Troy Lester

Welcome back to another episode of Logistics Remix. Today's guest is a founder who's reimagining how freight moves through the middle mile. Troy Lester is a co-founder and chief revenue officer at Warp, a logistics company that's raised 22 million and built a tech powered network with 1500 plus LTL lanes, over 50 cross facilities, and a fleet of more than 7,000 carriers. Warp already powers freight for retailers like Walmart, Sachs, and Go Puff, helping them cut costs, gain transparency, and improve service. Now, before starting Warp, Troy built and scaled multiple logistics ventures, including Jitsu and Covet shipping, and has been deep in solving middle mile pool distribution and parcel zone skipping challenges. Troy, great to have you on the show. Thanks for having me on. excited to talk about all things middle mile. Awesome. Me too. Uh look forward to this conversation. Let's get right into it. The where I wanted to start first with you is kind of um take you back in time and and learn more about how you got even got into this space like how did you first get into delivery and logistics? Yeah, started when I was 18. you freshman year of college decided to create a food delivery company where I was reusing meal kit point uh or meal points that you had uh for cafeteria food and I I wasn't eating the cafeteria food so I ended up delivering to a bunch of kids on campus using my meal points and that's when I got first bit by the delivery bug I like to say uh wasn't a good business but started to help me understand what was important about last mile delivery then went on to start more food delivery companies, grocery delivery, and then eventually got into more of up the supply chain into micro fulfillment and then final mile e-commerce delivery. Wow. So, all started with with food and delivery and um and and so what what made you fall in love with it? Yeah, just from the the idea that you know people had such high expectations around delivery even when I started back like 2015 the the precision that was required to get something delivered within 15 minutes 30 minutes where food had to be hot right and customers had to be notified in advance right all all the pressure points that are involved in a successful delivery that really got me excited gave mean an adrenaline rush for my first delivery. Yeah, it's like playing a game every every day. You get to go out there, you get to play the game and you can lose, you can win. Mostly want to win and and in delivering within that time frame is is winning for your customers. Uh I agree. I think there's a lot of really cool challenges across the entire space and I feel like once you get in, at least for me the experience once you get in like there's so many challenges to solve, so many problems to solve and and you just want to just want to continue building. Um, but you did you you actually built several companies. Um, I want to talk about that. Let's go back to those early days. Um, and I want to uh what I wanted to ask you is um since you helped build Jitsu uh and Covet Shipping uh those early ventures of yours before starting Warp like what did you learn from building uh two new ventures in the logistics space? Yeah, I think the the biggest learning lessons were around kind of those those lessons that I learned from early food delivery, right? how important customer obsession was and how little of it I really saw as I crawled up the supply chain into you know more complex logistics solutions that were prepared for enterprises less so delivering to consumers right so there were instantly right off the bat even with covet shipping which is a micr fulfillment company you know a little before its time the bar just in logistics was so low right compared to what you would see in like food delivery logistics. And I thought that was just the greatest opportunity. Like when when I saw that I was like, "Oh my god, wow. The standard is is really that low where people aren't getting back to you on time." I think like early on the microfilming company, one of our biggest value props was that we offered Slack, right? So that you could get instant communications, right? Now more of a standard, right? But they were like, "Yeah, we can't get our reps on the phone." And I was like, "Oh my god, I I didn't even realize this was that bad." And as I got into parcel delivery at Jitsu, I saw it was even worse. Right? Like FedEx, we were competing against like FedEx and UPS predominantly. We were a gig economy last mile fleet. When I joined, we had about four sort centers just on the West Coast. They acquired the micro fulfillment company. I go, I become our head of partnerships and business development. I'm tasked with trying to build out this middle mile infrastructure that connects our sortation centers across the country and then gets our sortation centers connected down to the delivery station level. Right? So, we needed box trucks with liftgates. we were dropping off and parking lots and you know uh these kind of storage rental unit areas where we'd have box trucks with liftgates dropping off you know 12 palletized routes for local last mile drivers to pick up and deliver locally and then as we started really scaling we were like oh my god we're now in 15 sort centers across the US brands that are wanting to use our great last mile service and this was the same thing that was happening with other regionals They couldn't justify actually getting either a full truckload to us or any other LTL service. They couldn't get that to us because the service levels would not meet our sort times for our for our parcel company, right? And so I realized very quickly, oh my god, we need to be able to give like per pallet rates that you would get from LTL, right? For all those folks that couldn't fill up a truck 7 days a week, 5 days a week. And then you also need the dedicated service levels that come from truckload. But if you couldn't fill up the truckload, why would you pay so much more per pallet? Your middle mile cost per box would skyrocket. Why would you pay so much more just to use Jitsu Axelire or or VHO or any one of these other great last mile providers? The math just didn't make sense. It didn't matter how much cheaper the last mile cost per box was. We were just losing at the middle mile. And that's when we realized that Jitsu, wait a second, kind of the whole key of last mile at scale is really that middle mile component. The what is your infrastructure? What is your network to get down to the delivery station levels efficiently and cost-effectively? And then really what's that connection piece? What's that trucking glue point, that nexus point that allows you to enable brands to get into your network easily? Totally. And you it's it's kind of like you you peel back the onion, right? You started with one layer of a problem and you solve that and and