Navan, Inc. ($NAVN)

Ladies and gentlemen, please welcome to the stage, Navan's Vice President of Investor Relations, Ryan Burkart.

Hello, everyone. Good afternoon. My name is Ryan, Vice President of Investor Relations. And it is my pleasure to welcome you to Navigate. Thank you so much for taking the time to be here with us today in person and thank all of you who are joining us on the webcast. Today is all about giving you a glimpse into the future of where Navan is going, why we are positioned to win and how we are driving durable and profitable growth. You're going to hear from 3 of our leaders today beginning with Ariel, who is going to share our vision and why AI is accelerating that vision. Then we're going to hear from our President, 5 Michael Sindicich, who will take us through our go-to-market motion and all of the amazing momentum we have there. And then finally, we'll hear from our CFO, Aurelien, who's going to wrap up with some financials and how our model is durable and high return. We'll finish off the day with about 25 minutes of Q&A, the entire executive team of Navan is going to join us for Q&A. It's going to be a great day. But before we get into it, I have to draw your attention to our safe harbor statement, and let you know that we will be making some non-GAAP statements today, which are reconciled at the back of this deck, which you can find on our IR website posted shortly after this event. And now without further ado, I want to introduce you to our CEO and Co-Founder, Ariel Cohen.

So, so, so exciting to have you here. And the main reason we met a lot of you in the past, NDR road shows post kind of NDR's, but you never had the opportunity to see our community, and this is really all of these customers and prospects and suppliers and our team that are out there. So this is actually a special for us. This is us really showing you what Navan is really all about. So that's exciting. And the other thing that is exciting, we're going to go way more under the hood today. So you're going to get to know our business better than probably we've explained it in the past. So I'm excited about that. But I always like to talk about this kind of statement when it rains, Navan shines. And this quarter, besides COVID, I'm putting COVID on the side because this was off the chart, was the quarter with most interruptions that I've ever experienced since I started to be in this business, we had a TSA shutdown. We had 2 major, major storms in the U.S. Lufthansa strike and then Dubai Airport going down, which is a major, major [indiscernible]. And with all of these interactions, the first thing, we didn't see any decline in business travel. So we didn't see spike in cancellation. We didn't see -- actually changes to what we are expecting to see from a numbers perspective. But we think that this is unique for Navan and we think that there is a reason for that. We know how to support it. The orchestration between AVA, this is our entire AI framework, the ability to take more than 55% of the interactions with AI, but then combining it with human agents, you cannot solve everything through AI. So combining it perfectly, orchestrating it with a human agents creates a situation that our travelers are staying happy while they are in the middle of New York, where all of the flights get canceled. And what they do, they are not canceling their trade, they are rebooking it. You cannot do that if you're waiting 3 hours in the line, waiting for the airline to pick it up. But if we solve issues with minutes, you're going to stay with us, you're going to rebook your flight, you're going to continue. You're going to be happy and you're going to bring more customers in that will tell the story. Business travel is so, so, so important and we are able to support it no matter what is going on there in the world. Which leads to our mission. Our mission is really to make travel easy for every traveler by being the best travel agency on the planet. This is what we are building. We are really at the beginning when we are talking about the best level agency on the planet. And you're going to see us delivering more and more and more services to this industry everywhere. So you can see this previous quarter to the [indiscernible] needs to kind of [indiscernible] previous quarter numbers because obviously, we are about to announce a new quarter. But really amazing numbers, and I'm always pointing out to CSAT and NPS. This is what drives everything. We're now travelers, when our customers are happy. We see more bookings. Therefore, we see more revenue. You can see that we added a significant amount of customers last year. GDP got to $9 billion, and we are growing extremely fast, and we're going to continue to see this coming from us. And why is it? We talked about it in the past, but it is very, very simple. First of all, we think about our customers. Our customers think about where their employees are. They think about compliance, they think about policy. And the icing on the cake, we are saving them money. We are saving them in average 15% on their entire travel budget. And just to give you a perspective, a company in the enterprise space that is spending $100 million on travel, we'll save $50 million from the OpEx on travel. This is not by giving them some discounts by giving different booking fees. This is real money that their employees will spend less when they are using Navan. It is extremely significant for our customers. But the users, the travelers, the EAs that are so important on the actual thing on being here in person, they love it because they participate in the savings. It takes no time to book your personnel trip, 7 minutes on average and minutes or seconds to get support when something happens. Just to give you an idea, in 30% of the cases, a trip will get changed. And you don't want to spend days and hours of changing these trips. With Navan, you go, you start to talk with AVA, you are quickly changing it. And the last thing suppliers, and you're going to see a lot of them here in the conference, suppliers in the private industry. Usually, you have this fight. Google it, you're going to see endless amount of fight. In the case of Navan, really deep supplier of the relationship, and there is a reason for that. First of all, we have access to the best users for these suppliers. They love business travelers. That's the first thing. The second thing, we know them. We don't have just a connection to these suppliers. We know them, we know them at scale. We know them online digitally. This is super, super important. So airlines, hotels, rental car companies and so many other suppliers. So this win-win-win. This drives this entire platform, and that's the main reason that we are growing so fast. So I know that AI is a big thing. So we've decided to actually drill down a little bit more to AI because I know that all of you and definitely the market wants to better understand it. So we, in the past, took a lot of time to explain our connectivity. This is how we are connected to everything. If it is out there, Navan is connected to this. Now we are connected to this in real time online, which means that we don't have agents that are calling an airline in Brazil to get the right price. What we do, we are connected to this online. It allows us to later apply AI to this. The second thing, we are connected to the entire fintech layer. So it means that we are connected to banks, we are connected to aggregators. We are connected to the ability to run payments for travel at scale in the smoothest way possible, which allows our customers to get the best automated reconciliation, removing a lot of mass from them. So this is just the bottom, the thing down there that by the way, took us more than 10 years to create, and we are still creating it. So it's a never-ended story. But it comes an entirely [indiscernible] allows us to create that service, the 7 minutes that I was