How Augment’s AI Engine Helps Supply Chains

(upbeat music) - Hi everyone, this is Lee Klaska. When we're talking transports, welcome to Bloomberg Intelligence Talking Transport's podcast. I'm your host Lee Klaska, senior freight transportation and logistics analyst at Bloomberg Intelligence. Bloomberg's in house research arm of almost 500 analysts and strategists around the globe. A quick public service announcement before we dive in, your support is instrumental to keep bringing great guests and conversations to you. Our listeners, and we need your support, so please, if you enjoy this podcast, share it, like it, leave a comment. Also, if you got ideas, feedback, or just wanna talk transports, I'm always happy to connect. You can find me on the Bloomberg terminal on LinkedIn or on Twitter at Logistics.ly. I'm very excited to have with us today, Harish Abbott, the co-founder and CEO of Augment, an AI productivity platform for logistics. Augment has raised about 110 million from Red Point Ventures, HVC, and leading logistics funds. Prior to launching Augment, Harish co-founded Deliver and e-commerce fulfillment platform acquired by Shopify for $2.1 billion in 2022. His career includes pivot roles at Amazon where he contributed to building global fulfillment infrastructure. He holds degrees from Indian Institute of Technology, Rookie, the University of Illinois, and an MBA from Stanford. Welcome to Talking Transport, Harish. Thank you for having me, I'm excited to speak with you here. Yeah, it's great to have you. Could you talk a little bit about what Augment does? Sure, so Augment is an AI productivity platform for the logistics businesses. Our first product is Augie. Augie, think of Augie as a teammate that can work 24/7 across all modes in a business, emails and phone calls and text and messaging and your systems of record like TMS or WMS. And it goes and does work that is given to Augie as SOPs or standard operating procedures. Augie's getting deployed to lots of freight businesses today, mostly large freight brokers, freight fleets, and now of shippers. Yeah, that's what we do. And Harish, could you talk about how a broker might use Augie versus a shipper? Sure. So if you think about what does a broker do, is their job is to match demand to supply and then execute against it. They get demands from shippers of the word and then the supplies are from several fleets, large fleets and small fleets. They try to match them and then execute a shipment end to end. Now in the execution of the shipment, there are multiple, multiple steps. Today, extremely manual, right? There are steps like, hey, I get a shipment in an email, I need to now put that load up in a TMS manually, maybe, or correct some fields from a load that came through an EDI. Do now I've got a shipment, I need to now go source capacity for it, I've got preferred carriers, I've got data on who's run these lanes before, or sometimes I get a load where I'd never run this lane before, so I need to post it on a public load board, like the AT or truck stop. Then there is this work of like, hey, I've posted the load, I need to now negotiate the load, go back and forth to figure out how am I gonna make my margin on it, once you figure it out, like, hey, this person is gonna do this load, this company, then there's a whole execution step off, I'm gonna dispatch the load, I wanna make sure the driver understands how to get in, I have to set up appointments for a load, right, on the pick-up side, on the drop-off side. Then I have to track the load, I have to keep updating the customer on where the load is at, if there's delays, what to do in terms of those delays. And then when the load gets delivered, then there is the whole paperwork cycle begins off collecting the delivery documents, collecting the invoicing documents, and making sure if there's accessories, detention, lumpers, those are collected so that the billing can happen fast. So if you look at this life cycle for a broker, you cannot deploy Augie in any or all of these steps, almost like a worker, and saying, hey, Augie, now you can be looking for all my emails, and look for tenders, and if you get a tender, maybe you start building the load in the TMS, and if you're missing information, maybe you can come in and ask somebody, and saying, hey, I'm missing this information, what does this shipper really mean, or go back and ask the shipper, so you can completely build the load, or once the load is built, like, hey, just start working on capacity sourcing, so collaboratively work with the rep, understand their preferred carriers reach out to them, maybe post the load on a DAT, get the incoming calls, take those calls, negotiate the load, some businesses say, like, pass the final negotiation to the rep, some people say, like, you can book the load, and then, like, let's start working on tracking the load, or tracing the load end to end, so making the dispatch calls, understanding, you know, the ELD events, making the scheduling appointments, and then, all the way towards the end, is like, hey, the load's got delivered, now we need to go and chase documents, and view, and I know nobody likes chasing documents, it's like, you could argue to work, and it can chase documents, or DSOs, or DS sales outstanding and shrink, and we can start, money can start moving in this business again, and faster and quicker, so those are like some of the use cases that people are using, argue for today as a broker, but also