what's interest I mean what is interesting about what you you mentioned about just getting back to people, getting back to your customer or where where a lot of requests are where's my shipment or what's going on here. You went from email to Slack. Now you're getting back to them faster and that at the time was a value ad. Unbelievable, right? Like how times fly. Uh but you peel back the onion and you just dis keep discovering another problem to solve, right? It's like, hey, we've solved this. Uh oh, now we've got to step one back, you know, one step back and solve another issue and then step another step back and solve another issue. Uh and that kind of leads me, it's it's interesting. It kind of leads me to this uh to ask you like how did you get the idea for Warp? Because it it sounds like maybe it came from that time, right? from some of your learnings there, but curious to hear how you got that idea with your co-founder Daniel and kind of but as you were starting Warp or or or just said like hey let's let's let's get this thing going um like what were you trying to solve uh at the very start I know you've evolved we'll talk about that but just curious about those like very early like ideation days yeah so I I think going back to just the micro fulfillment days right My earliest micro fulfillment centers were basically underground basements and shipping containers on surface level parking lots and I needed box trucks with liftgates to deliver. And this was just a nightmare to procure and and get reliable service, right? And it was that same thing that was happening when we were setting up the delivery station to sort center connections for Jitsu, right? And so we saw these kind of middle mile pain points for years. that was seeing these pain points and it got very obvious that LTL couldn't solve it. I tested all the LTL carriers. Like literally, there wasn't any LTL carrier. I didn't test for these problems. There wasn't even the the new tech enabled ones. I I tried everybody and nobody was able to give us, you know, exactly what we needed. And I'd say the confluence of seeing all that when we were trying to connect networks, right, and then seeing so much demand while we were at Jitsu for B2B freight. So think like store replenishments, right? Think like big and bulky, you know, moving around, big and bulky freight moving around in your network like inventory transfers. This stuff was moving via these parcel carriers networks, right? They were benefiting off the sort center efficiency and the middle mile trucking efficiency like a FedEx and UPS was for years, right? Like there was some stat UPS 42% of the volume that they're delivering last mile. It's B2B, right? There's still a bunch of these last mile providers. I remember like lasership, you know, on track, they're delivering so much payroll, you know, like checks are still going out. Like people don't even realize how much density they're getting from like mail and and payroll and stuff like that. So when we were thinking about, okay, how do we make jitsu better? How do we make jitsu better? We knew it had to get into this B2B world. We knew it had to get more drop density, right? How how do you compete with a delivery per hour metric where Amazon, FedEx, and UPS have all this this B2B or or Amazon just massive e-commerce volume. Nobody's got the same e-commerce volume D to C that those guys do, right? And so it became super obvious that the bigger problem to solve was actually like this this middle mile piece. So the earliest ideiation was okay, hey, let's see if we could just resell cargo vans and box trucks, right? That initial micr fulfillment problem, box trucks with liftgates, cargo vans. People need access to these smaller vehicle types. They'll use them for agile store replenishments and they'll use them for small carrier injections, right? Going into, you know, regionalized last mile providers, right? And so those were our first 10 customers. And then naturally, we were like, okay, wait a second, layering in the cross stocks, we know how to build sortation tech for pallets and for parcels better than anybody on the planet, I would say. let's let's actually build a network where you could consolidate freight. And so we added in the idea that you could pick up with 53s and drop trailers. We added in the idea that you could have freight come to our cross stocks. You could slow it down to get a more consolidated freight rate, right? And so that was from customer 10 to 50. And that's where we really started to take off. So ultimately you you basically found um it sounds like you found that you needed density in the middle mile to make it really economical to make it really work. B TOC was not enough. B2B was going on its own. What if you combine both and now you you can fill out that 53 ft trailer and utilize that equipment even more. Exactly. Like we were we were realizing, okay, we could help better enable more last mile equations for retailers and brands that are trying to access new and better last mile services and then also massively enhance their B2B freight and that's really our sweet spot. Yeah. So you're basic you're taking two and you're combining them into into one solution to to fill out the truck. I mean I'm I'm just simplifying it and we'll get into some of the technology you guys have built which I'm I think is really impressive uh that that you were able to build in such a short time frame. Um but I want to talk to you about something that you guys call flow skip. So I read you have this solution uh flow skip uh which is really it's taking kind of the I think it was what we're talking about here. It's taking the best of zone skipping and pull distribution and just combining into one solution. uh for for the for the retailer for the business and um you I just wanted uh you to share some more about it like what is it why is it beneficial to shippers why why are they using it is really the same exact concept when you're trying to build a freight parcel network just applied to a brand's retailer specific pickup truck right so there was this old notion that used to exist in early days in parcel and freight It's called this dock sweep, right? Where a provider would come in and say, "Hey, we're just going to pick up everything that you have on your dock. We're going to sweep your dock clean. We're going to pick up your parcel. We're going to pick up your big and bulky that needs to be delivered to consumers. We're going to pick up your B2B freight that needs to be delivered to your stores, your wholesale partners, right? We're just going to pick up it all and make sure it gets where it needs to go." Right? And then slowly over time you started seeing the unbundling of that. Folks like XPO, GXO, RXO now, right, realized, wait a second, there's more shareholder value. If you split these companies up, people want to be