Expensify, Inc.

Hi, welcome to the Q4 2024 and fiscal year 2024 expense by earnings. Before we get started, we're going to kick it over to Nikki, who's going to read the legal disclaimers.

Please note that all the information presented on today's call is unaudited, and during the course of this call, management may make forward-looking statements within the meaning of the federal securities laws. These statements are based on management's current expectations and beliefs and involve risks and uncertainties that could cause actual results to differ materially from those described in forward-looking statements. Forward-looking statements in the earnings release that we issued today, along with the comments on this call, are made only as of today and will not be updated as actual events unfold. Please refer to today's press release and our filings with the FEC for a detailed discussion of the risks that could cause actual results to differ materially from those expressed or implied in any forward-looking statements made today. Please also note that on today's call, management will refer to certain non-GAAP financial measures. While we believe these non-GAAP financial measures provide useful information for investors, the presentation of this information is not intended to be considered in isolation or as a substitute for the financial information presented in accordance with GAAP. Please refer to today's press release. or the investor presentation for reconciliation of these non-GAAP financial measures to their most comparable GAAP measures.

Great. Thanks, Nikki. All right, let's get started by reviewing the Q4 2024 financials. Revenue was $37 million. That's a 5% increase both quarter over quarter and year over year, which is great. Love to see revenue go up. Average paid members were $687,000, which is also up slightly. It's essentially flat, but last quarter we were essentially flat and down very slightly. Now we're slightly up, so we'll take slightly up over slightly down. Okay. And interchange was $5.1 million, which is a 62% increase year over year. The car continues to grow at a great clip. It's a bright spot within the company. Operating cash flow was $7.4 million. Free cash flow was $6.3 million. Our net loss was $1.3 million. Almost there to get profitability. Hopefully we'll get there soon. Our non-GAAP net income was $8.7 million. And our adjusted EBITDA was $12.4 million. Now let's talk about fiscal year 2024. Our revenue was 139.2 million. Our average paid members were 686,000. And our interchange was 17.2 million. Our operating cash flow was 23.9 million. Our free cash flow was 23.9 million as well. Those numbers aren't usually equal. There's a reconciliation in the appendix that you can take a look at. It's just coincidence. We double-checked that. We thought it was kind of funny too. Our net loss was $10.1 million. Our non-GAAP net income was $23.5 million. And the adjusted EBITDA was $39.4 million. Great. Now let's dive into free cash flow. For Q4, free cash flow was $6.3 million, a 272% increase year-on-year. For fiscal year 2024, free cash flow was $23.9 million, a 4,200% increase year-on-year. I know that's a large percentage increase. We're just highlighting it to really signal what a night-and-day difference our free cash flow situation is from where we were in 2023. Now let's talk about guidance. So last year, our initial guidance was to pre-cash flow at $10 to $12 million. Obviously, we tremendously outperformed that. That is due to a couple of reasons. One, the company performed significantly better in 2024 than it had in 2023, so that was great. And also, we implemented a lot of efficiency improvements with AI and other things like that. David's going to touch on the AI piece in a moment here. When we look at 2025, our initial guidance is 16 to 20 million, which is obviously significantly higher than it was last year. I do want to note that there's some conservatism baked into this number because we're not sure how the macroeconomic environment is going to play out for our customers. This is a number that we feel good that we can hit even if things don't go great. But as we see these kind of policy changes and everything play out and our confidence grows, we'll update this number accordingly. Now let's talk about the Expensify card. We had strong growth. Expensify card grew 11% quarter-on-quarter to $5.1 million, and Interchange grew 54% year-on-year to $17.2 million. Also, very happy to announce that the card program migration went off without a hitch. We are now fully migrated. The migration is over. We're very happy about that. It simplifies the accounting story, all the energy changes going into revenue, where everyone expects it to be. So this won't be something we need to discuss going forward. Our fiscal year 2024 interchange that was included in revenue was $9.2 million. In Q4, the interchange in revenue was $5 million, and the total interchange for the year was $17.2 million. Now, we always talk about how the latest month went from a paid user's perspective. In January, we had 665,000 paid members, which is lower than we saw in Q4, but that's to be expected. We've highlighted in pink on this chart previous January, so we usually see some significant seasonality in Q1, and we're seeing that again this year. Now let's just talk about some business highlights to round everything off. SquintSight Card, like I said before, it was a great year, grew 54%, producing $17.2 million total interchange, and we fully migrated the card program. Our free cash flow increased year-on-year by $23.3 million compared to fiscal year 2023. We launched Expensify Travel, which adds fee-based and transactional revenue opportunities for the business. We're very excited about this. Customer enthusiasm has been super high, so we look forward to getting you more updates on that in the future. And last but not least, we reduced all of our debt. We paid our debt down by $22.7 million, and we're now debt-free, something we're very proud of.