using, argue for, like coding, you know, like, they get lots of requests for calls, like, hey, could you respond to my code? - Okay, so just a quick commercial, you know, you mentioned DAT and truck stop on the Bloomberg terminal, their data is available, and it can be found at the BI space TRC KGO, on the Bloomberg terminal, so if you're interested in that data, we have a spot, contractual, workload market data from those two data providers. You know, you mentioned, so there's like three main customers, you guys are going after the shipper, the carrier, and the broker. Right now, kind of what is the mix, and kind of, you know, 'cause I know you're in growth mode, what do you think the optimal mix would be, three years from now? - Yeah, I mean, today, I think I would say, 60% of our businesses are brokers, right? And they tend to be the largest brokers, about 35% are fleets, and then we're just starting to work with shippers, so we're not as deep with shippers yet. I think as we go forward, you know, our vision is that, it is not about, I think step one, Lee is about taking this tedious, repetitive work off humans' plates, so they can focus on more important things, right? That's what we're doing. But if you imagine, you know, Oggies on a broker side and a fleet side, and the two Oggies can now synchronously collaborate, then we can eliminate the need for the phone calls and the text and the emails, right? So our mission is to make logistics better, and we think the way to make that better is to get to more real-time synchronous collaboration between shippers and brokers and fleets, so that people can act faster, right? So if you, you know, your question was like three years, so now I would sink three years from now, most of our businesses, like we're equal mix of, you know, shippers and brokers and fleets, and we are really moving towards getting all of them, the signals they need, so they can act in their business faster, right? Like, as example, it would be, well, today for trucks running late, what happens is, and it's a broker load, the, maybe the offshore team or the overnight team at the broker has to, like, keep an eye on it, if they get alert, the trucks running late, and they might decide, oh, I need to reset the appointment, the resetting of the appointment means sometimes you can log into a portal, sometimes it means you need to email somebody, and let's say if it's an email scenario, you email somebody, and if the receiving warehouse is not working at that time, well, till 9 a.m. next morning, you're not gonna get a new appointment. So now, the labor that the warehouse had planned is going to get wasted, because the truck's not showing up on time. The truck, when it shows up late, doesn't have an appointment, so it might need to wait for hours, sometimes days to be unloaded, and in the meantime, the shipper's inventory is unsalable, so everybody's hurting. The shipper's hurting, the warehouse was hurting, the truck guy's hurting, and the broker's performance is hurting too, because the shipper holds them accountable, and, but now if you imagine a scenario where an auge on the broker side detects the truck's running late, calls or synchronously communicates with an auge on the warehouse side say, "Let's do finally a new appointment that appointment is now booked." The labor that was allocated for the first appointment is re-allocated to something else, more productive, and now that when the truck shows up, it gets unloaded, and the inventory gets available, right? So that's sort of the future we think, like AI has the potential to enable, and we are starting to build for that, and that's a better future, because now you're making supply chains, more efficient, you're taking waste out of supply chains, right? - Sure. - And so, you know, AI, obviously it's a very sexy industry, especially if you're doing it for transportation. It's becoming a crowded space, 'cause there's a lot of competitors out there, Eddie differentiates what augment is doing versus maybe some of the other competitors out there. - Yeah, it is a crowded space, partly because it's so easy to build a prototype. You know, you can vibe quote a prototype, and if the buyers are uninformed for them, it's very hard to tell the difference, you know? The difference really comes in when you actually take a product to production at scale, and test to handle all these extenduating circumstances, and then I think prototype starts to break. The way we differentiate it, we have few core concepts or principles. The first one is that, you know, you can really do mean, you can only do meaningful work when you have end-to-end context. So point solutions of voice bots or email, or like a code solution, if they don't have a full context of like, here's my customer, this was the customer's SOP, this customer is an enterprise customer, this has been my performance so far. If you don't have that context, when you're building the load, tracking a load, when you have to make a decision if the truck's running late, or you have to make a decision to book a load at a negative mod, you can't do that. And to bring true productivity, you need to have end-to-end context. So augment is building, I would say, an order to cash set of use cases for all of our businesses, but that it's brokers or shippers or fleets. And through those order to cash use cases, we can now carry context and do a better job at, you know, each step of the way. I think two is we believe that, you