able to invest in these, you know, less diversified assets, more concentrated, you know, niches, right? And you're seeing that same conversation happen with FedEx LTL and everything that happened with UPS and UPS freight, right? And we actually believe it's just created so much more doc congestion for every brand and retailer. These docs are a mess. It's jacking up detention times and detention rates all over the place and it's really slowing down the flow of inventory. So that's that's where the name flow skip comes from, right? We think we could better enable the flow of inventory. We've done this with tons of brands and retailers already. It's it's one of our hottest solutions where we can pick up from a brand's fulfillment center their zone skip parcel volume, their store replenishment volume, their big and bulky last mile volume and their LTL volume. Okay, we could throw that all in one truck, one pickup, one dock door, take it to a Warp cross stockck in that end destination market where you're trying to really zone skip your DSC parcels, drop off at that Warp cross stock all the B2B freight, the big and bulky last mile freight, right? The store plans, the LTL, and then hightail it over to FedEx, UPS, VHO, Jitsu. pick your last mile carrier so that you're getting the benefits of that zone skip for your DTOC parcels and then you're also getting a faster and high qual higher quality B2B freight delivery from that Warp cross stock where we're outbounding big and bulky last mile. We do a lot of big and bulky last mile ourselves and then all that LTL that we're doing ourselves in our network too. Yeah. And with the way you say it, it it it's it I'm connecting the dots and it seems to make sense, but it does sound so counterintuitive to the status quo because if you have your parcel volume, it gets picked up by provider A. If you've got your LTL, which is less than truckload volume pallets, it's get it gets picked up by provider B. And then if you've got big and bulk, provider C. And if you've got store punishments, it's provider D. And so you've got like four, let's say, call it three, four providers coming in and picking up. That's why I say it sounds counterintuitive, but the way you describe it, it actually makes sense because you're able to full more fully utilize the middle mile. Exactly. And like, you know, as as you're mentioning all the different providers you need, I'm just thinking about, oh my god, how do you check all these invoices all the time to make sure they're right, right? Everybody wants one consolidated invoice at the end of the day. That's like another nightmare that we've heard we've been able to alleviate. Right. So, it's worked really well for a lot of these brands and retailers that that we've started this with. And, you know, I'm I'm excited to see more and more folks try it out with different, you know, permutations, right? Store replenishments plus zone skip parcels, uh, zone skip parcels plus big and bulky last mile. There's a bunch of different ways that you can mix and match. Yeah, you can mix and match is what is what uh it sounds like you could do. Now, cross dock is is a really important function for this, right? Because at the end of the day, yeah, you could do the dock sweep and I love that term. you know, come in, do a docu, pick it up, utilize the full middle mile, uh, to your advantage, take it to the destination market. And but the the cross dock is is the next stop. And so, uh, let me ask you this. So, as you guys are building this out and when it came to the cross docking, like was there a network you could just plug into uh on day one as like one cross crosscking incorrup? Yeah, there were we got very lucky. So there were companies that tried to stitch together cross stock networks before us and it's almost like you know when trucking got digitized there were companies like Convoy that came in and really pushed the idea that you needed to use an app to get business right like if we like every single one of our carriers, every single one of our drivers is using our driver app. It has live tracking 24/7. I don't know if it would have been as easy for us to do that if folks like Convoy didn't come before us. Although they failed, they pushed this digitization, right, that really needed to happen. The same thing happened with all cross stocks across America where other people came in and said, "Hey guys, you need to use tech to get business." And a lot of these people didn't have comprehensive solutions that they could sell a brand and retailer where transportation plus warehousing is really what folks need. If you're making it so that people need to talk to you know you as you know crosstocking you know inc that has all these cross stocks and then they need to piece together the transportation and they need to figure out you know these touch points. It was too messy. We tried doing it at Jitsu also. We tried using third party sortation. It didn't work out. We use those learning lessons to to create the technology needed to create our thirdparty cross stock network from scratch. Right? So we onboarded hand-to-hand combat. I was on the ground for the first 10 cross talks, right? I was shaking hands with the owners. Dan and I were shaking hands with the owners. We were giving the app tutorials to the labor on the ground to the management shifts, right? And so there was a lot of trial and error and scan compliance in year one was like not pretty, right? Then year two and year three, we really perfected it. Now it's 100% scan compliance across our network, which 55 cross stocks was not easy to do, but we were able to make that happen. And and then really understanding what the specializations per cross stock was, right? Cuz everybody in logistics wants to say they could do everything, right? It's just natural, right? You know, we've heard it a bunch. We just really had to figure out, okay, these providers are great for parcel sort and pallet sort. These providers are great for just parcel sort, right? These folks, you know, just in and out typical cross stockcking. These folks, okay, storage of more than 30 days, right? So really figuring out and feeding that information into our routing algorithm so that we could dynamically decide what freights routed to these cross stocks. That also took some perfecting. And that's why like year four like our network is actually extremely usable for any mixand match scenario and it's more algorithmically driven than a human making the decisions. Yeah, total it totally makes sense. U and look not easy to uh to build something like this from scratch. You have to be on site to train everyone up and also enable it with technology. But if you had to travel, right, and be on site to train them up and you're launching, I mean, you talk about three years and you launched 55 plus cross docks, like how did how did you decide