Now I'll pass it over to David. Great, thanks. So as we just saw, Q4 was a great quarter. If every quarter were like that, everyone here would be incredibly happy. But what we decided to do is more than just have sort of just uptake quarters. We decided to do something really big. So kind of going backwards, if you will, to the start of our IPO and talking about what has happened since then and what's changed to now. So starting with basically what hasn't changed. Basically, the opportunity size is still enormous. If we look back to kind of our initial TAM and so forth, so much of it is still untapped, and so much of it is still largely the same in terms of competitive dynamics. Viral Lead Gen is still the most scalable model out there, and we're the ones that are focused on bottom-up adoption, so the core acquisition model hasn't changed. Likewise, the Payment Super Act is still a huge hub of data that captures basically the same viral lead gen, transaction revenue, subscription revenue, all packages still in place. That strategy is still a great strategy. So fundamentally, the core tenets of what we set out to do are still in play. The strategy is still a sound strategy. But some things have changed in a very significant way. I'd say AI is finally here. And most interestingly, AI is based upon chat. It's not called email GPT, it's called chat GPT. It's called that for a reason. Because the language of AI is English. And the way you communicate on English and computers is primarily through chat. So I know there's been a lot of questions as to why we've been leaning so heavily into chat, basically. What is the chat-based expense report and so forth? And a lot of people would say, well, I don't really want to replace Slack. I already have a Slack. I don't need a new Slack. And I would say, it's not about Slack. It's not about business chat. It can be. You can definitely use Expensify Chat to collaborate with your colleagues. But the most important thing is using Expensify Chat to collaborate with concierge. Concierge is our primary sort of AI-first experience built throughout the entire product because I think we've learned early on that the UI of the future is a chat-centric UI. What we have building with New Expressify is what everybody's going to look like in 10 years or bring it to you now because it's not just about talking with your colleagues. It's about basically having the super intelligence built into the app in a simple way to communicate with it. So just like ChatGPT, you can talk to Concierge in the direct conversation. But unlike ChatGPT, you can also talk about data that is unique to you. ChatGPT knows maybe everything about the public world. Concierge knows everything about your private world. And it's not just basically general conversations about the data you have access to. It's highly contextual conversations. When you talk to the concierge inside the context of an expense report, for example, You're talking about that expense report, about that employee. We're talking about a particular approval flow. So it's telling you things that are actually unique to that particular report, and you respond to it. You say, like, maybe I want to approve this but not that, and can you forward it to this person? Our concierge AI has the context and understanding to actually do this more sophisticated action for you. So I think we've seen the chat you can see can bring a tremendous amount of efficiency across the board, but it's limited by what it doesn't know. Concierge just knows more, and we're bringing it to you in that context. And so when you kind of think about what happens when you take a super intelligent chat and you combine it with super app data, you get Expensify. And it's a very unique combination because fundamentally expense management is special because we process all of a company's payments. We know where every dollar comes into and goes out of the organization, whether it's expense, card, bills, invoice, and so forth. That's a tremendous amount of awareness that our AI has that no one else has. Likewise, we're actually in the pockets of every employee, whether it's not just the finance team. This is a sales team. It's a C-suite. Everyone in the company basically is using Expensify, talking to concierge, and getting that sort of efficiencies built into the financial experience. Likewise, we physically know where people are. Through our travel duty of care functionality, we know not only where you are right now, but where you're going to be in the future and when. Similarly, we know how the company is organized. We know not just basically who you are, but we know who your boss is. We know how many entities are in the company, what the departments are, who's in those teams, who your clients are, and so forth. Finally, I would say we have access to basically anything we don't know. We can reach out to outside organizations and pull it out because Expensify has already tapped into accounting systems, HR systems, CRM, payroll, and so forth. And so expense management is really the nexus of all of a company's data, and we're powering that data basically with our super intelligent concierge AI. And so as such, when we kind of think about, you know, fundamentally the future, what we set up to do with the IPO. Now, Ryan and I, we're both Midwestern. We're pretty humble. We don't like to make, you know, big boastful claims, if you will. So the humble goal that I think we set up for the company is total FinTech AI supremacy. And now... We've talked a lot about basically how the entire industry is converging in a number of ways. We said early on that the industry is going to go towards real-time extension ports. Everyone kind of followed. We said a long time ago that the industry is going to move towards sort of product suites and super app designs, and everyone's moving that way. I'm telling you now the whole industry is going to move towards a chat-centric design. We're already seeing early signs of that in sort of others as well. And I think that's because fundamentally everyone's following the same kind of process of bringing superintelligence into their legacy applications. And it kind of follows maybe three easy steps, if you will. First is everyone's going to start with what we call kind of deep AI. And that's taking all of the minimal judgment, repeatable tasks. and viewing them as a training basis for the AI itself. Because when you have any sort of AI, it all comes down to it's only as smart as the data you can train it on. Any sort of legacy incumbent player has, in our case, 15 years of receipts and human-generated data that no one else has. It's a very, very defensible, unique asset that no one else can have access to. We use that to train our AI on the nuances of our domain, and in the process, We also happen to create huge cost savings. Now, I'll dig into that a bit more because it was a big part of the Q4 story, or really the fiscal year of the 24th story. And so we talk about our AI is delivering strong cash flow gains in a few different ways. Starting first with SmartScan. What we did is we were able to increase the speed and accuracy of SmartScan and dramatically reduce the cost. Basically, taking the system that was previously a combination of