know, work is inherently multi-modal, right? So it's not just about calls or emails or tech, it is about all of that. Like in some businesses, a lot of work gets done on teams or Slack, and there's a lot of discussion that is happening between people on Slack, and it's extremely important for all of you to be part of that conversations because there's a lot of context in that, right? Somebody might say, hey, I just got a call, this shipment has become an urgent shipment, we need to now start tracking it 15 minutes. Okay, if Augie is part of that conversation on Slack, it now knows that it can start adjusting its work, so the shipment updates need to be sent every 15 minutes, you know, and so we believe that you have to meet where people are and the work is not just any mode, it's all modes all the time. And so we had to build a teammate that almost looks and feels no different than a remote employee, but it can be added to any of these channels, you know. Third is we are unlike many horizontal players, like we're saying, hey, we have a AI platform, generic platform and you can now implement it, we are purpose built for the logistics space, right? So we're saying we understand the ontology of freight and logistics, we understand what a multi-stop shipment is, we understand what a flatbed track is, we understand what does appointment times mean, we understand sort of all the key players that a highway fraud score or a, you know, a marker on DAT or truck stop, what does that mean? You don't need to teach Augie that. It comes learned about this space. So those are the three differences. One is very holistic, order to catch set of workflows, people can pick and choose of course, but our hope is that they end up deploying Augie across a vast majority of order to catch workflows and you get to see more benefit, very multi-modal and you know, very purpose built for logistics. - So when you say multi-modal and you talk about carrier, so you will do not only truck load, but less of truck load and intermodal and other things that brokers are, you know, broken. - Yeah, I mean, we today do all modes, right? All modes across all channels of communication. Like that's what I mean by multi-modal. So LTL, FDL, flatbed, auto, you know, DRAGE loads, we do that and we build the ontologies for all of those, but then the channels of communication are also quite varied, right? Like if you're looking at LTL word, one in the LTL word, there's really only 25, 30 big LTL carriers in this space and the way you communicate to them is a lot through their website actually, right? But in the FDL space, it's a really fragmented industry and the way you communicate with them is maybe through emails and text to the dispatchers and calls sometimes to the drivers. And so Augie needs to have sort of all of those channels of communication, but needs to also understand the oncology of all of these different modes. - And you mentioned the truckload market being very fragmented and very large. What does success mean for you guys in terms of market share? Like how much of the market do you think you guys could actually control of, I don't know if you would measure it by the AI spend or just how many, the percentage of the industry that you're working with? - Yeah, great question. So I think like first, let me, the way we measure success is delivering ROI to our customers, right, like, hey, by deploying Augie, what is the return on investment? What is the meaningful financial gain that you are able to get, right? And we actually measure that with every business we get into. And I think if you do that, we hopefully acquire more market. But the way we see market share is freight under management. Like how much freight under management is Augie assisted or Augie powered? In this last year, we've gone from almost nothing to now about 40 billion of freight under management that Augie is getting deployed at. Of course, you know, this market is close to 800 to 900 billion. And then if you count the LDL market, you're getting closer to a trillion. So we're still very early, right? Like we're maybe like barely 4% of the market. There's so much to go and build. I don't know what the upper limit is. There are some players who are going to try to build this in house. There are some players who are going to try to buy it. We're just really obsessively focused on saying, hey, if you deploy Augie, can we get you a 7 to 10 X ROI? And if we can, we think more people will use us, you know? - So that's the goal of 7 to 10 ROI for our businesses, that's right. - Can you talk about some of the productivity gains that you have generated for a broker, or for a carrier, like any anecdotal stories that you might have, or if you have some consolidated stats, that'd be great too. - Yeah, let's do that. So I mean, I like to go specifics, I think, because that's where the, like, people get, start to hopefully learn from the color here, right? So one of our customers is a fairly large, LDL focused, you know, 3PL. They have about a billion dollar of freighter management. And their first sort of issue that they came to us was that, that they had a 40% team, mostly offshore, that was reliant on spreadsheets and fragmented homegrown systems to do tracking for urgent shipments, LDL shipments. So what they were really doing was they were logging into all these different LDL carriers websites, maybe calling them to get a pro number. And they had done a fair bit of tech investments leave