contribute back to the data sources, that carrier sales agent is putting back all the options it's capturing into the TMS, the track and trace agent is actually cross-selling In a track and trace call, maybe it's proposing another load because it has access to the data layer. We're enriching that data layer in such a way that the action generates information. That information is creating an insight. That insight is actually generating a proposed next action. For example, say that your Calling your vendor to see if your trailer is ready. You have your maintenance team just calling vendors all around. Is it ready? Is it ready? Well, the sooner you know your trailer is ready, the sooner you can put that back into your network. And that's capacity that you have on the asset side. That's capacity you have to keep your business running. But if you have people just executing that, it's going to take a little bit longer. So that's a lot of the value of connecting the data layer, of knowing, like this interconnected business, of knowing when a trailer has been repaired, you put it back into the network, it's improving other parts of the house, maybe the brokerage side of that asset company. So by connecting the agents to the data piece, we're enriching it and we're generating even more insights and more potential actions for those agents to do. Now, these 2 layers only cover so much. You still don't know what your agents are doing. You want to have an observability layer on top, like an intelligence layer, so that you could almost just talk to your data. This is what we call a— I mean, it's basically like ChatGPT for your business. You connect by connecting to all of the data sources and by connecting to the work the agents are doing. You can now talk to your data. You can ask, hey, what were my margins for these type of, for these particular lane, for carriers that have more than 10 trucks? Like this particular query, our system can reason about that query and get it for you versus these folks having to like click everywhere on the TMS, like filter by carrier size, like the lane, blah, blah, blah. It's proposing insights, it's surfacing the edge cases, it's surfacing the insights to the business leaders. We had one of our freight broker customers, they wanted to build their own GPT almost. And this is really where the idea came from because they wanted to ask questions. And in one of these questions they asked, they uncovered $3 million of potential savings based on some stuff that I cannot talk about, but they uncovered $3 million of potential savings just by asking a question to the system. that has gathered data from their human teams, but also from the agents putting data back into those data sources. So this combination of the execution, the data layer, and the intelligence layer or the operability layer, almost like the one glass pane, was really, really powerful, especially at that enterprise level to go and be more helpful versus just like a, point solution of carrier sales. So that's been really the unlock for us. And I think it connects to the second question, how big can this be or where are we? Man, this, I mean, we're really, we've grown a lot since we did our Series B. I can share numbers, but I mean, I've shared in other podcasts, we're over, we're in the 8-figure realm. In revenue, which is fantastic. I couldn't imagine in less than 2 years just generating the revenue that we're generating and working with customers like, I mean, DHL shared a press release recently about their partnership with us. We're working with 8 of the top 10 freight brokers today, 2 of the largest 3 ocean carriers, a few airlines now on the air cargo piece, many of the top fleets in the country. The sheer amount of work that we are doing today is nothing compared to what's, what's coming. Now, to your point, I think we're minute one.

Minute one. I love it. That makes sense. I was hoping for an early, early, early time. And when you say that, do you mostly mean in terms of I guess answer this 2 ways. One is, is it more minute one in the sense of the products that you can build or will build, or is it more minute one in terms of the application or implementation of your products within the— you said 8 of 10 brokers. Is it a lot still of, okay, let's start with the track and trace, let's do that. Or is it, hey, you know, we've got these 3 layers that work across multiple functions in the business, and a lot of customers have fully deployed all of that and it's functioning? Um, like, help me understand that a little bit better.

Mm-hmm. I think it's mostly on the latter. It's mostly on the deployment of the products we've already built. We struggle so much about whether we open up, for example, our agent builder or not to the world, because we don't want to expose our IP, but we know that what we have is best in class. And when we show it to billion-dollar company X that have looked at Palantir, They're like, okay, well, this is on par or better, depending on what type of the equation you're talking about. If you're talking about more of the data piece versus the execution layer, actually many times now we're finding ourselves putting in the execution layer that we've built on top of Palantir's Foundry for very large manufacturers and enterprises we're working with now. If you see it from that perspective, we're definitely more on the latter. Basically, it's now more of a go-to-market approach. It's more about embedding ourselves in the customers and helping them realize the value that they're either paying for or have not even uncovered because they've not even gotten creative enough to work with our teams, with our deployment strategists, and with our forward-deployed engineers to think about new use cases and new applications. But then it's about connecting them all because you might be doing something again in maintenance that connects to the brokered side. And then a lot of lights start blinking and like, oh, boom, this is a lot of good ideas. And if you connect them all, it's really the interconnected business. So that's what we are really striving for, to interconnect these businesses.