OCR, our hand-tuned parsers, human fallback, and so forth, and largely replacing it with this new LLM technology. Going back to those IPO slides, basically. This is where we're talking about SmartScan looked like then. This is what it looks like now. Basically we've augmented our OCR technology with new LLMs and in the process almost entirely removed human review from the process. And so this is a big deal. Similarly, we're basically doing the same for concierge. We've brought a concierge LLM technology and in the process we have faster chats, more natural chats, and most importantly 80% fewer humanist interventions. Kind of going back to those IPO slides, we talked about how our concierge system is a hybrid AI sort of multi-tiered system where the user writes in a request. It goes to someone called a first responder who evaluates a series of sort of canned repeatable responses. And if they can't do it, it goes to second responders and so forth. With our new upgrades in the last year, now we've almost entirely, no, actually completely replaced the first responder tier, and just using not just repeatable responses, but bespoke and very customized responses to the user that have been trained not only in our sort of public health documentation, but our extensive body of historical conversations, all those sort of repeatable conversations and all the expertise that we've built up over years, that is a unique proprietary training data set for our AI. That is one reason why the concierge AI is so smart, because it understands everything about how the domain of expense management works. Likewise, and this is kind of a smaller one, but it's an important one, QA is obviously important for any high-quality product, and so we have a bunch of, you know, you can talk to a salesperson, you can talk to a account manager, and all this happens over the phone. Now, obviously, we've been trying to record these calls, QA these calls in the best possible ways with industry standards. We randomly sample calls. Some person listens to them, fills out a QA checklist, things like this. We've switched over to a new method where 100% of our calls are transcribed using AI. We will review these calls against best practices using the AI itself so we can score them all and then actually do proactive coaching. So it's not just a matter of saying, hey, you should pitch the Expensify card every time you talk to a customer. Because it's easy to say that, but it's hard to apply that feedback. If the call's not really about that, it's hard to figure out how do I naturally bring it in in a way that's actually going to work. And so what our AI coaching does is it'll take the transcript of how the call actually went And they'll say, here's specifically how you said this. This is how you could have brought the conversation in. And so with this proactive coaching, best practice scoring, and so forth, in the past month alone, we've nearly doubled the number of perfect calls, meaning calls that successfully hit every single point that we're trying to touch on the call and do them the right way. That's a huge increase for a single month and we're just getting started. Last but not least on this list is engineering. So now engineering is obviously an incredibly important part of any SaaS business. And one aspect that we're investing in, we've been investing in for a while, is trying to use AI to the maximum for code generation, automation, testing, and so forth. It's not just us. You might have noticed that actually OpenAI selected us as the basis of their most recent coding benchmark. Because the way that they train their AI is by creating a new, more difficult benchmark than anything else, and then evaluate all the different models against it. Now, in order to do that, they need a rich, open-source ecosystem that has well-defined tasks, that have all the management steps laid out there, because it's not just about can you generate the code, but can you understand the requirements, the design docs, can you take the feedback from the company, and so forth. We're the only company that has an open source repo like this, where we're actually creating issues at this scale and then paying freelance contributors around the world. So it makes a unique asset for how do we evaluate open source models. And so this is why OpenAI picked the Expensify codebase as actually their most important model for how they train the next generation of AI engineers. And I think this is just a sign of kind of the things to come. Across the board, I mean, AI is obviously here, it's growing, and we want to make sure that we are on the leading edge in every single part of it. Now, and I think a reason we can do that is because our company is special. Now, we've talked about how the company is an unusual company. We've got about 120 people right now, and if we're able to do what we can do with 120 people, Any company doing something similar with thousands of people must be doing something catastrophically wrong. And Expensify is able to do this with a team that we have because we've organized in a way where everyone on the core team is trying to focus on innovation, automation, and outsourcing. Everyone's job is to replace themselves some way possible. And we can talk about this in a few different ways. So, yes, internally we can, you know, initially we scan receipts, we figure out how to do it. But when we do that enough and we innovate the process to make the best experience for the customer, then we're able to bring in sort of like U.S. agents and international agents to sort of process receipts, scale up on a massive basis, create this huge archive and repository of receipt data that we use to train our artificial agents. And now we can bring them in basically on an equal basis to our historical human agents and dial back the number of human reviews that we need dramatically. And so our smart scan technology is still a hybrid system involving humans reviewing the AI, AI reviewing humans and so forth, but more and more we can lean on the AI to scale up even further. Similarly, when it comes to concierge, we talked about how we started off with the first responders and we can escalate to second responders. We've used this training resource to eliminate the first responders here entirely and no longer basically have a team devoted to sending the best repeatable response, but rather using AI to generate a bespoke answer to every single question and do it quickly and very low cost. And also we're working on the engineering side as well. And that basically we've long had basically this freelance community of thousands of engineers around the world. We've augmented that with what we call expert agencies. Truly the best of the best. The people who are developing React Native and the technology that we have are all working with Expensify. And in the process, we're also creating this huge training data set that we can use to build the best artificial contributors. And obviously, we're not alone in seeing this opportunity. We've already been talking about how OpenAI identified us as one of the leaders in this open source opportunity. And so, the company is unique because we've got this very core team, and everyone in the team is leaning forward towards