before, but 30% of their shipments sort of still lacked pro numbers. And then after, if you don't have the pro number, you can get the real tracking. And so they deployed Augie and said, like, hey, we want Augie to now go get the pro number, go get the latest update, keep posting those updates into our system of record, the TMS. And now they have sort of freed up 12 out of the 40 people and repurposed them for more growth related activity, right? So, I mean, what is that? About 25, 30% productivity improvement on a 40% team. And now the team has gotten confidence that, oh, I can hand over all these urgent shipment tracking to Augie. Now they're like, okay, what is the next thing that keeps my team busy? And how can I help them focus on things that I really want them to, relationships and margins? So that's like an example of that. There's another example of, you know, like a very large customers of ours, they do about $4 billion in revenues. They have a network of close to 80,000 carriers. And they had spent a lot of time and built a fairly extensive automation. Like their automation budget is in tens of millions a year, right, almost 16 million. And they were, you know, for a small percentage of their loads, call it 15 to 20%, that was being covered through the public load boards, like the ATO truck stop, you know, they were getting about 3,500 calls and 5,000 emails on a daily basis. And a vast majority of them were getting unanswered, right? Because, like just to imagine, like the workforce you need to answer 5,000 emails or 3,500 calls. So what did they do? They put algae on it and they say, "Hey, I want you to be the first line of defense." You know, answer these calls and then, you know, follow up on emails as you need to. And so now, you know, it's taking all of the among calls today, right, every single day and negotiating and booking loads that's freeing up people to do, I think, more higher judgment things. Now, if everything in a call is passed, like their fraud score is cleared up, they're not on the do not use list. They meet the criteria of the business. They have a good price, you know, the freight fleet company has a good price. Then, algae is passing that. But now it has done all the heavy lifting of qualifying, getting a good bid and then passing that bid and transferring their call to a rep. So when the rep takes the call, they're now really making use of their time well, right? And the same thing with email. So through this, plus about eight or nine other use cases, like this business, you know, expects somewhere between six to 12 million improvement in EBITDA, over 12 months, 12 to 18 months. So you know, so we were working with them on that. You know, one of our customers is, this one we have, I think, published case study with them. So it's called Armstrong Transport Group. They, you know, they are a large TPL. It's sort of mixed. They have agents and then they have W2 brokers. So they have a bit of both. And, you know, they had also a fairly, fairly large team that was doing day to day tracking and tracing of loads. And, you know, they're now putting algae as the first sign of defense to, you know, track these loads and areas, escalations and whatnot. And it's made a remarkable difference. Like I think the last we touched now, 40% of all TMS updates on track and trace are now made by algae from zero, you know. And roughly, it's in just like thousands of updates that needs to happen on a daily weekly basis on these. So it's that many meaningful numbers. The other thing, the, you know, ATG or Armstrong did with us was put algae to work on document collection, right? I was actually the first to use cases. Like, hey, truly, no one likes document collection. So I was one doggy. And their, their key was like, hey, when, how many, like, what can I do to reduce my DSO hours? Because every hour, as you know, like brokers can only get paid after the invoice. They have to pay in some ways to the fleets before the shippers pay them because of the term. So almost everyday counts because they're otherwise running on, you know, credit facility with the bank to cover this gap. So we actually brought down, algae starts to chase these, these documents through emails and phone calls and texts. And was able to, and then verify these documents whether they are signed or signed or they matching the load number so that the billing can happen in a really good way. And they were able to bring down, you know, the DSO from, I think, 58 hours before algae to now, like, less than 36 hours. And it's still coming down, they're optimizing it. So across the board, like, we're starting to see, you know, meaningful impact in, in like, true financial metrics for these businesses. - So given the focus in the truck industry recently on a non-domicile CDL holders and, you know, English language proficiency, is algae able to help brokers kind of like exclude those, that, I guess, population of the driver market? - Yeah, so, you know, we are a law compliant down. So if that's the law, then we follow the law. And there's a couple of cases. There one is, it auto today actually, it works with a highway real time. So as soon as we get an email or a call, algae, before we even take or pick up the call, we work with highway, if you know who they are, like, they're already a great company into fraud detection, but they've also added CDL verification for the carriers themselves. So you can get that signal from them. - We had one of their