talking about. It takes no time to book -- sorry, to rebook you and many more applications that you're going to see us releasing soon. It starts with our own model. Today, 30% of our calls are actually running on our own model. A month ago, it was 20%. So grew 50% in 1 month. But it gives us way better service, which means that you're going to see it in more deflection. Later, you're going to see it in gross margins. So as we are using our own model, you're going to see us becoming more and more efficient and more and more relevant, but another important thing, it will be cheaper, right? And you're going to see us over the years going directly to our own model. The second thing, Navan now, and it was a 3 years experience, is a complete agentic platform. When you are booking a flight, you are using our flight agent. When you are booking hotels, the same. When we've developed our restaurant capability, which is new, completely agentic. And on top of all of this, if you look at AVA, AVA is kind of an example of kind of a super agent that supports all of these other agents and know when to call it an agent. But where this is very important, and this is [indiscernible]. The ability to know seamlessly when we are using a human agent when you are using somebody that works in operations, when we are using an AI agent and make it seamless to the user, that's the entire magic here in this platform. So the orchestration layer, I would argue that's actually the magic of everything because creating agents, a lot of people can create it. Creating a model, this is very unique to us because we have our own data. Connecting to all of this, nobody is doing it and super, super hard to do. The orchestration of all of this. This is the secret sauce. And how do we present it? We presented in something that now we call Navan Classic. This has been even that most of our customers are using. There are agents that are already introduced there through a functionality that we call book with AI. But the place that we are really showing off with these agents is in Navan Edge. I will, for the first time, show you a demo of Navan Edge. And very soon, you're going to see it in new products. One of them is LSTMC. LSTMC is us delivering us everywhere, no matter how you want to consume it. This is not some -- I know that a lot are talking in the industry about [ headless ]. This is not some theoretical project. This comes from our customers and partners. Some of our customers, definitely the more cutting edge would like us to deliver Navan not through AI, but to a headless functionality, same as partner. And really, I would say, the future, and we are really at the beginning there. And [indiscernible] and I keep having this debate if it's the future ready like in a week. So we will continue to have this debate. I think it's in 5 years. Aurelien think it's in 5 days, but really [indiscernible] which is basically an agent that we'll just figure out everything for you automatically. And all of this together, this is really Navan in a nutshell. So what does it give us. First of all, the ability to accelerate, to accelerate sales because of our AI functionality, more and more customers in an accelerated pace are joining us. You saw we said last in Q4, we told you that we grew GBP by more than 50%. And we think that this kind of momentum will continue, supporting with Ava and continue to raise the inflection towards the 60% to 65% to 70%. That's something that you'll see from us. This alongside the conversion of [indiscernible] into the platform will show you a completely different gross margin profile from Navan. And then the delivery. I've mentioned it in the last call that AI accelerates us as a tech company. We can now develop more faster, policing more stuff, things that were in our road map for the next 5, 6 years are on this slide. We talked about [indiscernible] It's now on the slide. We talked about Navan Edge. It's now on the slide. And you're going to see so many things coming from us because you can actually develop stuff faster. And when you do it on our platform, you get this entire travel agency delivered through so many things. So we are definitely, definitely excited about that. And now probably the most exciting thing I'd like to do that, but this is a recorded, maybe next time I will do a real one. So I'm going to actually show you what Navan Edge is all about. I know that I shouldn't have clicked on it, right? Okay. So this is actually Navan Edge. And here, I'm actually booking the trip to this place to navigate. And what you're going to see is that the system figuring me out, they know that I like United. They know about the conference, the dates everything. They know the destination of the conference, and they're allowing to orchestrate the entire trip. Book my flight. The system will tune to United because I'm a united global service. Navan Edge is automatically connected to a wallet that we are creating automatically. So we knew that you are United Global Service. We know that you are [indiscernible] gold. We know all of this automatically, and the trip will completely aligned to this, to your past trips to everything that we know about you in our memory. So super, super easy to book. And what you can see is this combination of a ChatGPT kind of experience with real life things that you need, for example, selecting your seat, for example, selecting the right flight for you. So I actually think that we are the first 1 to figure out how to combine chat with real life objects to actually do something that makes sense for you. So this is actually an example of a flight, and now we're going to see hotels. The system will recommend the 5 hotels that are relevant for me based on what I care about. I care about loyalty. But based on this location, it would even highlight the fact that there are events right now in New York, so book it fast. We'll give you your Marriott point. We'll give you a band point basically really, really driving loyalty here. And I think where it comes super, super cool. You'll see it in a second. I had an extra day last night, and that was something that I planned last week. So I'm asking, what can I do? I have a free evening on the 12. What can I do? And the system will actually -- now you can see that we recorded it last week, we recommend to go to the [indiscernible] game, but the NIC's game didn't happen because we won last week. So of course, no 5 game, but you can see the point that it will know the event, it will suggest to book them. Instead of actually want to go for a restaurant. It will book the restaurant that is relevant for me in this area, in the location of the hotel based on the type of restaurant's that I'm booking all of this in my face. And the last thing proactive summarizes everything the day before the trip. This is my trip. This is my loyalty. This how long it will take me to get to the airport. This is everything that I need to know about this 3, and it will keep updating me. Keep updating me means that if something has changed about the flight, it will not be a flight notification, [ Ilan ] and I used to call it, [indiscernible] notifications, basically your flight got canceled and then what? It will be -- this is the notifications [indiscernible] about it. This is how we're going to change your flight. This is how long it will take you to be on the TSA line, everything that you need to know. So this is really an advantage. It's cutting-edge, it's something that we're super excited about, relist months ago, see good traction there and so many good things are going to come there, which eventually brings me to what we are trying to do. We're going to be everywhere. We're going to do everything. We're going to do it globally. And if you want to be there and it's important for you, we're going to be the agency for you. So with that, I'm going to invite Michael, who will actually explain to you how do we sell Navan. So thank you.