innovation, automation, and outsourcing. And you might say, Every company is like that. And I wouldn't say that's necessarily true. If you have thousands of people in your company and 500 of them are actually no longer the most effective way to do the job, that's a huge tension inside the company. And so we don't have any of that tension. Everyone here is focused on viewing AI, viewing outsourcing, viewing automation as a way to supercharge their own jobs and not a threat to their jobs. So first, super intelligence and three-digit steps. Deep AI was the key for delivering free cash flow gains. And I think everyone's going to kind of start with that. That's sort of like the low-hanging fruit. Then we go to what we call surface AI. And that's where you take the AI, after you trained it on the basics, now it knows enough about your domain to identify opportunities and then reach out to the user with some sort of request and then accept the response. Now, to give an example of how that works, one of the pieces we're building we call conversational corrections. We're basically, imagine you create an expense, seven bucks for something called Subway. Now, Subway That's kind of an ambiguous request. It's pretty straightforward to say, oh, okay, Subway, is that for a sandwich or is it for a train? And does it categorize as meals or does it categorize as ground transportation? And you can do that, and then also you can, you know, make it really easy to take those two options. But where the AI becomes valuable is when it allows you to respond to the third option just unexpectedly and say, like, actually, no, that was a typo. It should be Safeway, and it's for, you know, snacks for the office. And the AI has to be smart enough to know what is Safeway? What does that even mean? How do I interpret this answer if we offered them two choices and they came up with some third unexpected choice? The AI is there to receive these more sophisticated responses and not just do it via the app. Because it's a task-centered guy, it also works over text and email. One unique design of Expressify is everything is designed to reach the user wherever they happen to be. Yes, it works best in the app, but if you don't want to use the app, if maybe you actually want to talk to us over SMS or just respond to emails, that works too. And our chat-centric design scales into all these different platforms. It's a completely different way to think about user interface design, away from trying to create a whole bunch of buttons that you can impress, and more towards just creating conversations with the users and allowing them to express to you in natural language what they want done. The third step here is we start with deep AI to build a baseline. AI starts to sort of show this more advanced functionality. But the real superhuman results kind of come in the third stage, what I call kind of elevated AI. And that's where AIs are doing things that are really just too big and too fast. And the analysis are too complicated to be done by humans. And then allowing you to detect this in real time and then escalate it to you while you can still do something about it. So it's not reacting to something after it happened. It's kind of reacting to something before it happens or while it's in the middle of happening. To kind of give some examples, one feature we're building into Expense By that we call kind of a virtual CFO. And basically, it should be doing a variety of the things that you would like your CFO to be doing, like real-time fraud protections and things like this. But imagine a conversation of concierge reaching out to you and saying, Alice's corporate card is showing some unusually large and frequent purchases, but she mentions over here in social that she's on vacation in Maui. Should I block her card to be safe? Now, this kind of conversation is something that requires you to be tapped into a lot of conversations in the organization to infer what's happening about someone. Even if, in this case, Alice didn't mark herself as gone on the calendar, she said that she'd be gone, and that information can be discerned basically from the AIs and then connected with the other data that we have to make an opportunity to prevent fraud that you just might have not noticed otherwise. Likewise, sort of... Most organizations will do some sort of flux analysis at the end of the month, but we don't have to wait until the end of the month. We can basically be doing continuous flux analysis so you can see what's happening and you can step in in real time. So, for example, if we say here, hey, I'm migrating this month's expenses, everything looks normal, but I see a big spike in hotel expenses developing. But don't worry, I think it's for this conference being discussed right here. So basically saying something's weird, but we think it's one time, based upon this information that appears nowhere else in the system, but does appear in chat. Similarly, if we can say things like, you know, cash forecasting is incredibly difficult, but if we actually are able to bring your income and your expenses and combine that with basically all the data from your organization, we can find things like, you know, hey, based on your invoices, bills, and historical card spend, it looks like cash might be tight in Q3, so you might want to pump the brakes on this ad campaign being discussed in marketing. This is the kind of thing where historically you would only know this after they go through the work, make the proposal, do some sort of a cash forecast, and you realize, actually, sorry, everyone just wasted a thousand hours of their time because we can't actually afford this. We can catch these sort of things earlier, or at least that's what we're aiming to do by integrating all this data together. And then finally, same thing for sort of financial management. If we see that you're building up a cash board and we see your intentions behind it and you're not going to spend it for a while, that creates opportunities to manage that money that might not be visible elsewhere. Fundamentally, we think that AI is a tidal wave. It's going to come and it's going to change absolutely every industry it touches, expense management especially, because it is so tapped in to every single part of the organization. So it's a big change that's happening. It's a scary change that's happening. And the only way to avoid being pulled under is basically to make yourself into a surfboard. That's what we're doing with New Expensify. We want to basically view this tidal wave as an opportunity, as an exciting ride that we want everyone here to take with us. So, in conclusion, last quarter was great. It's been a super exciting year. We completed some major investments in deep AI. We've really improved our profitability. We're debt-free, which is a huge accomplishment. We transitioned all of our spend away towards the new Expensify card, which is so great. We launched Expensify Travel. I mean, now it's really a complete T&A solution. We're migrating customers methodically from Classic to new Expensify. And then overall, it was just a great quarter. And it's an exciting year. And I think that 2025 is going to be even more exciting still. So with that, are there any questions?