I'm curious, what's the easiest function or element of the business to build and what's been the most challenging? And by challenging, I mean the most unique Differentiated piece versus the piece that you'll have 50 competitors that can build or have built the same type of thing?

I think our biggest struggle so far has been putting it out there that we are not the 10th or nth. I mean, actually, we're probably one of the first ones really in this space, specifically in freight Freight brokerage, but putting it out there that we are more than just another canned agent. I was talking with someone from from Gardner, the research company, and they're like, "Okay, so like, see you sell like these pre-packaged agents." Like, no, we're we we are the we can be the functioning workforce, the AI workforce for these enterprises. I'm not gonna be selling to a 20-person broker. I mean, that's, I'll put it out there, you know, like we're not there yet and I don't think we'll ever wanna be because it takes a lot of, takes almost like a B2C approach. I mean, like it's a B2B approach, but it's very harsh to sell to a 10, 20-person broker that can actually go out of business tomorrow. I mean, I'd love to, but today our focus is on the top enterprises in brokerage, top fleets, top freight forwarders, ocean carriers, airlines. So we're looking at the enterprise layer of supply chain, also at the shipper side, if you will, like retailers, manufacturers, distributors, and how can we make their operations more efficient? How can we build an AI workforce for these enterprises that interconnects the way they do work? Because we want to reimagine how they do work today. So that's been the biggest challenge, you know, like getting it out there that we're not just like your nth carrier sales agent. There will be folks that might even build a bit of a load board looking platform where they maybe like do some offering for carriers and brokers and they connect the capacity and they do that load board game. We decided a long time ago, we're built for the enterprise. We're your workforce. We're not sharing your insights with the nth broker because that would be messed up. All of the insights I get or you get in your workforce, they should remain in your workforce. And the compound intelligence that you get by having agents working on your data every day, It's exponential. The sooner you build an agent on Happy Robot, it's learning every day. It's learning about what it should do, what it shouldn't do. It's getting feedback from the users. It's improving the network. And this is really interesting across the agentic space. If you actually improve an agent or a piece of an agent, it actually scales across the board. You don't have to retrain your workforce, your Your human teams. Now, your humans team, your human teams are still going to be helping those agents get better. So it's that combination that is really specific to every company. The way Broker X works is different from Broker Y, and that's their unique ability to to grow their business. We're just empowering the way they do data work. That's been the hardest part: getting it out there versus oh, here's just like another. Carrier sales or tracking trays, uh, free canned.

But holy shit, that is what everyone says when they see Genlog's truck intelligence platform for the first time. Founded by ex-CIA operatives and fueled by 15 million daily images across a nationwide camera network, Genlog gives you the power of total market capacity while also defending you against fraudulent carriers. Holy shit is what Genlogs customers say again when they see the ROI from covering loads faster with fatter margins. Holy shippers! That's right, Genlogs unveils the locations and lanes for all the shippers in America. In the era of artificial intelligence, nothing beats actual intelligence from verified by video data. See what Genlogs can do for you. Check out a demo at genlogs.io. Again, that is genlogs.io. Tell them Andrew sent you and they'll include their carrier compliance suite for free. And if you haven't already, I interviewed their founder Ryan Joyce last year, and it's one of my personal favorite episodes that we've done. Check it out. Now let's get back to the show. Yeah, because I think the— at least my initial understanding of Happy Robot is it's a voice bot. That's certainly what I was early on under the impression of it was. And it sounds like you're telling me it's a lot more than that today. And I'm really interested in the idea of kind of reimagining your workforce at the enterprise level. So let's just make up an example. Let's say I'm running a brokerage that's got 500 carrier reps and 500 customer sales reps and 300 customer operations reps doing scheduling and all the other fun stuff that operations folks do. And I come to you and I say, let's go all in together, reimagine my workforce, and I deploy all of your tools, your whole thing. I say, here's the check, reimagine it, what, what is that? What changes? Like, what, what is— what does the end goal look like? And you can generalize it, obviously, because this is a made-up example, but help me understand, like, what about each person's role changes? Am I, am I repurposing people? Am I going from 500 care reps to 250 and moving these other people to do other things? Like, what does that look like? And if there's a change, if you think today versus a future state, I'm curious what those 2 might look like differently.