on that balance for for growing the network? Like was it better for you to expand geographically or did you like deepen certain lanes? Like how did you decide where to be because you can't be everywhere at once? We spent years then and I fantasizing about how we would do this. Okay. So, we always knew, okay, you start local, right? The second you start taking national volume, right? And we learned this the hard way early in our career. Second you start taking national volume, unit economics and service are going to go down, right? And you're going to end up having to traditionally everybody's throwing bodies, right? Okay, add another ops manager, add another dispatcher, right? We knew we were not going to allow that to happen. So we built each part of our network is built around a local cross stock that we knew could handle good volume and there was like a good 10 that allowed us to start going national where we just built local pickup and delivery density. Like year one and year two we throttled growth. We did not grow as fast as we could have. Right? We slowed onboardings. We've been slowing onboardings more and more as we've scaled actually. Right? because the enterprises are asking for more and more complex complexity. There's more deeper technology integrations than than what we were seeing year 1 and year two. And so building that local pickup and delivery density and really selling into smaller shippers, right, that that also had these kind of local regionalized needs allowed us to build really good cost structures and service levels in each major metro, right? these like 10 pinpoints on a map. And then we started adding enterprise volume that allowed us to connect the dots from like LA to New Jersey was the first dot we connected. Fast forward now we have 1500 LTL lanes that are running 5 days a week. Almost every single lane's running 5 days a week now. And service levels are tight. You know, performance is where it needs to be. Time and transit's reliable. But that couldn't have happened if we didn't build the cake properly first and get that foundation of local pickup and delivery density with our cargo vans and our box trucks and our 53s around those those cross docks and perfecting our sort and seg. And so as you're adding more cross docks, are you having to kind of replay that same formula or like like how how are you still doing the things you did in day one which really I mean don't really scale because you traveling to the site and training everyone up or or have you kind of figured out a formula of like hey we can double this uh because we've got uh this uh repeat uh turnkey solution. What's good is the cross stocks we onboarded early on, while they weren't like these massive companies that had, you know, five or six facilities and this massive footprint, they weren't like your tiny mom and pop that had like 10,000 square feet and, you know, a retiring owner. It wasn't that. It was somebody that was ambitious, right? It took us time to vet these folks. Somebody that's ambitious, excited to grow with us, right? So we've tried to not onboard more cross stocks over the years, right? So like you heard the numbers at 55. It's now more about stuffing them more and adding in more of a robotics and automation to make them more efficient than it is adding new providers. There's a very crazy wait list um to get into our cross stock network, right? Everybody's trying to fill their warehouses more. you know, since COVID, there's been this massive warehouse glut, right? You know, demand really hasn't come back, especially with all this warehousing that we built up, all the new warehouses, all new people that got into warehousing, cross stockcking. So, yeah, we're really focused on feeding, you know, the folks that have been in our network the longest. We're we're big believers in in loyalty in that sense. All right. So, you mentioned something interesting. you mentioned uh some automation and and uh um technology in the cross stack. I want to go there a little bit. Um actually I want to uh let's let's start kind of um how that's how Warp is building software that's enabling your network. Like I'm curious like did you go off the shelf? Did you build it internally? like tell me more about like your software journey because it does not sound like there's a tool that was going to do what you're doing uh off the shelf but I might be wrong so you you tell me. Yeah. No tool you know to just plug and play unfortunately right so we were very lucky uh the same CTO that we had at Jitsu before uh Daniel myself and that CTO we've been all working together for the last seven years now. So we were able to come together here and we we knew exactly what we wanted to build. We have been through so many experiments together, right? So we knew right off the bat what we needed and how to build it all inhouse, right? So we said, okay, we know we need a driver app, right? We know we need scan in scan out functionality. We know we need our system to understand peace level freight, not just pallet level. Right? This was a big mistake of folks that went into, you know, tech enabled freight, right? They really just understood freight at the pallet or truck level, not the piece and the skew. And we know, we made sure our system like understood skew, piece, pallet to truck, right? And and that those building blocks in our core system, our core TMS, right? Those were massively necessary for us to consolidate freight in unique ways and to give data visibility to brands and retailers that they never got from like a transportation provider before. Right? So our TMS is a mix of it connects to that driver app. It connects to a cross stock app that all our cross stock facilities are using for scan in scan out and then we have an IMS built into that TMS. Okay. And it's a WMS built in there, too. It's like really custom software that has everything that you'd want to manage multiple warehouses, multiple different vehicle sizes across transportation providers and multiple different types of inventory, right? And so then, you know, we gave customers a simple like customer dashboard so that they could see these things. Where's their inventory moving? How is their transportation network functioning? Right? you know, do you want to swap vehicle sizes? You see, maybe there's, you know, uh, lanes that you could optimize more, right, within this customer dashboard. And that's something that we're we're building out more and more, but we really had to make sure every single part of the chain, right? So, cross stock operator, driver, right? uh people internally like our our dispatch team which has gotten you know a lot smaller over the the last year from from AI and we could talk more about that but those even supervisor agents that we have now all the functionality that we built for our system to understand that was necessary