Great to see a lot of you guys again and the people on video, Hello. I'm Michael. I'm the President here at Navan. I've been at the company for a little over 10 years now since we were a little less than 10 people. So I've been along the journey for a while and kind of know a lot about the inner workings and how we do things. I'm responsible for basically how we go to market at the company these days, changes everyone. But as the team has mentioned and we all know this, it's a massive TAM and a big opportunity for us to take that we are at the early stages. So what the team asked me to do today is walk you through a little bit about how we go to market and how we think about continuing to penetrate into the industry. You've probably seen this before. But as we kind of talk about how we attack customers, right, we really think about it in 2 different segments. So there's what we call managed business travel. So companies that have solution today and unmanaged business travel. And the way that we think about this is typically the larger the company, the more likely it is that they have a managed travel solution today. The smaller the company, the more likely it is that they don't use something today. And so when we attack managed business travel, typically, it's through our go-to-market channel called SLG or sales-led growth. Today, that's for companies that are more than 300 employees. So we deploy a salesperson. We've got 3 different segments: commercial, mid-market and enterprise, and we'll get into how we distribute the sales team across the world. When we think about unmanaged business travel, sometimes it's sold by sales-led growth because we have that 300 employee and above goes through SLG and we might find a company that has something, but doesn't really use it or doesn't have anything at all. That's kind of the DIY, and we do see mid-market and commercial companies sometimes that have just scaled really fast and didn't have a solution in place. And then when we think about the SMB or companies that are below 300 employees, typically unmanaged, and we attack that through our PLG motion or our product-led growth. So sales-led growth is a more traditional top-down approach. We go deploy a salesperson, do some account-based marketing. We go target the CFO, sell the CFO or the travel manager, the expense manager, and then they deploy our products throughout their employee base. On the PLG side, it's more performance marketing ads, TikToks, Instagrams, targeting admins who will sign up, create a policy, self-serve and start using our product to manage their travel and we'll get into that as well. Edge, obviously, which Al talked about, is much newer. That's our way to get to frequent travelers. They might have -- they might be at a big enterprise company and hate their solution, so they can come and use our product. They might be at an unmanaged small company and need something to help with their business travel. When we think about how we separate the teams or how we go to market across the globe, we've got these major offices, san Francisco, Boston, New York City, Paris, London, Berlin, Boston was a new office that we added earlier this year. And that one we see really exciting things coming from that office and we'll keep -- continue to scale it. And the breakdown from the sales team or the go-to-market team is above. So I don't think -- sometimes they get questions about how do you distribute the teams, what positions. So we wanted to give a bit of a breakdown here. There's acronyms, by the way. So it's typical of a traditional kind of software sales team where you have an SDR sales development rep. The AE is the account executive that's actually doing the demoing and the selling. We have solutions architects when you need to get into the details of ERP integrations and really complex travel nuances. On the marketing side, right, we do field demand gen product marketing and brand. On the operations side, this is really revenue or go-to-market operations. So you can see the functions here. And then in account management, the main functions are customer success, so managing and implementing that customer after it's been sold. And then consultants or solutions consultants that also have specialization in the various products that we sell so that they can implement and get into the details there. One really cool thing that we've been doing is product-led growth allows us to scale globally and test out how our products resonate in the market. So it's as simple as saying, okay, let's translate the product. Let's make sure it works through that market. Let's go deploy performance marketing ads. And if we see a lot of success there, great. Now let's go put a top-down sales team in that region and continue to grow. So we expect to continue to scale globally. PLG is kind of everywhere in attacking different markets all over the world. When we think about the sales-led growth sales process, [indiscernible]. He's done an incredible job to build a repeatable sales process across the globe in all the various offices that we operate from. And so I can talk to a rep in Paris and say, where are we with this deal? And if they say we're in CI, I know we're identifying a champion. I know we're doing DISCO meetings. I know we're starting to do demos to various folks. But the essence of this is basically, we have stages for every single sales process. The larger the customer, typically, the longer the sales process is. The smaller the customer, typically the faster, and we can get through these gates really quickly. So our sales cycle on the SLG side is roughly 2 to 9 months depending on, again, size. Once we sell that customer and sign the customer, then what we do is we start launching and implementing that customer, we launch it and then they ramp up until they get to full ramp or terminal velocity. And then during that launch process, whatever we haven't sold them during the sales process, we start the upselling motion. So maybe they bought travel, "Hey, you're implementing, why wouldn't you just use our travel payment. Or they bought travel and travel payments, why wouldn't you use expense? And also, hey, let's talk about our VIP services. By the way, now that we're working with your team, do you do meetings and events, and we can align with that schedule and start selling our meeting and event service. So that's kind of the sales-led growth approach. The interesting thing is what we invest in sales and marketing today takes time to show up in the actual financials because we need to sell, launch, ramp that customer until they get to terminal velocity. So I think Aurelian will talk through some of that as well. When it comes to PLG product-led growth, think about that similar type of sales process, but things happen in days or hours and not actually weeks or months. And it's very, very AI-based. So it's all performance marketing ads. It's not field events. It's all lead targeting, it's GEO, it's SEO, and it's really smart, personalized onboarding for the types of people that are signing up for the product. So we know your industry, your field, we can organize the product in a specific way. We can handhold you through our AI help to be able to onboard your company, create your policies, will suggest policies for you, add your users, integrate your HR system, turn on your SSO and really be on your ramping. And we've talked about a few case studies or stories where a company like Anthropic might have signed up through our PLG motion. And what happens is as they scale, we then graduate them to our sales-led growth account management to then work on continuing to manage the program and upsell more and more products to them. So it's a beautiful motion actually for the company to capture these fast-growing companies that then we can bring on to more managed services. And I think one thing that we talked a lot about during the last earnings call is our RFP volume. So typically, a mid-market or an enterprise company will likely do an RFP, request for proposal, where every couple of years, they reach out to a bunch of companies and say, "Hey, tell me about what you do. I'm going to learn about this space because I'm in procurement. I don't study travel all day long, and I want to see what else is out there and if I can make improvements to my program. So we get these RFPs. And sometimes companies are doing it because they might want to negotiate a better fee with their incumbent. They don't have a real reason to switch or it wasn't in their head that they were going to switch. But a lot of times, it is we're going to make a change, and we need to see what else is out there. This is a very, very interesting leading indicator to see if there's more activity and more sales that we can be driving by the growth in RFPs. And I think over the years, as we've continued to scale globally to go upmarket, to create more buzz about the company and have more active champions, more people are reaching out to us to understand what we do, which gives us more [ at bats ] to go and sell them. Why are we getting so much more RFP volume these days? I kind of break it down into a few reasons. Number one, we are the AI-led T&E platform in the space. And all travel managers, expense managers are getting orders from their leadership to start leveraging AI to be able to drive more efficiency in their operations. And that means that you must look at Navan. We see a lot of consolidation in the space. You guys obviously study the industry. I don't need to tell you what it is, but there's a lot of consolidation, which then means, okay, I need to see what else is out there because I don't want to go with whoever just got bought or this or that. So when there's changes in the space, usually people are opening up to see new solutions. I talked about our support and product continues to improve and move us up market. For those of you who are on video, you probably don't see this. And for you guys here, hopefully, you do see it. But we do really interesting events like this. So today, we have customers here, 320 at Navigate 2026, 320 customers that spent last year, $2 billion of GBV on our platform. We have 120-plus prospects that are here with these 320 customers that are talking, having drinks, doing breakout sessions with us, sharing about their program. We had on stage earlier, Ofer, our COO, was meeting with -- or had an open conversation with the travel managers at GE Healthcare, DoorDash and MUFG, talking about the success that they've had with their programs. And when the people who are prospects that we're trying to sell to are sitting there eating this up, talking about how amazing Navan is and how great it is, that just helps us sell more. And so we see that word of mouth really, really getting out there, which then kind of goes to the integrated marketing that we're doing. And I definitely think that there was an IPO and a brand effect on our name, which got more people to pay attention to us, which is a great branding event and something I'm really happy that we did. If you look at our customer base, you guys have seen the slides, but relatively diverse across the industries, we'll break it out for you, I think for the first time in a very -- ever probably. 38% of our revenue comes from outside of the U.S. So international customers, so good diversity there. No single customer makes more than 2% of our platform revenue. So there's a good distribution there. This is the breakdown of our revenue from the customers from different industries. So tech was the big one. If you were to ask me 5 years ago, I think the tech portion would be pretty big because new start-ups buy from tech companies and they buy from each other. But you can see we're expanding quite a lot more into professional services, manufacturing, finance, media and all other industries. And that will continue to diversify as we go and as I look at the pipeline coming up. So the other piece is, and Ariel mentioned this, but we are becoming more efficient. One of the biggest metrics that Grant and I and Eric are always looking at is how much sales and revenue, does the salesperson generate now versus some time ago. And it's called productivity of an account executive. So if I sold $1 of revenue last year, how much do I sell this year? That year-over-year AE productivity is up 50%, which it's hard -- no one can plan for that it's really interesting to see such a big increase in productivity. And so why is that happening? Again, we are investing more and more in the enterprise segment. So moving upmarket, which has very strong momentum, and it has a very high return for us. We expand more efficiently. So I don't need to send -- start a big office and get a bunch of salespeople and management layers and SDRs and open up a region, I can test it first with the PLG motion and then decide if it's a good place for us to go. We're leveraging more and more partnerships. So in a lot of our sales today, it's not just our salesperson selling. We actually involve the ecosystem to be more efficient and to be able to penetrate into the account. And we talked about the brand impact. And the other thing that we're spending a lot of time doing is actually setting up the deal so that it can launch and ramp faster. So what we do is we start bringing in account management earlier in the deal cycle, to make sure that the customer is ready to go, we've aligned everything because the faster that we can launch the account, the faster we can get revenue onto the platform. So at the end of the day, we have all these inputs, we've shared a few things. Ariel mentioned this, but in Q4, year-over-year, we sold new GBV more than 50% than we did the year prior, which we think is a very, very good sign. We see the explosive RFP growth. We said 200% plus. Our average deal size and our win rates are increasing year-over-year, which is great. And when it comes to the PLG side, our revenue more than doubled in FY '26, which is a really good sign. We're on to something here, and we will keep on scaling and investing in those products on top of the launch of Navan Edge. So we think all of these things create a really nice flywheel effect for us to continue to scale quickly and grow our revenue. So with that, I think I'll pass it to Aurelien to get into the numbers. Thank you.