Perfect. Can you hear me? Yeah. Okay, great. Let's get started with Citi. George and Stephen, I think you're both on the line.

Yeah, you got Steve on here. Thanks for taking the questions from our end. But yeah, really appreciate the deep dive on the AI side. I just want to get a little bit better understanding for Kind of, you know, where the capabilities sit today, kind of what's still on the – what's still kind of in the pipeline that you're working on and have everyone engaged on? And I guess secondarily, you know, for these initiatives to work the way you're thinking – Like, does all of this need to sit within the Extensify app in the chat today? Or, you know, can you go out to third-party systems like Slack or other financial systems to integrate all that data view there?

Great questions. And so as for what sort of exists today and what's coming, I would say the things that we talked about in the deep AI already done, basically the concierge smart scan and the QAing of calls, that's all done in practice right now. I mean, obviously we're improving on all of these, but I'd say that these are real systems creating real benefit right now. For a lot of kind of the more surface AI stuff in terms of user interactions, I'd say, you know, that's all under active development. It hasn't been released yet, but it's real. It's coming. As we think more of sort of the future virtual CFO stuff, it's like prototypes. It's basically, you know, it's demonstrated, but it's not kind of like, you know, production-ready. Everything here, however, is, you know, it's fundamentally real. Like, this stuff exists. We know that it can be done. It hasn't been launched yet, but it's coming. And so we don't have a specific timeline for that. This stuff needs to work really well before it's worth launching. But it's, you know, this is not paperware.

Okay, that makes sense. And then from, I guess, an integration perspective, like, does all this – adoption need to happen within Expensify itself, like specifically some of the chat stuff, and I guess, again, like the third-party system integrations. Sure, sure, sure.

One thing we certainly talked about is integrating with other sort of chat systems. We're not opposed to that. As was sort of mentioned earlier, one of the advantages of a chat system is that we can kind of meet you where you are. And so right now we focus on our app, email, and SMS. But we certainly talked about WhatsApp and Slack integrations as well. One of the challenges, however, is that as we do this, the benefit of the technology is that we can build it in the context of the expense management itself. So, for example, with SmartScan – I'm sorry, with – Slack, we can't show you your expenses inside of Slack, and it's weird to have a conversation about your expense report outside of your expense report. And so I think there's some kind of an impedance mismatch for how our data is structured and how our traditional chat application is structured. But fundamentally, I agree with the thrust of where you're going, and that is we need to meet the customer where they are. And for the right use cases, it would make sense to talk to concierge in different chat systems the more deep you get into the expense management stuff, the more it just makes sense to be part of the expense management system itself.