Many of our customers paint this super clearly to us. We want to grow. We don't want to— we're not going to fire anyone because we love our people, but we want to grow 2x in the next 2 years, whatever the number is. How do you do that without having to scale with people? And not because it's bad necessarily, but because if you don't scale now with agents, with AI agents, you will not grow 2x next 2 years, but someone that is leveraging those agents will. So folks that are thinking about it the right way, they see that unfair advantage that whoever leverages that AI workforce will have versus them. And this is a bit of a dilemma that I have. Like, if smaller brokers don't get access to this sort of technology, which I think they will, because there will be players that maybe cater to them, and they will be maybe hyperscalers that go from 10 people to being the next $10 billion broker, or 5. It's a bit of a dilemma that I have now. Like, if we're only serving the enterprise today, we might be missing out on some unicorn. But the way these folks are looking at their workforce is we don't want to fire folks. Obviously, if there's low performers, they're just going to naturally just churn themselves. I mean, from a class of 100 reps, maybe 5% are going to stay there in the next 3 years.

Of course.

And talking with some folks, Recently, we're like, man, like every every new class is like harder to train. Like they're just like more scared to do stuff, to send that email that to a customer, to to make that phone call to the carrier. They get shouted at. They're scared. They quit. So how do you grow a business that is inherently messy and complex? And and my goodness, I kudos like hats off to all of those operators that are like running these businesses every day. And I don't know how. how they would— how I would manage that. Like, it's amazing. They're thinking about how do we bring this technology to help us accelerate? That's how they should be thinking about it for this transformation, right? Maybe we go from 500 reps to, in the carrier sales rep side, to 400, 300, whatever it is, while people just naturally turn. But we're going to grow the business 3x, even if we have less people doing the work, because the ones that remain will be those Guardians of exception, which I like calling, managing a bit of a workforce of their own executive assistants, almost like their own reps, their own, um, assistant reps, if you will. So that, that's how I think people should be thinking about it.

And, and what does that day-to-day change look like? So if I, you know, before I hire you, my carrier, my 500 reps, each one of them has 25 carriers that are their owned carriers. And I'm Joe Smith, I work in the Chicago office and I'm responsible for inbound freight to the Midwest, any loads delivering into Illinois, Indiana, Ohio, Michigan, and Wisconsin. I'm on a team of 20 reps and we cover every load coming into those states. I have 25 carriers based in those states. Each of them has 20 trucks. I call them every day or they email me where their trucks are every day, and I'm booking their loads in advance as much as possible. And I book about 20 loads a day. I'm a good rep. I make pretty good money. And there's 499 other guys like me, Joe Smith. And my day is mostly dealing with those carriers. And then I also have to call new carriers every day because we have A bunch of spot loads and different loads that aren't on routing guides. They come in, my system does some automation. So I've got some primary lanes with some of these carriers that when Kraft sends us the load, the system sends it right to the carrier and I just talk to them about issues. I deploy Happy Robot. How is my life, how's my day changing? What roles and responsibilities are you taking over for me? What am I doing now?

