Ep 126: Will AI Spark A New Logistics Revolution? Harish Abbott Helped Pioneer Next-Day Shipping — And He’s Betting Yes.

0:00It's not an option not to use AI. Some people are going to try to delay it. Those folks are in trouble. Tell us a bit about working at Amazon. You would get emails from Jeff Bezos when a customer got their book laid. It's one out of maybe 50 ,000 books that were shipped that day. You were previously the CEO of Deliver, which sold for$2 billion to Shopify. How did you come to the idea of Deliver? There was this new set of merchants that were on Shopify-like platforms. These merchants cannot offer the fast one - and two-day shipping that Amazon offers. Let's go build this for the non-Amazon world.

0:29What was the inflection point to get back into the arena? Basically, what we're building is a teammate called Augie. It works 24-7. It can handle emails and text and phone calls and telegram or any other means of communicating. Rethinking logistics and reducing waste from this industry, that's a$300 billion opportunity in the US, almost a trillion dollar globally.

0:56Aarish Abbott is a serial entrepreneur. He's a billionaire from selling his last company to Shopify. He knows the world of logistics as well as anybody. Doesn't need to work again, but he's back in the arena. He already has 50 engineers. He's hiring another 50. His new company, Augment, has an AA agent named Augie that's already directing tens of billions of dollars of commerce. It's a really interesting time. The whole world of logistics is being transformed. Humans are being augmented. What is happening in the world of AI and logistics? What is this going to look like for America? Excited for you to meet Harish.

1:26Welcome to American Optimist. Really excited to have here today Harish Abbott. You're the co-founder and CEO of Augment and previously done a lot of other things. You were previously the CEO of Deliver, which sold for about$2.5 billion to Shopify, right? That's right. Yes. We were lucky to be the first big investor, I think, for the very first round of both these companies, Deliver and Augment. So, we're excited to partner with you again. So, I'm coming from a very biased position here. I'm a huge fan of Farish, but it's very interesting because you've built in logistics a couple of different times now.

1:53You were at Amazon before. Tell us a little bit about your background. I think you started off at IIT, I think was originally where you were at. Yeah. So, well, thanks for having me here. Excited to speak with you and thanks for your trust in my businesses so far. Yeah. I grew up in India in a small town, in a middle-class family and I think went to IIT and then schools here in the US, Illinois, Urbana, Champaign, and Stanford. And I studied mostly like graph theory, maths, computer science, logistics. So it was just this combination of - You studied some logistics as well then along with computer science.

2:25That's right. Yeah. Yeah. And so this is this idea of like supply chain logistics. It was sort of the amalgamation of kind of like graph theory applied to physical word is very interesting. So I studied that. I mean, if I look back, like all my studies, I would say, I think growing up in a household, I'm just so thankful for my mom and dad, like just inculcating like hard work, grit, perseverance, you know, those values is like, that's what I think at the end of it, that keeps you going, you know. Those are very American values as entrepreneurs. You know, and that sort of keeps you going. And, you know, I just can't be more thankful, you know, to just be born in a household where those values were just, you know, sort of always inculcated in you growing up, you know.

3:07Are your parents, would they come over here as well now? They sometimes come, they're settled in, you know, they're in their 80s, they're settled in India, it's much harder to get, you know, to travel now. But they used to more often than now, but now they're settled in India, yeah. What brought you to want to come to America? The opportunity to build, right? Like there is, you know, in the, uh, so, you know, IIT is like this engineering school in India where it's really hard to get into, you know, literally millions of kids apply, a few thousand gets in. And, and in those times in India, you know, like in 96, 97, there's just not that many opportunities graduating to work for like young companies to go build.

3:48And America was this place to say like, Hey, this is a place where you can go build. People will trust you. People will trust your ideas. There's sort of merit matters and merit speaks. And there was this notion in people's head. And so I think I just followed that and came here for that. Well, we're glad to have you here. There's a lot of controversial takes now on immigration and people who come here. I think there's a lot of anger about certain things. But for me, if you hadn't come here, there would have been a lot less wealth created, a lot less jobs for people. And so we want the best and brightest here.