Yeah, that's great. I guess that then kind of leads to the next question of, you know, if your customers aren't, I guess, necessarily using the chat functionality to that degree today, like, you know, what's kind of the pull to get people to then, you know, I guess for a broader adoption of Expensify and use it kind of the way that you're hoping that they'll take that on.

Well, first I would say I don't know if I agree that people aren't using it this way right now. The process of migrating customers over, and one thing we found is that it's pretty sticky. Like when customers migrate over, they typically stay in New Expense Apply. So they like what they find. Now, I would say fundamentally what we're doing is we're scaling it up for larger companies. We're scaling it up for more advanced flows and so forth. But it's a working system that people use and enjoy today. Again, everything's getting better, but it's already pretty good according to the users who are using it right now. But I also do think that what's nice about this sort of chat-centric stuff, especially like I would say some of these virtual CFO functions I'm super excited for, because they start to show how we can pull customers into a chat context by giving them something to talk about. So, for example – ChatGPT right now, it just sits there idly waiting for you to have a question, and then it gives you an amazing answer to that question. But because ChatGPT doesn't know anything about you fundamentally, it's just kind of idle. We're different. We're basically working for you 24-7, and so as a result, we have a lot of things that we can observe. And I think this creates the opportunity for concierge to reach out proactively and in these different contexts and then pull you into these highly contextual chats, thereby demonstrating the value of this integrated contextual chat. That was maybe a lot of word salad there, but fundamentally I'd say I think that this functionality is a way to demonstrate the value of chat rather than having to sort of imagine what would I do with this chat function. Did that answer your question at all?

Yeah, no, that's great context there. So definitely appreciate that. And sorry, last question for me, just on the travel side of it, you know, good to see that that's out there in GA now. But I guess how are you kind of – what adoption have you seen so far? How are you kind of thinking about what that could look like over 25? Yeah, so the –

initial group, we saw a lot of enthusiasm. We saw a very large increase, you know, month over month in trips booked. Now that was for a small portion of our customer base. Now that we've launched to everyone else, I think it's, we launched this week. So it's, you know, too soon to be drawing friends, I think. But our account managers and everyone, they're basically being overwhelmed with interest and, you know, a million different questions and all that. So I do think that it's going to be exciting since I travel in terms of will it be material to revenue? I think it could be. You know, is it – I think it will be likely like the card where for some period, you know, where it's going to keep telling you it's growing, you know, it's at this amount now, it's at this amount now. And everyone's like, okay. And then eventually it's like, actually, it's gotten quite large and it moves revenue in a meaningful way, even if, you know, subscriptions aren't necessarily going up. So I think I view it the same way as the card. Does that help?

Yeah, no, that's perfect. So I appreciate you taking the questions from our end and send back in the queue here.

Great. Next, Aaron from JMP.

Hey, thanks so much. Hey, guys. So we've talked in the past, you've discussed trying to get to a new normal by summer 2025. What does the new normal look like in terms of the day-to-day of the business and progress with New Expensify? And does summer 2025 still sound like a reasonable time frame?

Sure. Maybe I'll take a crack at this and I'll see what you have to say. I would say new normal means every customer signs into Expensify and sees the new brand. and it goes through basically a new Expensify-centric sales model. And then we have a sizable contingent of customers basically talking about new Expensify, because fundamentally your brand is what your customers tell their friends. But most of our customers today are using our classic product, and so classic is still kind of our brand. And so the new normal would be when we get enough customers over to the new product, that that becomes our new brand, that generates the new word of mouth, and sort of creates the expectations of when someone comes to Expensify, they're coming based on a description of this AI-centric expense management application as opposed to kind of a traditional travel expense tool. And so I think that by summer, like, now summer is obviously a big deal for us. As you probably know, we're sponsoring the Apple's F1 movie. it's going to be a big deal. And so it's kind of like, I know we did a Super Bowl ad a while ago, but that was 30 seconds. This is two hours of seeing Team Expensify's name on the giant screen in front of you. The impression upon that is just so much bigger. And so we expect that that's going to create a lot of awareness, and we're going to try to capture that awareness. So all of the first half of this year is building up to make sure that we're ready to absorb that interest. And the second half of the year is really about converting the interest into action. I agree with that.