The biggest one is you're managing by exception. You're getting pings from your, let's call it a tracing agent, telling you, hey, I already made sure that the 15 loads that you're overseeing today are all good. No worries. Like everything's looking at, it's looking good. They're going to deliver on time. There's this one load that might actually be late. I already called the driver and I let them know that if they have any trouble, they should call us back. So I'll be on the lookout for that. I'll keep you posted. That's the message that that rep gets in the morning. Like, okay, things are looking good. Maybe across the day, along the day, there's an issue. A driver breaks down. They cannot fulfill that load. We have to that they're not going to make it for pickup. We're going to have to find someone else. The system will tell Joe Smith, hey, I'm Tracy again, Joe. This driver, he broke down. We have to find someone else to make the pickup. I'm going to actually already reschedule with the customer for an hour later so that we have a little bit more bandwidth. It's possible in this facility. They already confirmed. And now I'm actually— Reaching out to these other 10 carriers that we have, uh, that have done these loads in the past. So I'm already reaching out to them and, and see if they have any capacity to, to make it happen. Obviously, uh, I'm asking you for confirmation that we can go a little bit over max pay here. Can you— do you agree? And maybe Joe says, hmm, no, actually I know who to call. Uh, go get these, uh, ABC Trucking because They're definitely going to do this for me. I'll actually text them myself. Then maybe Tracy doesn't have to do anything, Tracy the agent. So Joe Smith controls that executive assistant of his, that agent of his. They can do 10x more work for him. I mean, it's a question for the broker to see how they commission their brokers now.

Mm-hmm.

It's a question for the broker to decide how many how many people they put on more of the mundane tasks like track and trace. I mean, I don't know who likes doing track and trace, but I mean, I wouldn't. I've done some stuff there and it's painful. Who likes calling for PODs? My goodness. Who likes entering load data? So I think the team that remains at these companies will be super strategic, will be very customer-focused or carrier-focused at the highest level, not at the execution level, not at the getting things done, but rather at the, hey, Mr. ABC Trucking, you guys have done this lane for us now multiple weeks, multiple times this month. Why don't we do some more of a formal contract here? We just have a bit of a longer-term commitment here. We have this customer, this lane, happening like multiple times a month. Like, let's do something more formal. So identify— we actually had these, uh, multiple— these identification multiple times of lanes or carriers that could actually be, uh, contracted carriers, if you will, uh, based on the spot loads that we were covering for them via email and phone. I think this is the, the how these reps will actually operate, by exception and by being a lot more strategic. I don't know if that answers the question.

No, it does. It definitely does. It just gets me thinking about a lot of things because just in this example, and this is a minor thing, but you mentioned, okay, so Tracy says, I called the facility and bought us an hour at pickup. So there's nuance to that. Where certain facilities, if you were to do that, it is the right— it's the smartest thing to logically get yourself the most time to make your pickup. But in some cases, a customer is going to ding you for missing the appointment and showing up later. And so there is a bit of a balance. And It's like, okay, it's 10:00 AM when we got this notification. Our appointment's at noon. We know the facility's open until 2:00, but this specific customer is really strict about the appointments in terms of the scorecard. And our most recent scorecards are just a few percentage points lower than we want them to be. So we really want to hit that noon. And I'm just curious how that process works and just how much can I teach Tracy to understand, hey, it's probably worth it for us to pay the extra $200 to get a carrier that can actually make the noon versus defaulting to the, let's buy the most time possible because that's logically what makes the most sense. Right. You see how there's nuance there?

100%.

Yeah.

100%. The short answer is this is for the customer to decide. This is their own SOP. This is in the rep's brains, actually. It's tribal knowledge that that one rep might have. So we cannot discard that. We cannot discount that piece. So it's going to be a mix of How much do we know about that facility? Do we know that it's like strict appointments? Do we know that it's first come, first served all day long? Do we know that Pete at that facility can actually typically just buy us an hour or 2 if we give him some pancakes next time we go there? It's that tribal knowledge plus the facility whatever facility know-how you have about that one particular example. But definitely Joe Smith should have a lot to say on that operation because Trace is working for Joe Smith. The way we put it on the Haprobot platform is you see a bit of an AI org. So you have— we don't give it a name, but it's like the AI chief of staff, if you will, managing everything. And then we have Tracy, AI head of track and trace operations. I have Casey, head of carrier sales. I have John leading sales, customer sales. So you have like this AI org, and then every one of those agents or heads of that I just described, they have like workflows. For track and trace, you obviously need to do a lot of different type of check calls. You might have to do like a pre-pick, Maybe in transit, a couple of them, if there's multiple stops, maybe like a pre-delivery for the POD you connected with the— right after you've delivered, you immediately call, or maybe you wait a day to call for the POD. Those things are in the tribal— are tribal knowledge, are in Joe Smith's head. So you have to incorporate that human in the loop. That's where the guardian of the exception comes in. So I fully agree that there's a lot of nuance, and that's where humans can actually communicate with their agents.