cuz we wouldn't even be able to use AI properly without those building blocks of our our main TMS understanding these components right so all of those components work in and of themsel elves, but that plus this massive routing algorithm that takes into not just uh takes into account not just vehicle size but capacity at a cross stock, right? And you know what the SLAs's are from a price and speed perspective, right? It's how we price all of our proposals. We're throwing it in a routing algorithm. It's not just some guy in a spreadsheet all day, right? So all these pieces work together to create this Warp experience and so and so you sounds like there's a lot of software was built internally for the internal operation. What about the external uh customer view? Is is the customer getting value uh outside of this process just being more automated by your software for internal operations? The biggest thing I think I've learned actually in all of logistics tech is front-endf facing tech is very overrated. the the main thing transportation decision makers need is the cost and service level enhancement that you know you just can't get without the technology right so our philosophy on the tech is really we call it McDonald's level tech right where McDonald's can go grab somebody off the street tomorrow and make you a pretty perfect quarter pounder with fries exactly as if it was the same person that was running that fry station for 10 years, right? And so that and that that McDonald's level tech philosophy allowed us to plug in AI and make our operations so much more efficient, which allows us to unlock this lower cost structure, which enables our customers to get lower prices. That's our flywheel. I don't have the fanciest dashboard in all of logistics tech. Okay? Somebody else can give you that. I don't have the fanciest bells and whistles, but I have transportation solutions that you can't create anywhere else. Right? If you're trying to build your freight network, this is the place you need to come and do it, right? It's it's not going to happen in, you know, the fanciest of dashboards, you know, but it gets done and it enables you not to need as many track and trace people, not to need as many dispatchers cuz all of those mini functions of where is my order and needing to check tracking links and needing to see if SLA performance is measuring up. You don't need the weekly calls if you're getting the reporting accurately and you trust the data, right? So we eliminate all those touch points and make your operations a lot more efficient. Yeah. And and I also I I would kind of have that analogy of like Stripe like you don't need to connect to American Express or Visa or Discover Card or or or whatever other payment form and hey here's your API. We make it super simple and straightforward and your c you're just going to know you're going to get paid and that money is going to hit your account and and for the customer it's a transa they they can transact however they want to. Um, and it's all done. It's, you know, not the best, not the most formal like dashboards and pretty dashboards and here's where where everything stands and the charts that go up and down, but it's like, hey, our network's going to work and it's going to be seamless. Um, I want go back to like what you mentioned about uh you and your co-founder and your CTO working previously together. Was there a benefit there? Do you think do you think that helped you move faster um to get through you know to to innovate faster uh because you had prior work uh a prior working relationship? I mean maybe personal relationship like did did that help you? Oh absolutely. I I think you know Daniel and I have been speaking every single day since the first day we met. Uh you know it's like six, seven years ago already. and the CTO, we we all have that same philosophy on technology, right? Typically, somebody that's on the commercial side in one of these logistics tech companies is asking for bells and whistles that maybe one customer wants, right? And you know, they or or maybe they don't even have customer demand and they're like, "Hey, we need this." Right? That happens all over the place. And then two years down you end up having this tech debt. You have all these toys around. Nothing's actually tied to any revenue. Nothing's actually tied to any operational efficiency enhancement, right? Which could lead to cost savings, which could lower your prices, right? And if we if we all been focused on building the McDonald's level tech to just get with the aim of getting prices lower like it's very simple, you know, people are forever. It's a commodity business, right? Logistics is a commodity, right? People will always want to see their rates go down, right? Somehow, you know, we've gotten in a way where there hasn't been enough technology in this industry where rate increases are the standard and people want to talk about inflation and the cost of goods and, you know, fuel, search charges, fuel is going up, so we need to charge more. Wait a second. Technology is the most deflationary force on planet Earth. It's proven, right? It's out there. We all know that. We should be able to inject enough technology into our operational model. Maybe it's not a fancy dashboard for a customer right now, but we should be able to get that tech into our operational model. And the output should be a lower rate for you so that you want to keep working with us forever. And our retention is massive. It's not just based on a relationship of somebody that golfs with some guy and that when they retire or when they, you know, leave their company, the account and them travel with them, right? Like that I think will all end. And I think that's why us three building together for so long. I think it's I haven't seen it. I think it's rare to see that level of alignment from a strategic technology building and go to market perspective in our industry. That's that's uh that's really interesting and um especially especially if if you can speak the same language and get through barriers faster and just get through agreements from a disagreement to agreement to alignment uh faster. It's all about the uh it's all about moving faster these days. We know that. Um I I wanted to you mentioned uh you mentioned automation and AI several times. So, I wanted to um hear more about like specific use cases and how you're actually using AI because it could be so many things to so many people, but how is Warp using AI? And I think you're using robotics as well. So, how are you using both AI and robotics today? Yeah, we'll talk about AI for for a second here. So, I'd still tell you AI is overhyped in its application in logistics today. I'll still tell you that. I think AI in general is just forcing a lot of efficiency uncomfortable efficiency conversations around hey wait a second do we need this