executives on the podcast earlier this year. So if you're interested to hear more about highway and fraud or to prevent fraud, it's success listeners to go back and take a listen to it. - It's a great company. We enjoy working with them, great partner to us. But yeah, we basically, you know, algae will like, algae in a highway will like ping each other on a higher sense of the information both on fraud but also on CDL. And then if it's failing the CDL test, you know, we will obviously not take the call and politely decline and saying, "So I don't think we can work with you right now." - What's nice, that's nice that algae is polite, I like that. - Yeah, you know, like, yeah, I think AI has like, "Yeah, it's more patient than human." (laughs) You know, and so, and then on top of that, like we have conditions where let's say we are making a call to a, you know, a driver and we are sensing that they're not able to pick up the language English on that very well, right? We can analyze the call and we can rate the call for proficiency in English. And that's a pretty good signal. We can at least mark that to the business and say, "Hey, this is the proficiency of English according to us in this call. Do you want to do business with this company or not?" Ultimately, it's, you know, the broker or the fleets or the shipper's decision or art, but we can give them the data much more real time than it would be available anywhere else. - Interesting stuff. And then, so, in addition to the CDL and the English proficiency. So, as you mentioned, you're in constant contact with highway about fraud. Are there any things that Ogi does by itself to detect fraud and make sure that fraud is not happening? - So, a few things. So, one is that it supplements what we get from highway and almost all of our businesses have their own do not use list for a variety of different reasons. They could be fraud, but they could also be for performance related reasons or paid related reasons, you know? So, it supplements the signals it gets from highway with, you know, the business data and does that. That's, I think, one thing. I think the only other piece where we can help or help is in track and trace, one of the frauds in this business is that the truck starts to go off route. And it's, let's say, supposed to go from Atlanta to Chicago and there is a route for it. And, you know, it starts there, but then it starts to go off route. Now, typically, you can try to set like, you know, a geofence and saying, hey, if it's within this geofence, I'm good, but if it crosses a geofence, I need to know. This is a fraud where somebody is literally stealing, right? It's like, it's the worst kind of product. They're just off. They know there's a high, like high valued goods in the container and they pick it up and then they want to just steal or go to a warehouse and replace the trailer and then still do a delivery but with like, very different goods or whatnot, you know? So in that case, you can put algae to say, like, hey, I want you to be like a watchman, right? And say, like, hey, watch these signals. And if it's on its way, we're good. But if it's not, please start to alert internally whether it's through calls or, you know, teams or Slack messages to people and saying, hey, I think there's a potential. We, you want me to call the driver, I could call the driver or maybe you want to step in and saying, hey, why are you going this route? I don't know if we can prevent the fraud but we can like, maybe by this way, we can at least alert the authorities faster. There's a little bit more, like knowing that this can happen also can prevent the fraud. - So just changing gears a little bit. So like the brokerage industry is pretty fragmented. Obviously there's a bunch of huge players but there's also a lot of little players and you know, the broker industry was kind of industry where all she needed was a rolodex and a phone and it could be fine success, you know, fast forward today. Obviously, do you need technology and technology is becoming more and more important? If I'm like a small broker and I, you know, understand that I really need to start using tools like Augmented & Augy, where are the first places that you suggest a broker to start using AI and their workflow? You know, as they're kind of dipping their toe in the water 'cause they might not just have the funds to go all in. - Yeah, it's a really good question. So I think there's sort of two lenses to look at it, right? Especially for small brokers delivering an incredible service level to customers is how they grow their business, right? So if you signed up for, you know, if you get some business from a target or a Walmart or you know, you line like, okay, how did you do on my 50 loads? Really matters because that's how you are on your next 50 and your next 100. And a big part of the service is that, hey, was there on-time pick up, on-time delivery and load detention? Like, okay, can you always make sure those happens? So the broker's obsessed about that and they care about like delivering those services. So I think one of the easiest and early places to put AI on is just that. It's like, hey, can you like take over the whole track and trace type of initiative for me? And put me in in loop or get me in and saying like, if there is an extendivating circumstances as an issue, let me step in. But like having a pair of eyes that is like 24/7 on it all the time is because like most small brokers also don't have