4:20Well, I mean, I don't think there is any country in the world like America where folks from all backgrounds, from all countries can come and can be trusted to go build businesses, endeavors, social endeavors, and just embrace of failure in order to go build these audacious things. And so I'm just, you know, I think it's the best decision of my life to come to this country. I'm glad this is a place that's still attracting some of the very best and brightest. You started off after Amazon for a while as well, right? Tell us a bit about the working on Amazon. Well, I think I was just lucky. Like it was, Amazon just sold books at that time.

4:59And it was like the tagline was Earth's biggest bookstore. Now you're going to have to believe they can sell almost everything under the sun. And it was just lucky to start, you know, as they were sort of building out their fulfillment centers, their logistics network, they wanted a group of sort of, you know, maths, computer science people who also had interest in logistics and to come and help build that out. And just learned a ton there, you know, obviously built a lot of stuff, wrote a lot of code, what is now fulfillment by Amazon. But also just, I think it was just like the most formative years of my life.

5:33I just, you know, I didn't work as closely with Jeff Bezos, but watching him from a distance, but did work closely with leaders like Jeff Bilkey, who ended up running all of Amazon retail and several other leaders under it. I mean, one is the customer centricity, right? Like this obsession about making customer happy, that customer is right. And just really, you know, at every single point, I remember you would get emails from, you know, Jeff Bezos when a customer got their book late and they would just forward it to the logistics departments with a question mark on why, you know, just one out of maybe 50 ,000 books that were shipped that day.

6:09Just this obsession about it. Just what's the exception handling? What went wrong or how do we make it better? What is the root cause? I remember this is the diligence or the execution at which the sort of diligent execution at which the whole thing operated. I remember, and I'll never forget, 8 a.m., we would have these meetings where we would have almost 30 metrics, and there were like seven or eight warehouses at that time. So it was about 150 metrics on a sheet of paper, and you were supposed to meet at 8 a.m. And the leaders at that time, like Jeff Wilkie, would come in prepared at 8 a.m., knowing the five numbers they need to talk about.

6:42what went wrong with those five numbers the night before. And the leaders in those warehouses had to come answer, like, what are they going to do today so this doesn't happen tomorrow? And just this rigor of that, to do that every single day to get to root causes. You know, I mean, there's so many other things I learned there. Like I learned like this embrace of technology, like Jeff Bezos, I didn't work closely with him, but I think Amazon was leader in books by 99, 2000. and then Jeff sort of incubated the Kindle project, which if you look back could be like, oh, you're sort of cannibalizing the biggest business you have because now you're gonna sell every single book for$9.99 when the average prices of books was like$20,$25 at that time.

7:29But his idea was like, if I don't do it, someone else will do it. Might as well let me lead in, let me shape this industry, let me shape this product. He was looking for what was gonna disrupt his business. He's like, I'll do it myself. Yeah. So comfortable disrupting your own businesses to go build new businesses and like really adopting it so fast. And I think there were like several attempts before Kindle got right. So just observing that, you know, from as, you know, as an early employee was just so useful for me. And I just sort of carry these lessons, almost everything I build, you know.

7:59What are some of the principles you've applied to your building? Is there anything else that you've taken from there and said, we want to do this like Amazon does it? Yeah, listen, like I think culturally there's this principles of, you know, hiring owners, right? That's very important. Amazon has this principle of like, you know, act as an owner, not, you know, owners like you can act as a renter or you can act as an owner. Owners take care of their homes. If there's a leak, they don't go and do the short route. They fix the roof the right way as opposed to patching it. They think for the long term.

8:30They're willing to take short term losses for the long term gains. You know, that's a big, big value that I carry forward. I like that. You know, this bias for action is a big one. I think tactically, I've just really learned that Amazon had this culture of like writing things down. So they had the idea like, hey, if you have a project or an idea in mind, take your time, write like several pages, sometimes five or six pages long document before you call the first meeting, share that document with everybody. Everybody comes prepared with that document before you start the meeting. what that did is like it forced on you to like really distill your thoughts because you knew other people are going to truly read your stuff so you don't want to be bad and like looking like stupid in front of like really good people and then by writing it a few times it made you like really rethink read still like bring clarity to your thoughts and so we carry that forward you know now like hey if you have a new idea you're you're good to do that but just write it down to think of it a couple more times, bring clarity to it before you start engaging the rest of the company and using resources.