Hi, everyone. Very glad to be here. I'm Aurelien Nolf, the new CFO of Navan. I joined the company 2.5 months ago, not 10 years ago, like Michael. And I've been like amazed by the pace and the momentum and the energy of the team, like so far, it's been a really amazing journey here. So we are hosting this amazing event for our customers today. And I thought it was a great opportunity to also bring the investor community to the table to just see this momentum and give us a chance to also elaborate on our story and our business and how we are making money, right? At the end of the day, this is why we're here just to tell a story and explain you what Navan is made of and why we are confident in our future in our growth. So I'm going to start with the very basic, like the model itself. How do we make money as a company. It's a very simple equation when you think about it. We have gross bookings volume, so people booking on our platform, and that was a record high of $9 billion in fiscal '26. We have a usage [indiscernible], 7% in fiscal '26, and that gives us a usage revenue. And on top of that, you can layer on some subscription revenue. That gives us those $700 million, another record high in fiscal '26. What's really great is we are monetizing across the board from our customers, from our suppliers, from our payment providers, across travel, payment and subscription. So a very, very diverse model. But what I think is not pretty well understood is the fact that most of this revenue is usage-based. And that's very important because what's at play here is not a SaaS efficiency model, right? The revenue we are generating is going to scale with the activity of our customers. And that's very important because that's what incentivizes us to keep innovating, coming up with the best platform and the best experience for our customers so that they book more travel on our platform, more and more customers are joining the platform, and we essentially make more revenue at the end of the day as a company. So that's really what helps us protect our moat in our business. So looking at the way we generate that revenue, I don't think it's also very well understood and I wanted to take a minute here to really explain that. What's super important is, and you can see here this dark blue part of the pie chart is the supplier revenue. That's actually the largest part of our revenue. That's how we made the most [indiscernible] as a company. Think about this as all the commissions we are getting from our suppliers, our partners, airlines, hotels, every time we generate some bookings on the platform, we get some revenue from our suppliers. And that, again, is the largest part of our revenue. Then obviously, we are monetizing our customers as well through 3 Ps and platform access fees. We are also generating subscription revenue, mainly on our expense business, and that would be on a per seat basis or on an annual basis depending on the customer. And then lastly, we make revenue on our credit card payment business by gaining some interchange revenue net of the rebates. So I think the really, the point I wanted to make here on this slide, the most important takeaway is that the majority of the revenue comes from our suppliers, and not from our customers. And that's why it becomes very interesting. It becomes very interesting because not only our revenue is diverse and we're very strong, we also have a very high visibility, which also is something I don't think is well understood. When we start the year, we have roughly 90% visibility on the revenue we're going to generate during the coming year. And there is a very easy reason for that, like 3 buckets of revenue for FY '27 fiscal year for the guidance we just shared with you. The first one is the installed base. We have a very strong installed base, thousands of customers that have joined the platform over the years and that are very predictable, repeatable patterns from a usage perspective. We see their bookings coming at a very, very consistent rate. And the reason is corporate travel has been very resilient over the years. Like companies need their teams to get together. They need to get to this sales meeting, they need to close a deal and people are just traveling a lot. And I mean, Ariel mentioned even in the moment when we see a lot of disruption in the world, travel has been very resilient. So that's just the installed base loan that gives us a very, very high visibility on the revenue we're going to get during the year. And on top of that, we add the customer we signed in fiscal '26, right? So those large customers that have been signing a contract with Navan this SLG population of customers, we know how fast they're going to ramp, and I'm going to come back to this in a second. But that gives us to the 90% visibility on the revenue. And then on top of that, the sales team is going out there and we keep adding more and welcoming more customer to the platform, but most of that revenue is going to show up in the future on our P&L. So again, 90% of predictability, that's why this usage-based model is so powerful. So now let me show you how that compounds over the year. What I think is very interesting on this slide is to look at the different cohorts depending on when those customers join the platform. What's really amazing is to see that the difference between a cohort when they joined the first year and the second year is 3 to 5x from a revenue perspective. So the customers are ramping through the sales motion. They are ramping up through the year. And again, in year 2, their revenue is going to be 3 to 5x higher than in year 1. And so you see those lines that are stacking up here year by year after year after year. And that's where the model becomes very, very interesting. We are lifting every cohort ahead of above the last cohort, and the reason is we are signing more and more of those very important enterprise customers that have very large T&E budget and generate a lot of revenue. Adding on top of that, our net revenue retention of 107% in fiscal '26 is adding to the flywheel and keeps improving the revenue for every cohort. So that's why we are very excited about seeing those new customers joining the platform because that gives us this huge amount of compounding revenue on our P&L. Now going to the gross margin. So we are already operating with best-in-class operator gross margin, 73% in fiscal '27. That's more than 1,000 basis points of margin expansion over the last couple of years, which clearly is coming from our technology leadership. Ariel described how AVA has been making us very, very efficient in terms of supporting our customer, and that's very efficient from a P&L perspective as those gross margin rates are showing here. But this is not a [indiscernible]. Like we are not done. We see a lot of room for gross margin expansion in the coming years. There are 3 main reasons for that. The first 1 is we're going to keep leaning into AI for our support and for -- with AVA, right? So we are automating the super service. We get great CSAT results every time AVA is jumping in to help our travelers. And so that's going to keep us being very, very efficient as we provide support to our customers. Then Ariel also mentioned the proprietary models we are using from an AI perspective, and that's very powerful. because today, 30% of the volume is coming through those models. I think he also mentioned that it was just 20% a couple of months ago. So you see the adoption rate of our proprietary model jumping very, very fast. And the reason why that's very interesting, not only from a customer perspective because they get faster and more accurate results. It's also very interesting financially because the cost is way less than paying tokens to future models. And then the last input to this equation is the migration of our [indiscernible] customers from their first-hand platform to the new Navan platform, which is going to increase our gross margin going forward. And that gives me a natural transition to the way our channel mix works here. So we've been -- on the last earnings call, we've been able to report amazing growth rate on the Navan core platform from an SLG and PLG model. We reported 40% growth year-over-year of the SLG revenue, 100% growth of our PLG motion. So we doubled the revenue on the PLG. But we also reported that the [indiscernible] customers, this revenue base is only growing in the low single digits. And what's great is we expect that to keep being the case in fiscal '27 and beyond, which is going to give us further tailwind from the gross margin perspective. So the durability of our top line growth and the margin trajectory here gives us a lot of confidence in our future, okay. So now let's look at our sales and marketing expenses and why we are very, very pleased with the payback we are seeing every time we deploy $1 of money in our sales and marketing motions. I mean if you're here in the room today, you can see the excitement from our customers here. That's what we do with our marketing expense. But what's really important for you to understand is the disconnect between the moment when we spend the money and the moment when the revenue is showing up on the platform. So Michael mentioned that most of the spend from a sales and marketing perspective actually is incurred around the time when we signed -- around or before the time when we signed a new contract with a customer on the SLG motion, right? When you think about it, we send our SDR or AEs to sell our products, then we pay a commission when the deal is locked. But we won't see any bookings on the platform usually to be -- for 2 months after they sign the contract, right? So there's already this natural gap of 2 months. And then after that, as I mentioned, customers need time to ramp and book all of their trips to -- on our platform. And that's usually something that takes anywhere between 5 and 6 months. So -- I mean the underlying unique economics we are seeing once the customer is fully ramped, are really amazing and they are very strong. And that's why we see this revenue compounding over time and those cohorts of customers stacking up every year. The same is true with our PLG motion, right? It's the exact same process, but the time line are compressed. It's way faster, right? So between the time when a customer is hearing from us on our marketing channels, there's only 7 weeks already before they are fully [ rented ]. And so this is a very short payback that makes this sales motion very, very attractive from a financial perspective. So in summary, from a sales and marketing perspective, we have a lot of front-loaded cost, the revenue is compounding over time, and we are getting more and more efficient as we are scaling that business. Now let me touch briefly on our research and development expenses. How do we spend on these? How do we create this amazing platform that is Navan? Because here, the productivity also matters. I know Michael walked us through the productivity of our sales team. But the same is very true with our R&D team. Our incredible team has been delivering more than 450 products last year, products and new features that have been launched across our travel or payment in our expense business. And we're doing that with a very disciplined budget. 14% of our revenue was spent on research and development, which is really cutting edge from a R&D perspective and new products. We've been releasing a lot of new features and a lot of new breakthrough products. In case of Navan Edge, it was very fun to see the demo from Ariel earlier today. The [ expert ] chat agent. We had the team here on stage this morning showing to our customers how this can really transform the way people process expenses, travel cloud or also our audit agent. So I'm trying here to put it all together, right? What are the 5 compounding levers that are going to help us really grow our profitability -- grow our business, expand our profitability and grow our free cash flow over time. So first of all, our ability to upsell and increase the attach rate of new products. When you think about products like cx Pro or VIP service, meetings and events, the payments and expense business or leisure, all of that are not necessarily something that our customers are using from day 1, but that's something we usually upsell over time. And it's an ability for us to increase our margin from those customers. I've discussed about how confident I am in our ability to expand our gross margin as well through the usage of AI and through AVA or chat and superbot. And then from an OpEx perspective, operating expense perspective, I also see a lot of opportunities for us to generate some leverage over time. So how are we going to do that? Where we are seeing a lot of momentum from a sales and marketing perspective, and so we expect to see a lot of efficiencies there and keep lending more and more of those very significant enterprise deals with those customers that have very sizable team and budgets. We're going to deliver some G&A leverage, right, through our infrastructure that is fixed. We have done a lot of investment to become a public company in the last couple of years. But that's something we expect is going to generate efficiencies going forward. And then I just mentioned our disciplined approach to research and development and how AI is going to help us doing that in a very, very way. So just to be clear, those 5 different levers are all additive, right? It's not like we're not going to select between one of them. That's what it's going to make our path to margin expansion and free cash flow generation very durable. That gives us the clear path to the future of Navan. And so here, before we start the Q&A, just wanted to bring everything together from the beginning of this presentation to the end. What are our goals, what are our mission here, right? First, Ariel mentioned very important. We are a travel agency, and we want to provide the best travel experience to our users, to our travelers anywhere in the world, everywhere they are. Super important to us. That's #1 mission. You have to know that every time we have a conversation as a management team, we start with our NPS and our CSAT score. That's #1 very core to the mission of this company. Number two, we have a huge opportunity, a huge market. Michael mentioned this $185 billion addressable market. That's the size of the pie. And obviously, as a business, we are only scratching the surface today. And how do we do that? Well, we have multiple revenue streams that are driving this durable growth with high visibility into our revenue for the years to come. And then lastly, we're going to do that very efficiently through a very, very efficient and high-return go-to-market engine that has very attractive paybacks. So I'm very excited to be here. Again, it's been an amazing journey for me over the last 2 months that I have been the CFO of this company, but I'm really looking forward to everything that's ahead of us. So thank you for your time. And with that, I will ask the management team to join me on stage for some Q&A.