night shifts, they don't have sometimes overseas resources. And so, and truck's run 24/7. So having an OGI as like your night shift person at a minimum or an AI that can just monitor things all the time and alert you when you need it to, I think can go a long way in earning trust of the shippers. So I would say it's almost like less about cost to serve for small shippers, but it's much more about delivering an exceptional service level to, sorry, for small brokers, but delivering an exceptional service that was through their shipping customers, you know. - So as the current freight recession, hindered growth for you guys or as a kind of help growth is as, I guess, people are trying to get more productive with the, you know, the lower revenues that are coming in on the broker and carrier side. I think that net net has helped initiatives and companies like ours and AI in general because there's just tremendous amount of cost pressure, right. Like, I think we have what, I think three and a half years into now for kind of a price recession definitely in this industry. And so margins are raised in across the board, right. Like whether you have broker or you're a fleet, I think these are some of the lowest margins, maybe ever, certainly in the last five years that you have seen. And so having AI, starting to take some cost out, you know, almost becomes imperative and a nice hope in this environment that, hey, and then too as it turns around, which it will, you know, as a market now, now you have like, you can handle the cyclicality better, right. Because like before this business was like, it's so cyclical that as market gets hard, you just need to hire so many people to cover freight. Well, now you have AI, all these are the word that can like help you not go so high and so low. It can be like this bridge, you know, that if the market gets hard, you can just put AI to work to more use cases and you don't need to hire that many people. So I think there's like two benefits. But yeah, in net net, if I were to take a guess, I don't think it's certainly helped, yeah. Great. And then, you know, what has been the biggest challenge you face since founding an augment? The biggest challenge, like I'll put it like in two terms, one is for augment, you know, as a building the company, but then also working with our customers, I think like delivering value. Okay. I think working and building the company, it's talent, you know, like very AI native talent is an extremely high demand. And there are companies in every layer, whether it's the LLM layers or the data center layers and we are in the application layer, right? Every layer, this is a tremendous demand of that talent and it's not a lot of talent. It's, you know, that talent is getting groomed, the AI native talent. So we've got gone from I think like zero to like 100 engineers or so, but, but, you know, in short amount of time, I would say like less than, less than a year's. And, but it was an easy, you know, to get 100 engineers, we probably looked at, you know, close to 5,000 engineers, you know, and then obviously people trickle through the interview loop and then when you make them an offer, a lot of people have several offers and you have to compete on why they should join, augment, you know, and not open AI. It's supposed to work with you, of course. I mean, that's the number one reason. Well, I hope that goes a little bit, but it is, it is not easy. And we spend an enormous amount of time to build that teams. Because we think like talent comes first in this space. Now, working with our customers, I think the biggest, there are two challenges, but the biggest challenges change management. Like, when AI does something, somebody is not doing that thing. And AI is taking on real work. Like, it is taking on end-to-end work, right? Which means that the company needs to go and repurpose and retrain employees that you don't need to do this, but I need you to retrain to do this, right? And that requires a tremendous amount of effort and fortitude and courage to do. And there are leaders who are leaning in, but there are also leaders who are like, wait and see mode. 'Cause it is so much work. And it is, it's also courageous. Like, oh, I'd like to go and take 500 people and retrain them on new set of skills, right? And they were just so set of doing this one thing, but now AI is doing that. And I don't need them to do this thing, but I need them to do this new thing. And so a natural instinct is that maybe next quarter, maybe in two quarters, you know, versus today. So what that does is for companies like ours, is it pushes the ROI, where you can show that people need to do less, but those folks are not repurposed and flowing through the books, it gets harder to show, you know? And so I think there's just a lot of work that needs to be done in this industry and maybe there are consulting firms and whatnot who can help or folks who've done this. Like, how do you take folks and we are technology providers? And we have some, now some experience a bit, now observing across several of our customers, how what good looks like, what not. We're not the experts of change management. Like we need folks who can take these very large, three, four, five thousand people organizations and run them through change management. But the reality is that every single organization in two years time is going to look very different than what it does look today. - Right, it's interesting, you know, you