9:34I like that a lot. I push a lot of my most talented friends when they're building things to force themselves to make the decks and do the thing as well and write it down. Because a lot of people, when they've already had success, they think, okay, this time, I don't need to write it down because I already know what I'm doing. But I think no matter who you are, it seems like a good process. I think it's a great process, you know, because we all have blind spots. And when you write things down and you read it again, it's like, hey, this is not connecting well. You missed something and others read it and they poke holes at it.

10:00I mean, I'm sure you guys do that in your investment memos, right? You ask people to deeply think about the investment memos and that's sort of a similar exercise. It is. We changed our mind on a couple of things this week and it's just very tough. It changes when you write it down and you iterate on it together. Yeah. I want to jump into Deliver. So we were very lucky to meet you. Actually, I think we met you because we were in our build program thinking about what needs to be built in this space, it was very clear that a lot of these businesses didn't have what Amazon had. And we were thinking about it and talking to all the smartest people and they led us to you.

10:33And so you led us back. How did you come to the idea for Deliver? What was Deliver doing? Yeah, this is like Deliver was like, our long-term ambition was like, hey, the physical world is such a large industry. It's not programmable today. And let's make physical world programmable. Let's build an AWS for the physical world. So people will be able to like, you know, use things in microspecs. Like today you can go on AWS and you can ask for a compute and a single cycle of compute and pay for a single cycle of compute. It doesn't happen in the physical world. The leases are really big or contracts are super long.

11:06By physical word here, you mean logistics, you mean shipping and warehouses. Yeah. Shipping, warehouses, truck capacity, you know, like cars, like moving things, ships, planes, everything sort of that makes, that people touch to get this mic in this room. And so that was sort of the long-term idea. I'm like, okay, but where do we start? That's a need. And we was like, hey, e-commerce at that time, and I think still is, sub 20 % of all commerce at that time, it was close to 12%. It was growing fast. And then there was this new set of merchants that were emerging on Amazon, but also on Shopify platforms or Walmart marketplace or eBay.

11:49And like these merchants cannot offer the fast one and two day shipping that Amazon offers. And really the only way to get to fast shipping is almost like building a CDN for the physical world. You need your own network of warehouses close to the consumers, which Amazon can do. If I'm a random merchant in Texas doing something pretty cool that I want to sell to people, I'm not going to be able to afford my 50 warehouses. You're not able to be afforded. On top of that, the logistics of it, you may not have enough inventory to then ship it to those 50 warehouses. you may not have the analytics and data science to then make promises on the end customer to say like is this going to show up in one day or two day or three day and so there's a lot of infrastructure needed to make that happen and so i'm like hey let's go build this for the non-amazon world right and uh and and the problem was somewhat unique because we were dealing with sort of smaller merchants that didn't have the depth of inventory so we had to go a several level deeper in data science.

12:45I'm like, hey, you only have 50 units. Where should I put it? How should I predict what New York is likely to buy versus LA versus Seattle? And so it became like both a physical infrastructure problem, but also like a lot of interesting data problem because we just didn't have the depths of data, but we could use data from several merchants who were selling similar products. And once you had tens of thousands of merchants on it, you then got this scale to get all of them better prices. And so it seems like it became a little bit of a network effect. It is. It does, right? Like, you know, like merchants would send us inventory on the port of LA if you're a small merchant, but there's no way to physically, literally distribute your inventory seven different ways in the country.

13:20But if I pull on that truck, like 50 merchants inventory, the truck is going out full. Everybody gets a benefit from it. So more people participate in it, you know, so it became a network effect thing. Yeah. It makes a lot of sense. And I guess one of the other biggest players to aggregate merchants was Shopify. So you became obviously very close to Shopify over the years. They eventually bought you. How'd you get to know Shopify? How'd that work? Well, I think a lot of our merchants were Shopify merchants. Shopify was the fastest growing e-commerce platform. It's now, I think, close to 12 % of US e-commerce.

