00:46:29 · Michael Sindicich
If you look at some of the other public players, there's not many anymore in the travel management space, but the gross margins are more like a 50, mid-50s percent or something like that. And we wanna be, you know, really efficient and have tech multiples and tech margins and things like that. And so we were leveraging AIVA, which is the evolution of CATI, and it offsets more than 50% of all of our support interactions. And it's not that it offsets, it resolves. We say it's resolution. It's actually users are reaching out to it. It can respond in any language in seconds, twenty four seven, three sixty five, and it can help you out. And it achieves human level CSAT scores. And it continues to improve and grow. And so by doing that, our gross margins went from 50% to, you know, 70 something. And I think that helps us a lot because then we can take that profit, we can invest it back into better solutions, products, and keep building. So that's kind of the use case that a lot of people know. But then it's like throughout everything at the company. So if I swipe my card at a Starbucks, like, I can now reconcile it automatically. I know that I'm on a trip. I know that it's a merchant for meals. I know, okay, here's the three policies and GL codes that have to do with meals. But because Michael's on a trip, this is probably the meals while traveling policy. And this is the, you know, this is the approval workflow. And if it's within this and all these realms, I can automatically expense that transaction without making the employee do anything. So there's that. There's like an admin companion where I can go into the product and if I'm a travel manager and I can say, you know, tell me the average hotel booking from directors of sales compared to directors of marketing in New York. And boom, it gives you like immediate like spit out and you can, you know, analyze your program or it can give you ideas and benchmark and figure out how to drive more savings. So it's really throughout our products. And then the biggest one is we have this enterprise product and we have this product led growth product, but now we've launched Navan Edge, which is an entire travel agent, like, EA dedicated to you for individual travelers. So we always distributed our products by selling to a CFO or a company admin, then they launch it to the company.
00:48:44 · Michael Sindicich, Miguel Armaza
But now what we've done is a fully agentic individual traveler product called Navan Edge, where you can just fill out some some information about who you are, how you like to travel. You can connect your email. It'll load all your loyalty programs and tell you how to optimize them and how far you are away from the next status and tier. And then you can chat with it and it can book your hotels, your restaurants, your airlines, your hotels, your car rentals, and even experiences. So you can say, Hey, I'm in New York this weekend. What's going on? You can actually see concerts or different events and things like that. And it monitors your entire trip for you. So if it knows I like window seats on the flight but none are available, it can actually keep watching that flight and change me when a window seat becomes available. Or I can ask it to, hey, can I get a late checkout for the hotel? It'll email the hotel and ask on your behalf if you can do a late checkout. It's a really incredible experience. [00:49:42] S0: With Edge, is the goal to go after much smaller companies? Or actually, are you targeting the consumer?
00:49:51 · Michael Sindicich
[00:49:51] S1: Yeah. So we're targeting what we call frequent travelers. It's not a play of like a booking.com or Expedia or a consumer travel platform by any means. It's really for that VP of sales or that VP of marketing that's on the road a lot that maybe doesn't have an EA. And if we can deliver that and then to every traveler at any corporation that's a frequent traveler, we think that that's a really interesting thing. And so I'm sure that that experience will end up becoming the user interface as it keeps enhancing to the company travelers. But right now what it's targeting is not businesses, but actually individual travelers. And really savvy travelers don't like to use legacy systems. And so, you know, if we go to a company today that we're working on selling, you know, we'll ask them, what's your GL data? And then what's your data coming from your travel management company? And usually the data coming from the travel management company is a lot less than the GL data for the travel, really showing that a lot of people are what we call leaking. They're not using the system that you've deployed for your employees to use because they'd rather use something else. And we wanna capture that something else, which then also helps us capture the entire company. We use a combination of different models, and we've started building our own models. And the reason why is because if you think about we're in business for now eleven years, knowing everything about travel patterns, trends, users, delays, all of the information that we have access to. And if you ask, you know, a travel related question in the context of corporate travel to a traditional model, like one of the big models, they need to sift through trillions of data points. And so what that trillions of data points does is it makes it expensive, it makes it slower, and it makes it less contextualized. And so the reason why we've started building our own models and are leveraging more and more of our own models is because we can get answers much faster, much more relevant. And then lastly is for cheaper. You know? And like I said, it's all about the user and then comes the different revenues and profit points. So by doing that, we upped our own model usage by quite a bit. We continue to use our own models more and more and we'll use, other models as well and and when it makes sense.