Got it. I actually saw the trailer the other day, and a great, great, great logo placement for you guys. Second question, on the spectrum potentially of kind of potentially in 25 so far out in the future, where would you say you are in terms of maybe being able to use price as a lever to drive growth when weighing kind of the choppy macro for SMBs? versus increased product functionality, what's been a sticky inflationary environment for a few years now, and then not having taken price, I think, in call it three years, if that's right?

I think my instinct is we're going to keep price where it's at for the near-term future. I think when we have all of our – when the platform's a little more mature than it is now and we have a broad suite of – super hardened products, then at that point, I think our price becomes kind of silly low. And we won't really see any backlash from customers on a price increase. But I don't think we're there yet. But just to remind you, so the plan is expense management, free corporate card with 1% or 2% cash back, full corporate travel management, invoicing, bill pay, chat, a whole bunch of AI functionality, and also P2P, you know, consumer money transmission for $9 a month. So that is the goal, and that's a steal. We're going probably $100 or something to buy all that individually. So... I think that we are building the conditions where we would have immense pricing power, but we don't want to put the cart in front of the horse.

Yeah, I agree with that. One thing we talked about internally is this idea of this kind of red ocean strategy versus the blue ocean strategy. Red ocean is where highly competitive, blood-infested waters are fighting each other to the death. But there's a huge opportunity out there that's largely uncontested. And so I think the way that we go after this large market is really about bringing a tremendous amount of innovation and producing it at an incredibly low price. And so I think that there's a huge opportunity out there. Our primary method, we expect, of making money in the long run is by growing to acquire new customers, not just basically squeezing existing customers harder.

Got it.

Thank you. Thank you.

All right, next we have Luke. I believe Mark is on the line with us.

Hi, good afternoon. Thank you for taking my question. David, let me start off with you. Could you just walk us through your investment priorities for the coming year?

Investment priorities for the coming year. So I'd say the most important, as I sort of mentioned before, is lining all of product and marketing and go-to-market basically up for this F1 release in the summer. And so what that means on a more practical basis, a lot of testing, a lot of QA, a lot of just polishing up functionality. One thing that we do is when customers come over, we analyze basically their usage of the product itself. We proactively, without waiting for them to report bugs, we find the issues, we fix them, we optimize and so forth. So a lot of mundane stuff. I mean, it doesn't sound super revolutionary, but it's really important stuff. And so I do think the nice thing about AI functionality is that if you have a platform like ours, which is a chat-centric design designed to allow you to communicate to an AI as well as allow the AI to communicate to you in every context, it's actually quite easy to bring in more AI functionality. We don't need to create a bunch of new UI elements and controls and so forth. It's already pervasive. We've done the hard work to build a platform to allow AI functionality to basically engage with you. Now, I would say when we roll in some of this AI functionality, it's relatively low financial investment because the hard work is done to get the data into the same place, to get the UI ready, and to get all this in place. So the bulk of our effort really is on just more mundane testing and migration of existing customers and supporting existing customers and dialing it in. But we sprinkle in kind of like the appropriate AI investments along the way. I don't know if that really answered the question because if you have any, it has to answer your question for you.

Does that answer your question, Mark? I can expound if it did not.

No, that's helpful. Thank you. And then, Ryan, a question for you. Maybe you just talk a little bit about how customer churn trended in the quarter.

So we did have users go up, which is good, right? We're not seeing a huge change in current. Obviously, the paper use users are always kind of volatile, but I think as our new expense side continues to get better, we've increased the performance of our sales team. Dave talked about kind of our investments there, and we're – I think it's seen some encouraging signs.

Yeah, I think so. We've put a lot of effort into account management, and I think that's really had good effects as well. Fundamentally, I think it's just a stable trend, I would say.

Thank you. That's all from me. Thanks.

Great. Next up, we have Lake Street Capital, I believe. Max, are you still on the line?

Yep, I'm still on the line. Thanks for taking my question, guys. Great quarter. Just looking at all these product launches, I mean, with AI, then you have expense or the travel product coming online. I mean, if we think about maybe after the Apple deal in 2025, like what areas do you want to go to next? I mean, what area haven't you tackled? Maybe that's in the back of your mind. Maybe that's the next area or space you want to get into.

I think next. I think next is invoice and bill pay. Yeah, sorry. I think next is – so we have invoice and bill pay. We know it can be better. We know what needs to be done to make it truly competitive. It's great for a small business, but there's really strong competitors out there. So I think in terms of investment – travels in a great place, expenses in a great place, I think building a voice agency is the next logical one.