We're going to jump right into Q&A here in a second. A couple of ground rules. We're just going to take questions from folks in the room. We have a couple of mic runners on either side of the room. So just raise your hand. If you have a question and say your name and where you're from for the benefit of the folks on the webcast. While everyone is compiling their questions, I am going to get us started here with a question for Ilan. It's the first time we've Ilan on stage. So welcome excited to have you. This is a question I get all the time from investors at conferences, on calls, and it is -- what is it about Navan's infrastructure and technology that is really unique, really proprietary and that acts as a moat versus existing competitors and would be competitors that are thinking about coming into the space?

I think to answer this question, it actually first starts with -- just with our back end. Ariel spoke about it earlier. Agents can do nothing without tools. And -- okay. There's some operations [indiscernible] it really starts with what we've been working on for more than 11 years now. Now we do not plan to serve our agents when we get some [indiscernible] So I get distracted very easily, I guess. We spent 11 years building a lot of tools. And as soon as the ChatGPT came out, we jumped into it really, really quickly. And in [indiscernible] -- so 2022, November is when ChatGPT came out in March was the first time that we had a proof of concept of an agentic framework that we built internally. We built it because nothing else existed back then. We call it cognition. And it's allowed us to deploy agents in a mission-critical area like booking flights, upgrading your seat, talking to you about the refund. So areas where the cost of an error is huge, huge. And the only way to do it was in a reliable way. And LLM are definitely back then had so many hallucinations. It was so inconsistent. It was biased and so we built this agentic solution. And fast forward to today, we are in a very unique position because we have been having agents -- having conversations with users, real conversations about travel with tool usage since May of 2023, we today have a very, very unique data set that no one else has that we use this data set to train our own models to deliver a more accurate experience because it is focused just on booking a flight or just on booking a hotel or just on this determining whether the answer that Ava now gave is within our SOP, standard operating procedures or not. And this is really important if we have employees or agents that provide answers that violate our SOPs, we fire them. So it's really, really critical. It's amazing because now that we have the data, we train the models, we get something that is largely 6x faster than frontier models, and we still are learning and seeing and we improve it. It is much lower latency, as I said, much higher accuracy and definitely much, much, much cheaper. It also, I would just give you like something to think about. But now that we have our own model, it opens the door to even further enhancements and optimizations in the world of interpretability, which is something that we are investing now in and it will allow us to fine-tune the model to perfection. We can talk more about it in some other conversations, maybe what super, super exciting. All of this together, I think.

Super, super helpful, Ilan. And I would just note that Ilan is doing his AI keynote right after this. So if you can, I would highly, highly recommend sticking around for that. All right. So let's go to the audience. First question over here from Samad.