Is this where the kind of term of the year in tech companies, forward-deployed engineer, is that where this creates value? Like in this tribal knowledge understanding, like, you know, it seems like that's all— I don't think a year ago we'd ever heard the term forward-deployed engineer. And now if you don't say you have one, You're way behind the times. Am I right? And that's kind of what's going on with a lot of these companies.

Yeah.

And I'm curious, but is this the application of that where it's you're understanding the tribal knowledge of a broker and how to apply your solution in a way that consumes that and applies it Effectively?

To some degree, to some degree. I would say that part of the beauty of language models is that it can actually understand the nuance of Joe Smith texting back on Teams. No, don't do that. I know that we can overpay for this one. I have like, I can override this max pay here and pay $200 more. So do the opposite, do something else. I think the FDE is more so for rethinking maybe the way things are done. For example, we had a customer where we're doing a lot of these track and trace calls, and one of our FDEs working with their reps, one day they're like, hey, I mean, we should definitely mimic what the reps are doing today, or some of the good reps are doing today, which is— see if you can get a backhaul for them, or see if you can actually get another load, the next load for them, and just keep them in the network. So by having someone that is just listening closely to what's going on in the business and going on-site, or going on-site multiple times a month, you uncover a lot of this tribal knowledge. And you can put it back into the agent, but the agent itself, will actually learn about your own specific way to do work. Joe Smith saying, this facility, we should not mess it up with these guys. The shipper will get angry with us and they'll just find another broker. This can be learned by the system. And next time that Joe Smith is talking to Tracy, Tracy will remember. It's like, again, your EA, a good EA, should remember these nuances. We're trying to implement the same for these companies. We're trying to reimagine the way they They do the more operational work and many times some of the sales work, which is super interesting. We have done a few case studies and use cases with some of the LTL brokers where we just do campaigns for dormant accounts where we're emailing and calling customers, like smaller accounts maybe. Well, some of the accounts they cannot reach efficiently or in a way that would have ROI. But when we get a lead and we pass it to a human, hey, let me connect you with my pricing expert. You guys shipped a pallet of potatoes a month ago. You guys want to ship again? Oh, sure. Okay, perfect. Let me connect you with my pricing specialist. Wait a minute. Boom. You have a sale there. So those type of use cases in sales are also super interesting and it connects to the other operational part of the business. Which is again why we realized we cannot just be building siloed agents. It has to be part of a broader network of agents, hence the AI workforce.

Yeah, that's very interesting. Where do I want to go from here?

It's a lot to ask. I mean, we've been thinking a lot about these things.

Yeah, I bet. Is there anything that you've kind of refused to automate or that you feel like you don't want the AI to touch, like the agents to touch?

That's an interesting question.

And refuse is maybe not the right word, but anything that you've looked at and been like, I don't think we should go down that path, at least not yet, or maybe something like that?

I mean, definitely all of the Cold outreach, like pure cold outreach, especially via phone. We just know that it's not going to fly in terms of regulation. Email, it's harder to track.

Cold outreach to who? To carriers?

To customers.

To customers.

Yeah. Like, that's one that, I mean, we all, both our customers and ourselves, we understand that as much as it'd be interesting to like do that Pure cold outreach. Uh, if someone has not signed up for a call, especially in certain states, um, it's, uh, it's trickier to do that, uh, cold outbound with, with AI agents, and we don't wanna go down that path.

Why, why is it, why is that, why is that different for a trucking company versus an olive oil company? They're both businesses technically, and both, it's technically you're cold outreaching to either of them. I'm curious why that's a difference.

Especially if you've already worked with someone, for example, with a carrier, you've already worked with them in the past. Also with a customer. I mean, what I'm saying is for both carrier and customer, if you've already worked with them in the past, you have more of a say or you have an excuse to call them.

Yeah, that makes sense.

