· David Senra, Alexander Taubman
[00:00:00] S0: You are. [00:00:00] S1: And our friend Zach tells us we're going to live to 150, so hopefully I could actually be doing this for at least 50 years, maybe 100 years. So in the fullness of time, we want to bring Long Lake and we want to bring AI to- Huge swaths of the economy. We're now operating in roughly $3 trillion of the $20 trillion TAM, so-called 15% of the economy in the US, and now we're global. And we just want to keep adding to that. So over time, we'd love to have a market-leading business and platform in many more industries. But there's moments in time when you can find a great asset with a great team at the right price, and then there's other times where certain industries are not attractive for long periods of time. And so we take a prepared mind approach, where there's this sort of our top 15 or 20 industries that we think are interesting for our strategy at any given time. We're constantly force-ranking those industries based on what we learn, based on how market dynamics are evolving. And the best way to learn about is look at deals in the space. And so we've built a really powerful sourcing engine where we're seeing 4,000 or 5,000 deals a year across all these sectors, and we'll literally do one or two new industries per year. So it's a big funnel, very tight filter, and I can talk to you about how we select industries and assets.
· David Senra
[00:00:00] S0: That's literally what I was thinking. I was like, okay, when you're select- I assume this is operating at the founder level, that you guys are the ones selecting, okay, this is the next industry we're going to to enter, correct?
· Alexander Taubman, David Senra
[00:00:00] S1: It's more based on opportunities coming our way. To go back to Buffett, you have a better chance of getting a date if you're bisexual public and private. You know that quote? Twice the chance of getting a date. So we like to kind of keep an open mind. If you're open to 20 industries, you're more likely to find a great business and a great team to partner with. We just keep that funnel going all the time. We're constantly looking, and when we see something, it's kind of like you know it when you see it. If there's a great business like Amex GBT with high 90s customer logo retention, 100-plus percent kind of dollar retention, with tremendous customer relationships with the best companies in the world, built average of 15-year relationship with their top couple hundred customers, deep, deep embedded partnerships where we sort of managed a mission-critical service for all these companies. Travel, I mean, it's a small cost and it's mission-critical. You don't want to screw it up. We talk about that century of customer trust, but it really means something, especially in travel, right? If you have thousands of people traveling in the Middle East and a war breaks out, or you have a huge snowstorm in Chicago and your entire team gets stuck in the middle of the night at 2:00 in the morning. Things have to work. Somebody's got to be able to answer the phone and know what to do. This is an amazing business. You just kind of know when you see it. [00:00:00] S0: So wait, was Amex your first entry into travel? [00:00:00] S1: Yeah. [00:00:00] S0: Okay.
· David Senra, Alexander Taubman
[00:00:00] S0: Okay. [00:00:00] S1: We'd been building a thesis in that space for like a year, meaning looking at several other assets. It was on our whiteboard of industries that we thought was really attractive for reasons I'm happy to go into and- [00:00:00] S0: Yeah, I want to hear about this. What do you think AI is going to do to this business? [00:00:00] S1: I think it's going to make the travel experience much more seamless. I think things can be much more personalized, more proactive. Right now there are a lot of legacy messy data issues and systems issues in the travel space. You have these supplier systems and customer systems and distributor systems, and it's a longstanding industry, so over many decades. And so it's kind of complicated, but one thing, as an example, AI is really good at this, is like our friend Scott Wu, obviously, like his Devin and things like that, integrating messy systems. And this is what we've built Nexus to do. So Nexus and our co-founder, Rasmus Wisman, comes from a data... He was the top computer science PhD at Oxford. He won the prize for the best thesis, and he's amazing at building, simplifying complex messy data structures. And so Nexus, our AI platform, is built specifically for this use case. And so we think about how do we make this a better experience? [00:00:00] S0: Hold on. Let's not move on because I know what Nexus is. The person listening might not. Explain what Nexus is.
· David Senra, Alexander Taubman
[00:00:00] S0: Hold on. Let's not move on because I know what Nexus is. The person listening might not. Explain what Nexus is. [00:00:00] S1: Yeah. So Nexus is our horizontal AI platform that we've been building across verticals. And so we've been investing, and part of the reason we're in multiple verticals is I think scale's increasingly important in an AI world in general. You're competing for the best talent. Being able to leverage innovation and building a platform across 5, 10, 15 different scaled industries allows you, I think, every single industry we go into, it makes Nexus better. We add additional capabilities. [00:00:00] S0: Why does it make it better?
· David Senra, Alexander Taubman
[00:00:00] S0: Why does it make it better? [00:00:00] S1: So as an example, about 70, 80% of the infrastructure is shared across the verticals, and that's kind of like our platform where on one side we integrate with all the models. We're model agnostic, and we kind of use different models for different tasks, or even one task might be using several models just to complete one task. On the other side, you integrate with the data, the workflows, and the different software systems of the business, and you have to build that deep integration. So Nexus sits in the middle. A lot of that intelligent model routing, security, sort of agentic orchestration layer, things like that, that's reusable across all the industries. That last 20 or 30%, which is embedding Nexus into the workflows of the business, mapping the workflows, integrating with these different data sources, it's a very hard technical challenge. It's why you need a world-class AI team to be able to actually use AI in the real world because you have all these systems. [00:00:00] S0: Hold on. This is a really important part, too, because again, I hear other people talk about Long Lake and I'm like, "Oh, they don't actually understand what you guys are actually doing." How many different states are your engineers operating in right now? They're not fucking just sitting in New York.