13:50Wow. So pretty meaningful. They power, I think, close to 250 to 300 billion of online commerce now. And a lot of those merchants started to use us. And so in order for us to make promises on Shopify platform or to give, you know, shipping things on Shopify's cart, we had to start partnering together anyways on a product side. And I think Shopify started to see us and saying like, ah, that's interesting. When people use Deliver, you know, they sell more because they're able to make faster promises one day and two day. And these guys are deeply integrated into our platform. And it's becoming a core and core part of some of their biggest merchants sort of, you know, presence online, right?

14:34And then that's how we started to work. Yeah. And they ended up buying you. You're still, I think, very bullish on Shopify. I understand you're a big shareholder there. Obviously, they bought you for over a couple of billion dollars. There was a little bit of a strange thing where they had some problem internally and they sold some of it to someone else. What ended up happening with that? It seemed like it was a messy thing there. Yeah. So listen, I think one is when we sort of got into Shopify, I think one thing as an entrepreneur that, you know, when you get acquired by a very large company, on one side, you get the benefit of the scale, the brand, like things that you never thought you would have, right?

15:11Like a$10 billion company would pick up their phone call and like call you and say, if you want to work with you, you know, so that, but on the other side, you're no longer the main agenda. There are other agendas in the company and you are a sub agenda. You are a sub point. You are a sub sort of, you know, mission in the company. And you have to sort of figure out how do you fit into the main mission. I think also culturally, there's like two different things, right? Like we were a low gross margin business, large, fast growing business, but it's not software. Shopify is a really fast growing business shop, you know, software.

15:44So when you merge the two cultures together, there are interesting things like the way we pay, the way we hire, you know, just bringing the two groups of people under single HR and aligning like their pay comms and stuff was a very hard exercise. Like I never appreciated that. But yes, on one side, when you have a high gross margin business like Shopify, and you merge a low gross margin business, the construct, all of that are very different. And you have to merge the two in a very interesting way. So I think it sort of became clear to Toby and me like, hey, this is an important thing for the merchants, but it is still a side quest for Shopify.

16:22And their main quest is to build the digital infrastructure for online in commerce. The physical infrastructure is equally important, but it's better done outside of Shopify than inside. And so then we worked with Flexport a little bit and a few other companies and figured out that Flexport was the right partner to go continue this mission. Is it still happening there? They had some trouble separately. No, it is. It is growing. It is growing. I think the last I heard from Ryan is their warehouses are full. They're adding more capacity. It's a pretty meaningful business now. Good. Well, I'm still a shareholder in Flexport, so I'm glad to hear that.

16:58Okay, that's great. Maybe it'll make us all money again. And so post-acquisition, you're rich enough, you don't need to work again, for sure. What was the inflection point to get back into the arena? Why are you going again? Yeah, listen, as most engineers in the world, AI started to come through and you start to play around and prototype stuff in AI. It's one of the most transformative things that I think I have seen personally as an engineer. And I think two things happened to me. I started to see the reasoning models come through and just the power of those reasoning models was incredible. And then second, the context windows started to get larger and larger.

17:33And so, you know, now I think Gemini just announced like kind of a million token context window, but it was starting to become bigger and bigger. And like, ah, the reasoning is coming through, you know, this context windows are getting bigger. Now you can pass a ton of context. You can ask these LLMs to take on fairly complex, you know, business issues and cases and see what you can do with it. And, you know, being from logistics, I'm like, wow, this is sort of a moment where you can finally like rethink this thing. Like if you take a step back, Joe, like logistics in its simplicity is about trading information so you can trade goods, right?

18:14Two companies, whether you are a warehouse or a truck or a shipper or a broker or a warehouse that is a receiver or a ship or a plane, freight forwarder, you're trading information and now so that you can trade goods. And then that makes the goods move from point A to point B, which makes civilizations happen, trade happen, markets happen. So it's very important, but that's what it comes down to. Now, the trading of information here, you know, in the logistics world has been has been mostly like asynchronous. Like it means it's going through methods like emails and phone calls and text because there's very high degree of fragmentation in this space, right?