Hey, Mr. Carrier, you guys should, you guys moved this load for us or these lanes, did this lane for us a few A few weeks ago. You're going to do it again? No. Oh, okay. What do you typically run? Now you get an insight from the carrier. Same with the customer. Hey, you guys shipped a pallet of whatever with us last month. You're going to— you want to do it again? That's the excuse that we have. So yeah, so like doing the cold outbound piece.

But are you guys doing cold outbound for trucking companies?

Mostly if they've worked with With the customer.

Well, that was— so I'm saying if they haven't, if, if, if I'm a broker and I'm like, I want you to be my carrier sales tool.

Yes.

Okay, so you can't— so like if I said I want you to call any carrier in Illinois for Carrier 4 that's on Carrier 401 that has ever been to Ohio with inspections because maybe they run Illinois to Ohio, you can call some?

Yeah. Customers feel like, our customers feel a lot more entitled to make that call because in a sense you're giving them potential business. Like if you actually have a load, if it's a candidate load, it gets a little bit more tricky because then you get the carrier angry and it's like, you didn't even have a load, you were just like phishing.

Yeah.

So it's a little more messy. But if it's, if you actually have a load already and you're just like really needing a carrier to do that, like they feel more legitimate How do you put that? It's a bit more of a legitimate call versus if it's just a pure outbound call. This is getting into TCPA regulations for cold outbound to sell services. That's where TCPA regulation gets a little bit more strict on that. And this is where we recommend customers to not do a lot of. But for the carrier side, you're right in that we do do that because you have an You have an offering for work. So it's a little bit— it's very interesting. It's a little bit different than selling services, basically.

Yeah.

No one wants to be sold. People do want to do work.

I get that. It's interesting because I'm a sales guy at heart. I love selling more than— you know, I've spent my whole career in freight brokerage and I've run a brokerage There's no function I enjoy more than reaching out to shippers for new opportunities. And part of it is the competitive nature of it and the fact that there are 30, 40, however many thousands of freight brokers out there doing the same thing. And they're largely undifferentiated, at least in the minds of many shippers, where that cold email, that cold call is generally very unwanted. And something about the chase for me is I just love it. And Specifically the creativity. I lean into creative sales and trying to come up with leveraging as much research as possible into the individual, the company, the market, whatever I can find out about this potential lead to come up with something that I think will be interesting for them. Maybe it's funny for them, maybe it's educational for them, informative, whatever. I enjoy that, um, and I, I— it has been successful for, for me. And I'm just curious, like, I think it would be interesting to leverage AI to help do some of that, um, preemptive work. Not necessarily doing the outreach itself, but saying, okay, you know, Joe Thompson is the VP of Transportation at Kroger, and I want to haul for him, and I want to know everything there is to know about him and his business and Kroger and the market that applies to Kroger. Give me all you got so I can come up with an email or a call that makes sense. Are you guys playing in that arena at all?

We started looking into some of that enriching of the data. I mean, I'm going to say there's a lot of sales-specific tools for that alone, especially on email, and they do good work. We use ourselves some of that for our own outreach, but we have had customers with whom we're already now piloting some access to LinkedIn for the agents, so then the agent can research a little bit on LinkedIn. And the reality is that web search or Browser search for agents is getting so accessible that now you can do those things, which is really interesting. But no, I agree. I think the— honestly, this is where I would still, as a— if I were, like, doing a broker, I would probably, as you feel very invigorated and, like, very excited to, like, be doing those sales, like, it's the hunt. I think that's where it'll be hard for humans to let go of because it's fun, no?

Yeah.

I mean, similar with killer shells, but especially with a customer, you hit the gong, right? I mean, you feel that energy. So I think that'll probably be the last piece where we can help. I would frame it more from an account management perspective, especially smaller accounts where the agent can actually be that warmer, how do you put that? The agent can be Warming up the account a little bit again after it's been dormant for a little bit. We've seen a lot of success on that dormant account type of use cases on the sales piece. Maybe on the cold outreach, I agree. It's really fun to research about that person and be like, oh, do I send them some chocolates to their office and then give them a call right after they get the delivery?