So I would agree with that, but also emphasize that I don't know that there needs to be a big next thing, fundamentally. I think the next thing is getting all of our customers to use what we currently have, and the next thing is really getting people to understand the value and capture that value and use the value that's already been created. So we think that So fundamentally, you know, again, AI is hard to talk about because it's so eye-roll inducing because everyone just says whatever they want and they make it up and it makes it sort of like hard to talk about and feel credible. But I'd say especially because everyone makes the same claims that they're going to like, you know, we've reinvented everything with AI. And then you look at their product and it looks exactly the same. Like every product, all of our competitors look the same and they all claim that they're like the most AI-centric thing in the world. We look quite different, and there's a reason for that. We look different because we are actually building a different kind of AI-centric environment. What we see here, the user experience that we're making, it might seem radical now in the same sense that ChatGVT seemed like a radical user experience when it first came out. But this is the future of user experience. And everyone's going to be copying this in 10 years or however long it takes them to catch up. And so I think that really the main investment is, yes, bill payment and invoice, absolutely. We need to dial that in. But really this comes down to we just need to – we've built out this broad product. We need to really consolidate it, get all of our customers on it, and just keep investing and improving that.

Yeah, and those two things aren't going to need to be mutually exclusive. Absolutely, yeah.

Great. Great. And I'm guessing the price tag that comes with that Apple ad, but, I mean, in theory, should we see any dramatic changes, I guess, to the non-GAAP operating expense structure throughout 2025? Yes.

So, great question, Max. I've talked about this in the past, but it's good to kind of go back over. So movie accounting, by Gap, is kind of interesting. You do not recognize any of the expense until the movie comes out. Because if the movie doesn't come out, then what do you do? So the money spent for the movie... That's already reflected in our pre-cash flow. That money's already gone, but we have not recognized it in our sales and marketing expenses yet. So what you can expect is a large increase on the expense level, but I want to be clear that money has been spent already. So it's kind of one of the situations where reality and gap kind of look a little different.

Understandably, yeah. I was just wondering if there was any other We're also doing additional marketing around the movie.

We're not just, you know, going to the movie and seeing ourselves there. So, in addition to what we paid for the movie, there's kind of additional go-to-market there as well.

All righty, guys. Thanks for taking my question.

All right. FT Partners. Matthew, are you still there?

Yeah. Hi. Good afternoon, gentlemen. Thanks for taking the questions. A lot of your questions are asked and answered. Maybe just quickly get to see the debt paid down and the reload of the share repurchase authorization. Maybe you could just outline sort of your capital allocation plans as a result above and beyond maybe stock-based comp and so forth.

Yeah, so... A big debate internally and also just in general is we've gone from not having much free cash flow at all to having a lot all of a sudden, which is great. And to what extent should that be put towards buybacks versus debt? We obviously decided to focus on debt, which, to be clear, we're paying a lot in interest. Interest rates went up. So we... and more free cash flow as a result of paying down the debt. And I think our first priority is obviously let's invest in sales and marketing to the extent that we need to. And we've done that, and we still think we're going to have a sizable amount of free cash flow after that. And I think that we're also hiring. But beyond that, we think buybacks are great. We've Done buybacks throughout the years. When we were private, we were doing buybacks, which is kind of strange for a private company. But we were long-handedly doing buybacks. We love reverse dilution. So nothing to announce right now, but, you know, we like buybacks.

Understood. That makes a lot of sense. And I think philosophically, you know, a lot of what you guys are doing probably answers this question, but just to kind of cement it as, you know, you get a lot of unit cost improvements through automation, AI, right, like the smart scan, 80% fewer escalations, et cetera. And so, you know, as far as the willingness to kind of drive more sales and marketing budget into customer acquisition, things like through the movie and otherwise, is that – are we thinking about that right, that, you know, as the sort of operating leverage in the model improves, it makes more sense to sort of push into paid user growth efforts going forward? Yeah. Absolutely.

I mean, we'll see how the movie goes. Yeah, I mean, I agree with you that when you have more cash, you can put more towards sales and marketing for sure.

Yeah, I mean, we've never been shy about taking big swings when we see a big opportunity. But I think that we run a very efficient shop because that free cash flow didn't come from nowhere. It came from efficiencies and discipline. And so I think we take big swings when we see the opportunity, and then we're not afraid to in the future.

But we also, I guess, it's not in our culture to spend just to spend. We need to feel good about it. It's not like, you know, you always hear use it or lose it. You know, we better spend this or they're going to reduce our budget. That's not part of our culture at all. So the dollars we spend, we feel good about. And if we don't feel good about it, we pull it back as quick as we can. Yeah. Great. Thank you both. That's all for me.

That was everyone.

All right. Thank you, everyone. No questions about the card migration, first time since IPO. So happy that we got that migration done. Thank you all for the time, and we'll see you next quarter. Thank you very much. Thanks, everyone.