18:55Like if tomorrow I decide like, hey, building businesses or software is not my calling and I'm gonna drive a truck, somebody would give me a loan and I'll be on the road driving a truck. So there's 1 million truck driving companies in America. So there's 1 million. 1 million, it's not kidding, 1 million, right? There's 40 ,000, 50 ,000 trucking brokerages in America. So there's just a lot of fragmentation in this space. 40 ,000 brokerages. The average brokerage is one person, but some of them might have hundreds of people. Oh yeah, and there are brokerages that have like 5 ,000, 6 ,000 people too, the very big ones.

19:26And so when you have an industry with this structurally high fragmentation and you have to inherently trade information to make things happen, what do you trade information with? You trade information with tools you've got, which is emails and phone calls and texts. It's a giant mess. And a lot of people in the past have tried to fix this, obviously. I guess Uber even built a business there for a while. Uber, Convoy. Convoy. And why didn't those guys just fix it and make a big platform for everyone to use? I think one is that one, to get a million trucking companies or to get on a single platform is a Herculean task, right?

19:59Like it's like, how do you get so many different companies to say, trust us, come on my single platform? Two, it is a low trust industry, right? People don't want to share all their information because they feel like that will be used against them for negotiation. I give an example to people and saying, if you're a truck driver with home base in Austin, and you've been out of Austin for the last three weeks, and now you're in Chicago and you want to get back into Austin, and you tell everybody that, hey, my home base is Austin. I've been out for three weeks. I haven't slept in my bed in three weeks.

20:28Now I'm in Chicago, and I want to bid on a load that is going from Chicago to Austin. Well, if you give them all that information, they're going to give you a really bad price because they know how desperate you are. Yeah, and if they're the only one going, then you're not going to get paid very much. And you're not going to get paid very much. So people are not that willing to trust and give all your information. They want to keep the information siloed to protect their business interests. So when you combine the two, it was really hard to get everybody onto a single platform. What AI changes is that, hey, you can be where you're at.

21:00If you want to be in your emails, on your phone calls, audio text. You don't even need a new system. We can get you an AI employee that can take on all this tedious work of coordinating. It won't stop from nine to five. It will do 24 seven, right? It will be more knowledgeable than over time, like any one of your single employees. It'll know everything going on. Yeah. It will know everything that is going on. It will be smarter in responses, but now you can have your people focus on more creative work, more value-added work, customer success, things that really, truly move the needle for you. The long-tail brokers, are they going to use AI too?

21:38Are they not going to be needed? Is it more of the big firms that use AI? How do you think about that? I think everybody would use AI. It's not an option not to use AI. It's getting to a point. Everybody would use AI. But I think there's sort of two things that are going to happen. Some people are going to, you know, try to delay it because it does require change management and it's tough. Right. And so I think those folks are in trouble. Like the more you delay this thing, the more chances are that your competitors are going to come eat your lunch. You know, that's number one. I think there might be some more consolidation because what I have seen is that technology,

22:22does create bigger, fewer companies get bigger with technology, right? Like if you think about just e-commerce, they're used, now you have between Walmart, Amazon, Shopify, and maybe few other players, you've got like close to 50, 60 % market share. And so I think few big players will get bigger, but everybody has to use AI. So this company that you're doing this with, It's called Augment. You started. You already have 50 engineers, I think you were saying? Yeah, 50 plus, yeah. So it's a big company. You're partnering with some major brokers to iterate with them. And fleets. And fleets as well.

22:57And I guess a lot of these really top brokers, they're top people, make a lot of money. And you're helping these people be even more productive? How does that work? Yeah, so I think basically what we're building is a teammate called Augie. And people rename Augie. They've chosen all kinds of names for Augie. It's interesting. They even created pictures for it and personas for Augie. Augie participates in people's weekly town halls and monthly town halls. So it's really interesting to see how an AI teammate is becoming part of a culture of an organization. But really behind the scenes, Augie sort of works.

23:33First, it works 24-7. It is multimodal. It can handle emails and text and phone calls and telegram or any other means of communicating. It does, actually. because there's a lot of dispatchers in this business that are in Eastern Europe and Eastern Europeans only use Telegram. And so Augie has to use Telegram to communicate with them. Similarly, like there's a large community of drivers that are from like Punjab in India. And, you know, those guys use WhatsApp. And the way to communicate to those folks is WhatsApp. So Augie has to use WhatsApp, you know. So it's multimodal, just like you would expect, you know, a teammate to be.