Yeah, stuff like that I think is interesting. Um, so given time, I want to move to kind of some of the market in general, the competitive market and the category itself. Like, this is where all of the VC money is going right now, into, into AI businesses and, you know, trying to find ways to be more efficient and do more with less. Um, our industry itself is, is one of the Most busy and noisy because our industry has always been looked at as the most antiquated, old school, pen and paper kind of business. I'm curious from your perspective, because there are a lot of companies preaching a lot, what are you hearing? What is the biggest BS claim that you're hearing from other AI vendors? That you feel like is misleading or not representative of what companies are actually doing?

That's a good question. I mean, I have to be honest, and we just don't instill any fear of competitors in our company. It's more of a, as we have a customer, let's win it over. Let's crush it with the customer. Let's deliver. And let's focus on that. They might, like, the customer might bring up, hey, we have, like, we're looking at a couple of folks, 3 folks, whatever. Let's see what you got. And I mean, so far we've been very lucky and I think we've nailed, I'm gonna say, like, 99% of the engagements that we've had. You cannot, like, do everything right, but, like, especially as you grow, there's things that fall through the cracks sometimes. and not responding maybe to a customer email that gets them angry. What can you do about that? But in general, with the enterprises that we have, we've been really, really white glove, and that's gotten us that access to their different use cases and ideas and operations. So, I'm going to say that from our perspective, the bullshit that the external world might be thinking that AI is doing, I mean, I've just seen folks let go of a team that was doing super, super manual work and put them doing sales. They went from 82 people to 4 people, okay? From 82 people to 4 people, it was a very mundane, repetitive task they were doing. Everyone just was churning through it. They were using a BPO in this case. But that type of tangible ROI is massive. And when, again, you can actually put the good performers, because who wants to have low performers in their companies? I mean, the good performers—

No.

You're going to promote to maybe customer sales, and they might be your VP next year. So, that's the type of company that you want to run. I remember one of our customers saying, hey, whoever doesn't engage with this bullshit AI, at the time, It's not going to stay here for long. You have to work with it because again, remember, this is what he was saying to his team, if you don't work with AI, someone will and they will win over us. So that's been the biggest mind frame or kind of set of ideas that we've had, like always be very practical in your approach to what we automate first. Because you want to see ROI. So even if I've been talking a lot about the high level, like 3 layers and like coordinating work, ultimately where we start many times is just executing for a certain use case. So honestly, no, not a lot of bullshit that we see around here, man. I mean, it's working. It's crazy. We're busy. Yeah, if you have good engineers, send them over because we need more to build more product and go-to-market people as well.

I'm curious, like, 5 years from now, what do you think this business looks like? What is— what does the impact look like? And what does your team look like? What is— how has the industry changed as a result of your business? Like, give me the— what your vision looks like down the road.

Mm-hmm. 5 years from now, we are working with the largest enterprises in the world where we've deployed an AI workforce that is doing that 90% of the work, having very smart humans doing the strategic and more important part of the work. And the TAM for that, if you will, the market for that, and we're talking about billions of spend in just people being the interface to the flow of data, like sending emails, making phone calls, entering data from a PDF to a database, that's not going to be the case in 5 years or even 2 years from now. Whoever is serving as an interface to the flow of information, is not going to be doing that job for long. Now, whoever is adding value to the company strategically, getting creative, as you said, thinking creatively around how to win more business, be more efficient, that's who you want to keep in your company. I mean, I tell my team of engineers themselves, like, guys, code is getting really easy to write. You better make use of those tools. You're going to be left behind even faster than the people we're helping on our customers, meaning you better learn how to 10x yourself. So, what applies to my engineers in Happy Robot applies to the reps doing all of the keeping the world moving, which is super important, but we're going to have to 10x the efficiency. So that's how I see the world, both the engineering world, but also like the operational world is gonna 10x themselves or 100x. I mean, things are moving so fast.

My concern, and this has nothing to do with— this is more macro. When I think about my last When I think about a brokerage, most of the people are not the 10xers. Most of the people are not the creative thinkers that are like all day coming up with the strategic solutions. Like, most of the people are the ones doing the kind of work that AI should be able to replace in 5 years.

Mm-hmm.