[indiscernible] I think a lot of the analysts in the room can definitely relate to all the travel. And I think Navan Edge is really cool as you think about it as a user experience. We all take a lot of the same trips every year and to be able to build that knowledge and be able to easily ask based on preferences really resonates with me as a business traveler. And so I understand the value of the experience and how it differentiates you from a user experience at this point. How should we think about what the opportunity there is from a monetization standpoint, right? Is it just that it increases either the velocity of GBV. Is it something that allows new monetization opportunities with suppliers where you get a different profile of the type of customer? Just help us think what the value there is.

I think, first of all, the first target of Navan Edge is this frequent traveler that tends to go home [indiscernible], Delta 360, all of this. [indiscernible] as this tendency to do whatever they want to. So they can be in a managed environment, but they'll go and book somewhere. In Navan generally, what we call the [indiscernible] is pretty good on capturing them, but not all the time. Our competitors never capture them. So we think that there is a great opportunity to hyper, hyper service to really get to know them on the level of loyalty, on the level of their behavior. I always like to give this example. I fly a lot from SFO to London. And I'm certain [indiscernible] certain BA flight, and there are 2 types of flights. And don't put me on the other flight. I hate it. It's certain configuration of business class [indiscernible] amazing. You come to any system or to any travel agent and the travel agent will basically tell you that flight is not a variable. Navan Edge will continue to work for you on this flight or will book you on the other slide, continue to offer you on this flight until I'm going to get you this flight. For the top [indiscernible] for the ones that really care about travel [indiscernible] and above, this kind of use case is the early go. Now if I can optimize your mileage while I'm doing that because I'm automatically connected to all of your basically credit awarded. If I can give you extra rewards by Navan, this speaks the language of that kind of [indiscernible] the frequent travelers that always do whatever is good for them. We view this as the huge opportunity, the ones that we make most of the money from. They tend to be very loyal while you've got their loyalty. And we think that only with AI, you can give them the service that they want to. Of course, you can create some cutting-edge VIP travel agency, very hard to scale. With AI, you can scale it. This is what Navan is targeting. I think that Navan Edge will become the future of travel. I don't think it's the future of corporate travel. I think it will become the future of travel. We need more iterations, you need more technology development there. But I'm -- actually, as a product guy, I'm putting everything that I know, everything that I got in this product. I think it's the future, the ability to later delivery maybe through [indiscernible] there are already features in Navan Edge that are delivered today in Navan Classic. We kind of embedded them in the platform. But then to deliver it to everybody through a headless platform, I think we'll make Navan without a doubt, the #1 in the market, an idea and say which market you can imagine, which market I'm talking about.

Maybe just from a monetization perspective, I would add that this is -- I shared today for the first time that the vast majority of our revenue is coming from suppliers. And that's how we would monetize Navan Edge that's mainly from supplier commissions.

Jared Levine with TD Cowen. I wanted to talk about the current state of travel cost inflation. Can you talk about how that will impact your usage yield as well as revenue? Is it a tailwind or a headwind?

Maybe I can take that. So yes, we're seeing some decent inflation in the market right now. In fact, since February, we've seen the flight on average the cost of a flight going up double digit, right? So there is inflation in the market. It's more -- I think we've all seen that it's more and more expensive to fly from one place to another. Given the way we make money, that's good for us in the short term, like it increases the commission we are gaining from suppliers. So that's a tailwind. So that's one way to think about it. And so far, we're not seeing any meaningful price elasticity from [indiscernible] perspective. I mean companies are craving in personal interactions, right? And we keep seeing that with every customer we are onboarding, those customers are very focused on bringing the teams together on meeting their customer face-to-face to close a deal. And so business travel has been very resilient. So is there -- is that yes, -- probably not. But so far, we've seen a decent amount of inflation and no real impact on demand.

Chris Quintero from Morgan Stanley. It's been really helpful. I want to ask about the ramping of the recent customer cohorts that you mentioned, really impressive to see that accelerating. So just curious from the go-to-market implementation standpoint, what are you all doing that's really driving that faster ramp than new cohorts?

Yes. First of all, it comes from a realization that we do not make money as a business unless our customers use our products. And so that's actually number one. We also have a payment mechanism for the sales reps who sell deals where they make some of their commission upfront on the signature, but they make a very good portion of the commission once the account is lodged. And so that incentivizes the sales rep to essentially do everything that they can to make sure that the launch happens fast as fast as possible so that they make their full commission on the deal. We also implemented what we call launch 2.0 internally, where every time that there's a deal, it needs to be validated by the account management team. So the account management team gets brought into the sales cycle. They meet with the customer, they talk about implementation. We align on an implementation schedule and what needs to be done. And that account manager is actually validating the GL of that accounts data, how much we expect them to spend on travel, that the deal is not doing a pilot. We're doing a full launch, what time we expect to launch the U.S. and then Europe and then APAC. And so we're aligning all of our incentives as a company to basically make sure that the account is first and foremost successful; and second, launches as fast as possible. So those are 2 other things, but it all really stems from like we're not going to make a dime if we get a signed contract. We have to get the account launched and operating on our platform, and then that account will help us be successful and go sell us to more companies.

That's something [indiscernible] in the PLG segment, if you could see that it's becoming significant, this is all AI. So we basically took something that used to be human intensive. And we are not talking there about SMBs with 10 employees. This is 50 to 300 people. 300 in some other organizations will be considered enterprise. We know how through AI to sell to them, to onboard them, to do customer success to them. So this completely accelerates the entire thing, the entire launch fast to get and also makes it in a cheaper way. So these 2 together are keep improving basically the average of time to make money.

Jed Kelly, Oppenheimer. Excited to see some of the product demonstrations after. I think you had a stat that your RFPs are inflecting up 200%. So can you talk about the underlying drivers there? Are you now involved in every big travel RFP that comes up. And then can you give us the reasons like what happens if you don't win? And then when it come up maybe 3, 4 years later, how those conversations are going?

Yes. So I talked about some of the drivers. It's brand impact. It's customers saying, "Hey, you really got to look at these guys. They transform my program. They made me a hero internally and now I'm promoted, right? So it's a lot of that. It's the scaling of the company. It's -- the narrative now is that, oh, actually, Navan can support a Visa or GE Healthcare globally. And by the way, they launched me in just a couple of months globally. And so that type of narrative has shifted in the industry. We were not in every RFP. And the person who told us that ran a consulting firm for these big travel companies, and she was helping run RFPs for them. A woman named Kim Hammer, we just hired her and said, "Okay, you know all the big travel companies. You know the consultants, we're not in them. She said we were in about 1/3 of them. So we hired her. And we're hiring more Kim Hammers, and we're getting out there at the local chapters and the GBTA and we brought Navigate -- our big event. So I'd like to think that we are in many, many, many more of the RFPs. I mean if you are on a BCD or an Amex and you're going to an RFP, it would be, at this point, irresponsible not to look at Navan. We IPO-ed, we're growing fast, way faster than the rest of them. There's got to be something good here. And by the way, our sales team is probably hit up you and your boss and your boss' boss and the CEO 20 times by now. And so I think that we're getting more and more into a lot of these RFPs. What happens when we don't win an RFP? The very vast majority of it is because they decide not to change. So it might be that they did an RFP to lower their fees on their current incumbent or to threaten their incumbent to try to get them to do something or to bring on a new partner in a new country. But that is why we would lose. It's very rare, very, very rare that we would ever lose an RFP from 1 big TMC to a different big TMC, because those 2 things are basically the same thing. You're using Concur, you switch the agency underneath. It's still a crappy solution, so.