much headcount right I is is this maximum efficiency per head right so like we'll start there and so that's where we started we didn't jump into oh my god let's get every customer service ticket answered by AI even though now 40% of our customer service tickets are answered by AI. We just built our own LLM for it, trained it ourselves. It took us a little longer than everybody else that was coming out saying it, right? That's the easiest application. How do you make sure the customer is actually happy with it? I've seen people come out with the metric, you know, hey, you know, we have a certain percentage of our our tickets answered by AI. What's the NPS on that though? I haven't seen that. Right. We made sure we measured the NPS, right? What's the satisfaction level compared to, you know, an overseer? What is Can you uh uh define NPS for us? Yeah, just like your net promoter score. How happy? Ours is a simple, you know, one to five stars, right? You know, we put that in from right off the bat cuz like getting responses off of people you off of from AI for these people is easy. That was the easy part. How do you make sure it's quality? How do you make sure like it's not pissing people off, right? because when we were testing it, it definitely, you know, pissed some people off, you know, right? You have to really fine-tune what your LLM is spitting out to these folks, right? So, so that we've done. And then where we've had a lot of fun is those major operational functions. So, think like track and trace, Harrier sales, and like dispatch, right? Those functions, there's many agents that handle micro tasks within there, right? So, think about like a bidding process for carrier sales. Okay, you want to send out uh a bid for a lane, right? Let's say it's LA to Texas. We send it out to a thousand different carriers that qualify based on the parameters that are uploaded in our system for vehicle type and geographies that they cover, right? And you know, newness of vehicle equipment. There's all these things even to be a carrier in our network. There's all these blockers like you have to have new equipment, right? There's like 10 different checks that go into becoming a carrier, right? So sending out that bid, it goes to specific pools. The pools used to be manually created. Now AI is determining what carrier belongs in what pool. Right? That was pretty scary for me to see. The first time I saw that, I was like, whoa. Right? Always thought the cultivation of the pool was like a unique role that would exist for a while, not needed anymore. Right? So that's one of those micro functions gone. Okay. The next micro function, the negotiation within carrier sales and procurement. We have AI negotiating back and forth with carriers on numbers. That doesn't need to be a human, right? Human incentive. We don't need a carrier sales commission plan. I don't have to worry when I'm giving freight to a customer when we're we're selling freight. I don't have to consider the part of the pie that was a carrier sales commission number. That's a nice chunk of the pie that's gone. Saves money. Yeah. I could give them a lower rate from that thanks to AI. Right. So there's all those micro processes involved in track and trace and finding out, okay, hey, this person instead of our first version was for track and trace making the track and trace rep more efficient. Hey, this person's off their route, right? they're going to be late for this pickup. And then the next level was, hey, they're going to be late for their next pickup that they have, right? And then level three, which is where AI came in, was, oh, hey, AI is just going to be able to notice all these things and send the messages themselves. You don't need a human in the loop, right? The only way we were able to get the human in the loop though and people that have really like played and built with this understand this, we had to create supervisor agents because agents do still misbehave. You know, these like micro functions still can misbehave. You needed like process check agents at each step that understood an overall SLA that was trying to be met across not just one load but dozens of loads and on customer specific agent supervisors we had to do that too right so the AI applications have been a lot of fun I don't hear a lot of people actually getting into the nuance of that I wish I did more that's why I say I still think it's overhyped um but That's that's our the bulk of where our AI work has gone. Yeah, it sounds it sounds like it's gone into areas and specific tasks there that are pretty manual. I mean, I remember myself, I started at CR, you know, CH Robinson 20 years ago, you know, training class was learning how to book a a load, learning how to book a carrier on a load. And that took phone calls, that took negotiation, right? And it was all done over the phone, very manual, not even email at the time. And now, um, a lot of that with AI and Agentic AI is is being automated. And it sounds like you guys are at the forefront of that automation and and leveraging AI for those specific tasks. Absolutely. Um, let's talk let's talk about robotics. I I I I want to hear more about robotics and what you're doing there cuz cuz you're not just in that information. I mean, you know, finding the right carrier pull and and and negotiating with them. A lot of that is just information going back and forth and communication now, but now the freight gets to the cross dock and and you've got to deal with and you've got to make sure it goes out on on the right um on the right provider for the for the right destination. So what are you doing with robotics and and um want to hear more about that. Yeah and it it it kind of goes back to our philosophy around just the AI and robotic deployment and supply chain overall. We believed for a very long time that there are too many like burnout jobs we call them in supply chain right from dispatch dispatch functions to track and trace to a lot of menial tasks that are going on in a warehouse right I I've been in you know over 100 warehouses in my life you've been in hundreds of warehouse like we've been in these warehouses you talk to these guys working there right and I'd always ask them like hey you know would love to learn about your family right we talking about these things and you know the ones that had kids say, "Well, oh, what do you want your kid to be when you when you grow up?" Nine times out of 10, I get an answer. Not working in here, Troy. Not in a warehouse. As long as it's not in a warehouse. Okay. So, with that, that like core philosophy is what drives all this automation, all this robotics, all this AI use, right? It comes from a place of this future should be where supply chains are really getting closer and closer to running themselves without needing so many humans, you know, missing anniversary dinners, missing