And I don't know that it's just so feasible to turn them all into 10xers, to turn them all into the creative thinkers. And I worry a little bit about where the world will be in 5 years in terms of opportunities when those just aren't the native skill sets of a lot of people. the functions will require. So I don't know if this is even a question, or if maybe it is a question of, do you agree?

I agree.

And what the possible risks are to the world or to the global economy if, I don't know. I mean, you hear some people talk about not even having to work and universal basic income, and that's a whole different conversation, I think, as we're running out of time. But do you have any thoughts As I'm kind of postulating here?

Uh, yeah, I mean, I think first of all, we better learn how to teach our kids to live in this world. We better upgrade our school systems. Right now we're just creating sausage producers, you know what I mean?

Yeah.

It's very like, you do this, you study this, you don't do other things. So we better learn how to today, like, change the way we, we teach our kids. I don't have kids, I don't know how to teach them, but my wife is a teacher, um, and honestly, I still see the same process, the same manufacturing process of perfectly educated kids that actually don't think out of the box. So I think that's a problem. I don't know how to fix it. Um, I think AI might actually help because what I was discussing with my wife is I wish I would have had a more personalized education, meaning I spent a lot of time, and my wife verifies this, just waiting for the teacher to calm the class down. It's like, Peter, don't shout. Matthew, sit down. And you're just like waiting there, like 8 hours of your day. Just waiting for people to shut down and try to learn versus if you could just learn for 2 hours what you want to learn about that day and you get creative, you're curious about that specific topic about the history of— I was talking to my engineers about the history of the Reconquista in Spain yesterday. That's what's going to get kids more creative. And then, they'll be able to have more time to maybe build stuff on their own and learn how to go and find new things to do on their own versus us trying to, like, put them in buckets. So, that's a little bit of how I think about the education system. It should be a lot more personalized, and I think AI will help do that. Now, how do we fix the potential problem that you brought up? I think we, as a society, we generally have overcome these problems very well. The scale is different. I mean, when my grandfather bought a tractor, I mean, he didn't need enough, he didn't need that many people like going to the field, but they found other things to do. So I don't think we're going to have a problem. It's going to be at a different scale, that's true. So it's going to be— we're going to have to adapt faster, but I think something we do well as a As humans, we adapt fast. So, even if someone today is like, you think they're actually not strategic thinkers or they're too narrow, give them a carrot. I mean, give them a carrot and they probably are going to chase that carrot and they'll learn, they'll get creative. I think necessity creates creativity, if you will. I don't know how to put it.

Yeah.

But I mean, for example, in our case, we have not raised— I mean, we have raised big chunks of money, but we have competitors that have raised, I don't know, you look at Sierra, for example, Sierra AI, this Brad Taylor company, it's like valued at $10 billion. They've raised like hundreds of mils. I think, I mean, we're working with some air cargo and airlines and folks that are telling us like, wow, this just works a lot better. So, I think that out of necessity, because we didn't have enough engineers, we've built better products than if someone just has infinite resources, similar with the government. Government has infinite resources and it doesn't work. Another—

That's a great comparison. Always compare your competitors to the government and you'll sound like you're making a home run point.

I brought up Sarah because it's also out of the spectrum of freight, if you will, but I mean, I don't have anything against them.

No, no, I appreciate what you're saying. I think you're making a very cogent point and I agree with it. Is that like, just because I can't see the solution today doesn't mean that when push comes to shove and we find ourselves in need of creating a solution that we won't find one.

I think so. I'm very positive about our societies. I think we can always overcome things even if If, if we mess up in other, in other points. But I'm positive about finding solutions.

Yeah, I'm with you. Well, listen, man, I've had you for over 90 minutes and I promised I wouldn't go over that. So, um, this has been a lot of fun and I feel like I didn't even get to a lot of things that I could ask you about. But, uh, maybe we'll do this again sometime.

Let's do that. Really appreciate it so much for having me.

To our listeners, I hope you enjoyed the, uh, enjoyed the ride. Welcome back. We are back officially with our first episode back in a few months and What a great guest we had. Thank you so much, Pablo. Um, to everyone else, we'll see you next time.

Thank you.