I think that could be very -- the technical answer that is very relevant to our go-to-market. I think there is another thing that is changing. The entire market is changing. You can -- we talk about AI all day long, AI needs to start with online. We are basically competing with calling a travel agent. So it is very strange to expect your employees today to tell them, "Hey, send an e-mail and get after 24 hours. This is the trip, after all [indiscernible] flexibility has changed or the prices. So I think there are actually probably a year ago, there was an inflection point in the market. People want to see something else. Business travel is important, travellers are important. So I think more than just the mechanics, I think the market has changed and it really plays for our favor because nobody else in the market is doing it. So we are in a very, very unique position.

Scott Berg here with Needham & Company. So how should we think about your monetization of customers by industry that they're in. I know the general rule of thumb is larger customers tend to have a slightly lower take rate than smaller customers because of pricing concessions. But we get some questions, and we've done some work on how to think of this by industry because in some industries, a customer might only stay at a 2-star hotel yet in some industries, customers only stay a 4-star hotels and the monetization on a 4-star hotel is usually better than 2 as an example, right? Because the pricing is higher, even though the take rate might be higher, the price is twice as big. Your monetization, obviously, is twice as large in that scenario. So I don't know if you've got the data by that type of [indiscernible], but it would be wonderful to hear what industries actually drive the highest absolute dollar in terms of monetization?

I think our strategy as a company is to bring GBV up as much as we can because this is about taking the market, taking as much market share as possible. So in fact, Navan historically is not that strong in a blue collar and what you are calling 2-store hotels, but I would love to be strong there. I would like to take the entire market. And it means that you'll have different yield characteristics for different segments, different verticals, different geo. But we're going to take this entire market. That's our plan. You're going to see us operating throughout the market. In some cases, it means 2 stars hotels, which may be actually usually higher yield on these hotels, but to your point, smaller ADR. In some areas, lower yield, but much more expensive ADR. So you're going to see us paying everywhere, and you're going to see us continue to drive GBV faster than the rest of the thing, but you can see the revenue continue to go up very fast. So our plan is to grow GBV as fast as we can everywhere.

And I think that it was awesome to see Michael share the diversity of our customer base right now is increasing. And I think this is really, really important to us. As you know, like cycle every single industry that have different pace. And so being available everywhere and have a very broad customer base, is going to make our model very, very strong going forward.

Let me -- maybe just to add, but when we look at a deal, we actually analyze what we think the yields and the gross margins would be. And we have something that is flexible, which is the trip fee. So if we sell to a larger customer that has a lot of corporate negotiated rates, which means we would make less commissions, then we would typically charge more on a trip fee basis. The less CRs that you have, then the more yield that we make and stuff like that. So we analyze it. I don't see many major differences that I can call out by industry. I see it by size of the company, which is what we've talked about before.

Steve Enders from Citi. I want to ask about just industry dynamics. You have Amex being acquired by PE and focusing on AI and their product stack. Just -- how do you kind of view what that means for the market moving forward? And as they try to incorporate AI, what does that maybe mean for competitive dynamics?

Yes. I think, first of all, I saw it as a huge complement that they finally kind of having the language, right, that described the vision that we've been basically sharing with the industry since 2016. So actually, say welcome to 2026. So that's the first thing. I think that the second thing, I think that AI will accelerate anything. So you can take something that is fairly old, fairly antiquated, fairly off-line and you're going to get some benefit by implementing AI. No question about it. I think it will take years. And I'm not talking specifically about GBV. Generally, I think it will take years. But basically, I think we are expecting to see efficiency from AI across every industry. The benefits from AI for a tech company is going to be completely differently because what these guys are talking about right now are the type of discussions that we had 7 years ago in Navan. And the discussions that we are having right now in Navan are so, so, so different that by the time that a certain gap will get closed, we are today in a different place to what they are talking about, okay? So that's one thing. I think transforming a company that employ 27,000 agents worldwide, that you need to call to make a booking or you need to send an e-mail. We talked about developing a model. Actually, I don't know how you develop the model from that. But to do that, that would be a really, really hard task to do. But they will have their business goals from their [indiscernible] firm. So I think this is a huge opportunity for us because I think the first thing that will happen there is a huge degradation of service because if I need to meet a financial goal, and the AI is not going to give me that benefit that I wanted to gain. I think that Michael and [indiscernible] are going to have really, really nice [indiscernible] in the next couple of years.

And I want to add something to that is that we have our vision, we execute, and we always believe that -- and executed against bringing the best technologies in order to create the best experience and best products and best services for our users globally. There's been competitors, some of them got -- some of them, some start-ups disappeared, some start-ups, new start-ups showed up, now Amex is being acquired, it doesn't change our vision. We continue to operate. So we don't care like there's new technologies and the pace of technology accelerates. Like now there's [indiscernible] which is from January of 2026. We do not wait for a competitor to be behind us to say now we need to push really hard. We already are. I already blogged about travel clone. I'm using it. We're going to release it soon. So that's how we operate. So there is market dynamics, we cannot deny it. It's there. It doesn't change our vision and how we operate.

[indiscernible]. So you guys really strengthened your balance sheet with the IPO. As you're thinking about kind of expense and payments cross-sell, can you kind of frame for us how you're thinking about it this year compared to last year? And any quantification would be appreciated?

Yes, absolutely. That's one of the main benefits from the IPO actually, which is getting the balance sheet we need to go after that opportunity at a faster pace. And so you saw in Q4, the volume of payment we processed on the platform was going up 19%, which was an acceleration versus what we've seen for over the year. And that's exactly because we had the right balance sheet. And so we could let the sales team go after that opportunity. And so I'm expecting to see an acceleration there. This is really interesting for us, right? When you think about us selling a flight or a hotel room to someone, if they use our payment solution, and I'm just making more margin on the exact same transaction. So it's a very, very efficient way for us to [indiscernible] expand our margin, also makes our customer even more sticky. So a huge opportunity for us going forward, absolutely.

Blair Abernethy with Rosenblatt. Two questions for me. First, can we talk a little more about the cost of goods sale opportunity with the shift to AI? And how much can you push that? Can you get well north of 90% service from AI in time, what does that mean to the model? And the second question is just around -- can you just give us an update on [indiscernible]. There's been some changes you're implementing there? And I want to see how that's going and if that's also other M&A that might be still out there? Do you still feel that's an avenue for you at this point?