keep on mentioning retraining people versus reducing head count. So you think most of the brokers and carriers that implement this are not going down that road. - So I think there's like, there's offshore labor and near shore, we're seeing their reductions. Like they're not, like there we are seeing very clear, like, hey, these teams I need to reduce and they're usually not on the payroll, their contractors. - Right. - But I think the folks on the payroll, most businesses are saying like, can I, the institutional knowledge, the culture, like can I actually see if I can repurpose them for growth? - Okay. - Like they are looking at it from that angle and saying, can I deliver more superior? But they already know who I am, they know how I work. You know, they know my systems, they know my customers. Like that would be a way to say like, I can have this talent go out of my door, right? It's just the skills that they were working on is not different. And now, so I think most companies are thinking that way. I don't know how much will end up on, can sales or customer relationships or new business expansion absorb all of that. I think the jury is still out on it. - Right. And so just, again, changing gears a little bit. Can you talk about a little bit how you got into transportation? You know, what made you kind of, as a technologist kind of gravitate towards the freight world? - Yeah, so like I started off my career at Amazon and I, you know, rode a lot of software which now runs a fulfillment by Amazon service. There we were big consumers of transportation, right? Like trucks coming from home. - They certainly are. - Distributing all over the country. I mean, now Amazon, maybe after Walmart or a couple other players probably will have the larger trucking consumers in, in, in, in, in, in the U.S. And then several of my businesses, whether it was a deliver or even the one before, we were either builders or consumers or both of the freight world. And so, you know, from that, I understood both the problems in this space. Like, hey, why doesn't, you know, why is it so tedious, right? Well, and like, why there are so many inefficiencies in this space, it's a large space. It's so crucial and important to, you know, for augment. So that's how I got into it. But I think the way, like I'm sort of a computer science math type of person, and I look at things as like this, node arc networks, you know, you have nodes and you've arcs and like arcs can be like ships and trucks and planes and, and nodes are like distributed warehouses or, or, you know, like small distribute, like last mile distribution warehouses or cross, cross-talking depots or, so there's like different nodes, you know. And truck is a, is an arc that connects almost all nodes in America. It says them the most common form of. So like, okay, if you want to go and change logistics in general or supply chain broadly, let's start with the most important arc. Through that, we will learn the ecosystem, the warehousing, the distribution ecosystem. And from that, we can now branch out to those verticals, if you will. We were starting off with trucking, but augment's mission is broader, right? Our mission is that can be make logistics better. And that requires coordination with nodes and arcs or different types. - All right, well, good luck with that goal in mind. You know, before we go, I always like to ask my guests, they have a favorite book about transportation leadership or, in your case, technology that's kind of close to your heart that you might want to recommend our listeners to take a look at. - Okay, so I think I'll give you maybe two. I think the first one is a very technical book. It's about how to optimize network, it's called network flows. The main author is Ravi Ahuja. It's like a seminal book on how you model physical networks into sort of this computer science way of thinking graph theory and then now you can optimize. It's very dense, but it's fascinating. And you can start to see networks in everything. Like, you know, like an airline is a network where airplanes going from one to the other. We also are networks, ships are a network. And how do you optimize it? And it's, you know, it's an interesting read. It might, for you to sleep if you read it on your bedstand. I think in terms of leadership, I like the hard things about hard things. You know, it's a great book about grit that how leaders just need to lean in on the toughest calls and the complexity of decisions that you need to make and the discipline of execution to build anything of meaning, you know? So I love that book. I read it, read it every time I have, you know, tough days, building businesses, I go back and read that book and say like, okay, it's not that bad, you know? Well, I really want to thank you for your time, Harish, and your insights today. This is definitely an interesting conversation. So thank you for having me. It was great chatting with you. And I want to thank you for tuning in. If you'd like the episode, please subscribe and leave a review. We've lined up a number of great guests for the podcast. So please check back to hear conversations with C-suite executives, shippers, regulators, and decision makers within the freight markets. Also, if you want to learn more about the freight transportation markets, check out our work on the Bloomberg Terminal at BIGO and on social media. This is Lee Klaska, I'm signing off, and thanks for talking transports with me. Bye. (upbeat music)