00:52:17 · Michael Sindicich, Miguel Armaza
But, but it's really interesting that our own models are are are very, very powerful. And so so much so that, like, you know, we're so AI focused that our our CTO and cofounder, he has an entire team of AI engineers, he works on AI projects full time. It's the only thing that he he and his team do, and it's how they can distribute it throughout our products and our internal employees and, you know, our different sales motions and things like that. [00:52:46] S0: When it comes to the AI front and the consumer products, right, where are the limits? What is it that your models or just, I guess,
00:52:59 · Michael Sindicich
[00:52:59] S1: the industry in general cannot do yet? Well first I'll tell you what it can do and then it leads to what it can't do. So is you know, the way we deploy it, if a traveler is showing up to a hotel after midnight, sometimes that hotel will cancel the reservation because you didn't check-in. AI, our AI, will actually understand, oh, Michael's flight got delayed. He's now gonna arrive likely near or past midnight. It can call the hotel and ensure that that hotel... Or it does. Not it can. It calls the hotel and emails the hotel to let them know that you're coming late. Can you verify that you'll hold the reservation? Then it sends me an email that I'm all taken care of. I can remove that from my task list. I don't have to worry about it because as a savvy traveler, you know you're gonna get in late. Now you have to panic and call the hotel. It does that. It can do things like cancel a United flight, book a Delta flight, and apply an unused credit that you have banked. And like, that is such a hard thing, even for humans, be able to figure out how to do with all the infrastructure. And and it can do that automatically. It can retrieve hotel folios for you. It can reach out to the hotel, get the folio, then itemize it. Okay. This is parking. This is minibar. This is, you know, hotel stay. So it is phenomenal at that. What it what it doesn't do, today in a broad sense, I think, is it it doesn't provide that human empathy and the human interaction that sometimes people are looking for. And I think that sometimes people are continuing to willing to pay for. Like there's a reason why AI music isn't as interesting for whatever reason. Because a human didn't make it and you don't want to listen to AI music or an AI reader reading an audiobook or an AI person on the phone that's like, you know, you're talking to versus you wanna just talk to a real human. And so for us, like, we're able to monitor these different chats, understand, you know, the user sentiment, be able to orchestrate the handoff between AI and human. And I think that's like, if you boil it all down, it's really good at doing tasks fast when you don't wanna actually have a human or talk to a human, or maybe you can't explain yourself perfectly and you need that human to be able to help you out. And that's what I think, you know, we leverage it for.
00:55:24 · Miguel Armaza, Michael Sindicich
[00:55:24] S0: I have a question, it's okay if you don't have an answer. [00:55:27] S1: You serve Anthropic, OpenAI. Cursor, Ligora, Harvey. [00:55:32] S0: All those guys, right? [00:55:33] S1: Sierra. [00:55:34] S0: Do they, and you see their data, do they behave differently than legacy companies? [00:55:42] S1: I think the difference in the context of when we're approaching a sale with them is it's just a common understanding that they want the best solution for their travelers. They don't want to wait for, you know, they don't want them to have to wait with support. They don't want them to have to wait for it to load all these types of things. And so for us, it's just like if I if I know it's an AI company and we're walking into a sale, like, there's a very common understanding that they want the best technology, the best service. They don't wanna deal with clunky solutions. And, like, that's why I think we serve all of them. It's less for them, I would say, about, like, saving money. You know, they're all really well funded. They don't they don't go after, you know, necessarily, like, negotiating. Unless the bigger ones, they do. But, like, they might not have a travel manager, they might not have a procurement team that goes and negotiates hotel rates for their specific employees. They want something in, fast, automated, self managing, integrated to the ERP, set up the policies and be done. And it's an amazing thing for us because it's exactly what we were built for. [00:56:50] S0: Amex GBT just got taken private. [00:56:53] S1: Yep.