kids soccer practices. Like this has been a crazy industry for way too many people, right? And it's just not talked about enough. So, so that's where we come from when we're trying to add technology into this stack. The robotics piece really exciting. So, we have concept sites. We call them concept sites where we've deployed robotics across the US into into warehouses and cross docks that that we control. And the most exciting part for me is it's a lot safer, okay, than what we would normally have in a warehouse, right? Like people that are showing up for warehouse shifts, a lot of times it's temp labor. They're running three shifts. You know, we don't talk about it enough. There's tons of forklift accidents and accidents in a warehouse all the time. And I, you know, there was so much technology invested in the safety, but it's not good enough, you know, in in our perspective, from our perspective. And so we were able to create an automated forklift, really exciting one, right? Where we could unload a truck and get it into its proper staged position in the facility with the use of computer vision cameras, okay, that we have a fixed at the top of these concept sites. And between those two things, we've eliminated multiple labor shifts. And there's still humans working alongside. These are not fully robotic facilities. The vision is to have all of these facilities in our networks be robotic and have minimum very minimal, you know, human involvement. But just those two pieces alone where computer vision is mapping out the facility dynamically and understanding what's going on with that facility without us needing to do WMS integrations with all of our cross stock partners, right? which would be hairy and some of these you know providers that are out there don't even have like a WMS you can integrate with right so we were able to get such a level of visibility and direct exactly where that automated forklift should drop off that pallet and where it should take it when it's going outbound for staging. Wow. It you know what it reminds me of this garage in and parking garage in my neighborhood where if I want to go watch a movie everyone's could go park in that garage and uh if I want to go to the movie theater and uh it used to be you pull up you you know you you press a button you get a ticket and the gate uh would lift up and you drive through find your spot and the way out you got to get that ticket validated you got to you know put in your credit card make the payment and put you know it's just all these steps and Now, there's this company that's basically automated it with with Vision. So, I drive up, it's already it's taking a picture of my license plate. It opens up the gate. I drive right through on the way out. Like, I I validate with my phone app and then and then when I drive out, I don't have to make a payment. It already has my credit card saved to the account. So, it's just like it saves like two three minutes, but those are like really frustrating minutes. And uh it makes the whole process seamless. It sounds like you guys are doing that for the cross dock with the cameras and uh auton. Are these autonomous? Are these autonomous forklifts or is there like someone like like uh like at a Saudi PlayStation like playing a game and like driving these around? When we started to test it out, we did have the the Sony PlayStation. Um it was a lot of fun. I had a lot of fun with it. Uh Daniel and I were playing with it for weeks. Uh you know, our girlfriends were like, "Hey, what are you guys doing?" "Oh, we're at the warehouse, you know, playing playing with the forklifts again." But, you know, we were able to remove the Sony PlayStation part, right? And that that was a massive step. So, that that's already months ago, right? And now, you know, as as we scale out the network, you know, and and build more into the existing cross stocks, we're going to be plugging more and more of the robotics and the computer vision kits into them. That's what's really exciting. get the whole network to be more optimized so that we could lower our cost structure at those cross stocks too and pass on those those lower costs in the form of lower prices to our customers. So it's really like a softwaredriven cross dock. It's not software in the cross dock. The software is driving the cross dock. Yep. All right. Well, I can't wait for my first visit. I got to come out and actually see one of these. Maybe we could do like a a live in person podcast where I can actually see one of these and experience it because I it sounds uh very futuristic. Um especially this concept site that you have now. All right. You you've sh you've already shared some futuristic stuff, but I'm sure there's more. Uh if if you guys have gotten this far, there's got to be something on the road map. I don't know what you're open to telling us on on this podcast, but you know, I'm curious if you could share any any near-term technology launches, anything exciting that's that's coming up. How are you going to use AI? How are you going to implement robotics? Like like it can't stop here. What What's next? Yeah. So, the McDonald's level tech is that piece that we're constantly perfecting. The that piece has been perfected the entire process through, right? All the way through this four years. Now we get to play with the bells and whistles a little bit more, right? So for the folks that want fancier reporting, right? And to be able to manipulate their network more in like a drag and drop and I'm not going to go too you you could see this too and maybe we could show this back on a podcast in the future as well, but to be able to actually manipulate that network and dynamically play with things without needing to call us, call a carrier, right? Cuz right now there's still that human-toum interaction with our client success team like, "Hey, we want to strategize together, right? We want you to have somebody come to our site and really give us almost like an industrial engineer level audit, right? Because we're really good at that. We have experts, right? Our client success team, we have former transportation decision makers that are, you know, on our client success team and then they're getting, you know, one of us in a lot of instances. And we've just seen so many of these different networks. We could provide really good consultative experiences. But what if you could upload that full consultative experience into an LLM and into a front-endf facing dashboard experience, right? That nobody has been able to do yet. Very difficult because we're talking for complex, you know, the largest networks in the world that we service, right? So, it might take some time to get in there. They'll still be our consultative arm, but we're going to try it with like more medium-sized businesses first. Give them the ability to generate their own proposals, right? And