Yes. Maybe I can -- maybe [indiscernible] here. I'm going to start with your question about gross margin expansion. So we've said today, roughly 55% of the customer support interactions are reflected to our AI agents, right? And have we seen an acceleration of that over the last couple of years, you've seen a very steep expansion of our gross margin. We're not going to stop there. I think -- and [indiscernible] here. I think we have a huge opportunity to deflect even more of those customer support interactions to the AI bought. And the main reason is the technology is getting so much better than we see CSAT scores that are on par. People are not even noticing sometimes between -- if they are talking to AVA or to a human agent. So I think we have an opportunity to get this to 60%, 70%, 80% over time, which means there's a huge opportunity from a gross margin perspective. And that's not just the only factor here. The other one is the usage of our own AI model. Maybe I would love Ilan here to explain exactly what we are doing. Today, 30% of the volume and the cost is significantly cheaper than using Frontier model. Maybe Ilan you want to elaborate here?

Yes. I want to say a few things. Today, Ava does chat and only chat [indiscernible] still have, I don't know, also correctly, it's like 30% of the incoming support contracts are still between e-mail and voice. We already released voice internally. So you can call AVA and chat with her and it works so well. And one thing that took us years to master is the -- Aurelien touched on it earlier. It's the orchestration between the AI agent and then the human agent, you'd be surprised. It's very complex. It took us years to perfect because the last thing that we want is for you to have a conversation with an AI and then when it takes you to a live agent, you need to start all over again. It's the most underling thing. It's like when I talk to some of the vendors in the U.S. and get service. Like what was the purpose of this AI just to annoy me more. And now I'm ready to talk to an agent when I'm really [indiscernible] the effort, yes. So we do now voice and we applied all of the learnings that we had from text to voice. So the -- moving the conversation from the agent to the handover is perfected. It works so well that the live agent can ask AVA before it gets back to the user follow-up questions, what did you say that the user wanted this hotel or that hotel? No, I mean this [indiscernible] anything else now, okay, I'm going to connect you now to the user. Amazing, it works so well. The second thing is email. We still have customers that use e-mail. And today, we have agents. So we are now starting to see how other AVA required even more adjustments than voice, was complex because not getting to the details. But definitely, we'll see an improvement in gross margin by deploying both voice and e-mail. And now about models. It's simple. It's -- we have the data luckily since 2023. It's very unique data, data of conversations with an agentic system that spits out its reasoning, which tool it needs to use, why, which parameters, super unique data, no one else has this data in the world. And we train the models on that. And then you have -- the frontier models have anywhere between 1 and 10 trillion parameters, call it, 1 trillion. The models that we use have 27 billion parameters. Imagine how much GPU power and we host it, how much GPU power you need to power 27 billion versus 1 trillion, imagine the latency. So cost latency, and it is so much more professional in that space. If you ask our models about quantum physics, you will not get the best answer, but we don't care, a ton of potential.

Just the [indiscernible] question. First of all, we are only now we have the technology that will allow to do that without upsetting a VIP customer that is willing to pay [indiscernible], okay? So it's completely different, I think, and also booking first class or private jet and so on. This orchestration that we keep mentioning only when we got it to a perfection, we could actually allow ourselves to move to the [indiscernible] customer to our platform. Now the reason that we are moving it is not a gross margin item. It's because our customers, these [indiscernible] customers asked us to have a better booking experience. For example, in a traditional travel management company, any traditional management company and really [indiscernible] similar to that. You have less inventory. There is no low-cost [indiscernible]. There is no specific hotels. Very, very hard to book a train. Now why these things are important, you say, what, the VIPs level will want to use [indiscernible], sometimes yes, because sometimes this is the fastest way for you in Europe to get to somewhere. So having this limited inventory, but I'm a VIP is a problem. Also coming and saying, I always need to talk with an agent, and I cannot just book something quickly. It's a problem. So our customers, the [indiscernible] customers asked us, can't we get an hybrid?" We can have an agent, the agent will know us by name, but also I can get from all of this goodness. So first of all, what drove us is service, is what our customers want. And then there is huge gross margin benefit. So it's a program that we run. We are running it slowly. The customers there are very mature. We already migrated several customers. All of them are happy. It actually opened up to a new opportunity that we didn't think about when we planned it. There are some pretty big enterprises there that have been using Reed & Mackay for a subset and now it's an opportunity. So the benefits there are, first of all, this is what our customer wants; second, definitely gross margin expansion; and third, a huge upsell opportunity. So very [indiscernible] we are very careful with this migration. But what we see so far, we are very happy.

I think we've got time for 1 more question. We'll take it from Gabriela in the back.

We've noticed a trend with software companies where we sometimes see an incredible demo for agentic. For them when you go to the customers, the customer readiness is essentially there's a spectrum of [indiscernible]. So I guess my question for you is how do you view your customer base's willingness and readiness to accept some of the newer generation technologies like Travel floor and the Vonage that you're previewing [indiscernible] generally available today? And then how do you close the gap between the customers that maybe are less forward thinking about how they want to travel and expense experience to go?

I think it's really, really important question and I will actually answer it as the product manager. You have technology, AI technology, nobody cares about it. It's not important at all. What's important is why it can give that user, that customer, does it create value or not. The reason that we keep mentioning AVA, there is no -- it's a no-brainer. I want to get my service faster. I want to get it in the most accurate way. And I want to be satisfied. So AVA, which we started to develop 3 -- actually 4 years ago, was in that context, it was to solve a customer problem. Our machine learning that is embedded in the platform that was developed in 2016 was developed to allow the 7 minutes booking. This is too -- so it doesn't come within headline. This is AI. But the reason that the search results are so effective is because of machine learning. The 15% savings, this is machine learning. The ability to book with AI on an agentic platform, very, very relevant on the demo that I showed you earlier with Edge. But another area that is relevant on what we are calling Navan Classic, when you try to book multi-seating, very, very complex thing to do on an online platform. This is AI. So we didn't come to the customer or the user and say, "Here is an headline. This is multi-city, powered by our model. We are basically solving problems for our customers, and we know to deploy the relevant things for them on the right time. If I will take an advantage right now and we'll make it the Navan Classic UI. We will not have customers. They are not ready for that. And I think -- I want to think that we are very, very good on knowing what is the value the technology create, when to deploy it. So I'm like you, sometimes at all following some of the SaaS companies announcement. And actually, as a customer, I was saying, I don't want this. So I don't want you to touch my sales person that is working extremely well and start to introduce stuff that they don't need. So I think it is very, very important to know what the customer wants, what the user wants and come from that perspective. And if there is something that is fundamental in the Navan culture, it's actually the first value, which is all about the user, all the users all the time, that's what drives us, not AI, not a certain technology, not a model, not a gross margins, it's actually the customer, and we are really good on tuning this customer they want.

And I want to say one thing about it. It's all about experience and less about AI and -- just 1 second. And I just want to share one experience that we've seen. We invite some of our Navan Classic users to try a conversational experience with the agents. And so I've seen a conversation that the user basically said, "I want to book -- please book this hotel in New Delhi for me on the 10th. And the agent replied and said, so I'll help you with that. However, I see that you land on the 9. Should you -- would you want to look it on the 9? And you'll say, "Oh my God, yes, on the 9. Thank you so much. And that's a difference between an intelligent and agentic experience versus forms and tables.

Well, that's an awesome place to end. Thank you all for your insightful answers. Thank all of you again for being here and for your support, and have a great day.

Thank you for joining us. That concludes our investor presentation today. Navigate sessions will resume on the launch pad stage at 3:00 p.m.