their own strategic analyses, right, that don't involve speaking to us. And that I haven't seen anybody in logistics nail this. I haven't really seen too many people try. I'd love to see more and more folks do this because, you know, another massive cost center in these business models, right? We don't have the cost that is the warehouses. It's really the people, right? So, if you optimize the people cost center, right? You just get to be able to give lower and lower prices and your flywheel spins. Customers, again, this is a commodity business whether we like to say it or not. The flywheel spins, the lower your prices, the more customers will give you and the more customers will come to you, right? And so that's that's if we could get that less human involvement to that piece, that's the next step when we look at, you know, futurizing our business. Yeah. And it sounds like it's also going to speed things up like like if I if I want something done, let's say I need a graphic design, I'm going to have to send it off to a graphic designer and and and get them to turn it around, but they it might be like, you know, midnight when I send in that request. they won't pick it up until 8:00 a.m. But if I'm, you know, talking to like one of these large language models, I can get that done within 2 minutes at any time. Uh, so that could that's really interesting. I mean, it's uh uh that's really an uh that's really exciting. So basically selfs serve uh to on the solution rather than you know trying to and it might you might still need a human in the loop especially initially right to just validate things but self-s serve on on a network solution rather than uh take submitting that email waiting for the next day etc and waiting for a solution to be sent to you. What what else are you guys working on? And we already rolled out, by the way, like a self-s served quoting functionality where you could get cargo vans, box trucks, 53s, and LTL. I know no other spot where you could get such comprehensive transportation solutions at an instant click anywhere online. Right? So, like that's rolled out and there's no human involvement whatsoever, right? There's you might say, "Hey, to validate there's an automatic credit check in there. It's prepaid and you could submit for credit. You want to get a line of credit with us cuz you think you're going to be shipping a lot? Sure. All automated. That used to be a manual touch with, you know, a finance department. No more. Right. So, the next pieces that we layer in are more of the comprehensive network proposal creation instead of just individual quotes, right? And adding in the cross stock functionality to be able to get quotes through the cross from a cross stock perspective, right? That's not in there. So, those those are kind of the next pieces that that we're playing with right now there. Can't wait to hear more about it. That's awesome. Um, all right. Well, we're talking about some near-term things that think the these are things that you guys are rowing out pretty imminently. Um, I know I've had you for a good amount of time today. I want I want to wrap it up with this. Um, you know, as you guys are looking ahead, you know, let's say five years, uh, what's what's next for Warp? Yeah, the most exciting pieces are getting those those cross stocks to be fully roboticized. We see that there's nobody really in a better position than us to take freight into this. It's really just the 21st century, right? Like who's going to be the next great American supply chain company? Has there been one since FedEx? We don't think so. UPS is 150 plus years old, right? Amazon's already 40 years old, right? You know, we're not going to call them a logistics company yet, right? So, we see it as our responsibility to take the mantle of being the next great American supply chain company. And it doesn't need to be American supply chain tech. We've been call like you and I have been in this industry a while. We call it supply chain tech and logistics tech. No, it's just logistics. It's just supply chain, right? It shouldn't even be thought of as the tech, right? The tech should just be doing the job of getting shippers lower and lower prices. Changing those conversations from being around, oh, inflation, you know, uh fuel search charges, we need price increases, right? You know, cuz our network's getting more exclusive, right? More volume equals more density equals lower cost structure equals lower prices. Very simple equation that this business model should operate under, right? You saw Carnegie Steel operate under the same business model. You saw Rockefeller, Standard Oil. This business model that we're talking about in commodity businesses, it exists. Yes. then maybe you could say there's pricing power down the road. But you know like as a Jeff Bezos said right your your margin is my opportunity right like we don't plan to take great you know right like we don't plan on taking egregious numbers which I think FedEx and UPS and all these LTL carriers got away with just because they had the networks they had the terminals the moat was I'm an LTL carrier I have 50 terminals who's going to compete with me right I'm FedEx I have 100 sort centers who's going to compete with me I could set the price. But that experience has worn on all these brands and retailers. They've gotten away with decades of price increases, you know, rate hikes, all these, you know, peak search charges. Somebody could come in and it's going to be us. Somebody could come in here and actually just break the model of how logistics commercialization has worked because there's finally enough tech and automation that allows you to get a competitive cost structure with these massive legacy networks without the capital intensive investment. We're we're definitely in interesting times. Uh I think there's probably never a better time to build than there is today. And so congrats on on your success so far. Uh and and thanks for sharing the story of Warp and where you guys are headed. It's very exciting. Uh I I think um uh I think everyone needs some pricing relief and so hopefully you guys can continue on that journey to to to continue to reduce those costs uh for years uh for years ahead. Well, Troy, look, thanks thanks for coming uh on the show and joining me and talking about the story of how you guys are reshaping the middle mile. Um uh where can people learn more about uh Warp and also if they want to connect with you? Yeah, they could go check us out at wewarp.com wp.com or connect with me directly on LinkedIn. Awesome. Listeners, as always, subscribe to the Logistics Remix channel on YouTube for more conversations that make sense of shipping. Give us a like on this episode and thank you for listening. Troy, it's been a pleasure. Thanks for coming on. Thanks for having me. And everyone, I look forward to seeing you next