Intuit Inc. (INTU) Earnings Call Transcript & Summary

Excellent. Thank you, everyone, for joining us. My name is Keith Weiss. I run the U.S. equity research franchise covering software here at Morgan Stanley. And really pleased to start out my conference talking with Intuit CEO, Sasan Goodarzi. Sasan, thank you so much for joining us.

Thank you for having me. Happy Monday, everybody.

Excellent. So maybe to start off, you guys just reported Q2 results last week. A really impressive set of results. Beat revenues -- almost every segment, beat revenues. Operating margins came in well ahead of expectations. EPS, you beat EPS by 12.5%. Can you talk to us about what's going on under the hood? Like what are the developments that you guys are seeing thus far in your fiscal year that's enabling you to put up those results, to beat on the top line and bring more leverage and beat on the bottom line the way that you did?

Yes. Sure. So just as context, we ended our last fiscal year growing 16%. And to Keith's question, first half of the year, we actually grew 18%, and Q2, which is what we just announced, was 17% top line growth. And really, the drivers were across the entire business. Our Business platform grew 18%. And within our Business platform, a key metric we look at, which is one of the company's big bets, is mid-market. That grew 40%. And our Consumer platform grew 15%. And the double-click that's really worth calling out there is TurboTax is off to a very strong start; it grew 12%. By the way, while the IRS returns, because early tax formation is not where we thought it would be, IRS is down actually 5 points. But our growth is up 12%. And then Credit Karma is really doing its job contributing to tax growth, and that grew 23%. So as you said, every part of the company, strong growth. The why is it's our AI-driven expert platform strategy, which we declared 7 years ago. And what we said we would create is a platform that combines technology and human intelligence, to deliver experiences that are done for customers. And that's unlocking our total addressable market. It's unlocking accelerated ARPC growth. And it's accelerating our margin expansion, because everything that we've done is tech-led. And I'll just end with the following, which is, in this environment, you don't hear much about AI and HI. In our category, the combination of AI and HI is actually what's unlocking the growth that we are experiencing, because our category is a category of one. We are in a category where customers make high-stakes financial decisions. And consumers and businesses and accountants, if they get those decisions wrong, the liability is incredibly high, which is why they demand expertise. And it's the combination of those 2 that's driving the growth. So we're excited about not just the quarter, but the years to come.

Yes. I'm going to go off script a little bit, but the AI-driven...

I love it when we go off script. Let's do it.

Excellent, let's do it. So that AI-driven expert platform, it's what's gotten me really excited about Intuit. And it took me a while to get it, right? You guys came into the market earlier before any of us were talking about ChatGPT or agentic or anything, talking about this AI-driven platform. And I remember the first time I really dug into is when TurboTax Live came out, right? And as a software guy, my immediate reaction is like, this can't be good, right? You're putting people into a software process. I'm a software guy, I like the gross margin behind software. I don't like the gross margin behind people. It's how we've grown up, right, in this industry. And then we started digging in and we did some surveys of people who were using the service. And at the time, this is probably like 6 years ago, the average interaction time was 12 minutes, of how long people spent using this service. The average payment was $89 at the time, right? So you guys were monetizing over $360 an hour, right, for this service, which is well ahead of what you would have paid if you went to, like, well, if you used your Full Assist service. But what it really dawned on me was how valuable putting a person into the process at exactly the right time, right? And that's what the platform is all about, is how do we make sure that we have that human intelligence part of the equation, how could we get the confidence of having a person to ask the question at exactly the right time so it has extraordinary value? Not just the value of a tax professional that bills out at $30 an hour.

Yes. I actually think that's like one of the most important question you could have asked, so thank you for going off script. A couple of things I would just amplify in what you said. One is our total addressable market is over $300 billion. And we have 6% penetration, up from 5% last year. And the majority of our addressable market is what we call nonconsumption. People are using still, believe it or not, shoeboxes, Excel spreadsheets, Google Sheets. And people, experts, whether it's a tax pro, accountant, bookkeeper for consumers' businesses influence half of that TAM. Why? Because it's customer-backed. Customers demand experts in this environment. In fact, right now, Gen Z are the biggest consumers -- one of the biggest consumers, of experts, because they want somebody to own the liability because these are high-stakes financial decisions. And the reason that's so important is that's what's unlocking our TAM. That's why when you look at 6, 7 years ago, we were, I don't know, growing high single digits. And I think we were, I don't know, $7 billion, $6 billion in size. And today we're over $20 billion growing twice the rate. That's not an accident. And that's accelerated in -- because of all of our investments in AI and HI. And beyond unlocking TAM, the reason it's unlocking ARPC is exactly what you said. These are -- we call them super humans. They're AI-enabled experts that sit on our platform. We do a lot of the work for them. But the customer ultimately wants the certainty and the confidence, which is why they want to engage with them. So they pay more. And because they're so efficient because we've automated a lot of things for them, for the experts, we make more ARPC, and then our company margin has expanded. I remember you pushing on me 7 years ago saying, "You're going to screw up your margins." And you look at where our margins are today versus 6, 7 years ago: bigger company, growing faster and margins are expanding. And it's exactly the premise that you articulated.

Yes. And that same analysis shows that gross margins for that TurboTax Live is actually be higher than your overall company gross margins. And now that same AI-driven expert platform is in other areas of your business. It's in QuickBooks Online with the Assist products and payroll. And Full Assist is the big one. So I want to start there on the tax business. Last year, you guys exceeded your original guidance. You exceeded consensus expectations on tax. But you had a -- you used the phrase constructively dissatisfied, with what happened in the FY '25 tax season. Can you explain to us, like what were the areas that you were dissatisfied with? Because I mean, you beat expectations, the margins were good. What were the areas that you saw for improvement, number one? And number two, how are you addressing those as we head into the 2026 tax season?

Yes, yes. So just for context, right, the total addressable market in tax, business tax and consumer tax, is about $40 billion. And the assisted is 7x bigger than do-it-yourself. Do-it-yourself is just -- is $5 million (sic) [ $5 billion ]. And the reason I start there is the -- last year, we delivered 10% growth in TurboTax. And we were constructively dissatisfied because ultimately we have created software-as-a-service that delivers certainty and confidence, and we were learning a lot, Keith, in terms of how to show up locally. So what was happening last year is we're within 80% of all the households and businesses in the United States, because we have a massive virtual team of experts that sit on our AI platform. But when folks Googled or used an app to say "A pro near me," we never showed up. And so we did a lot of work, that started actually paying off like the last 3 weeks of the last tax season, where we started showing up locally. The second is we've been a software company. And so when they engage with us, we made them go through a bunch of screens before we engaged them with an expert. This is a service. They just -- they wanted, "Where is my expert? And do my tax for me. That's why I came to you." And so the last 3 weeks of the tax season, I would say the last 2 weeks, we had a lot of momentum that delivered that 10% growth. And the reason I shared we were constructively dissatisfied is because there were so many of those things we didn't get right. The experience was not where it needed to be. Customers couldn't find us. They would see our ads in a digital environment and on TV, and then they'd go try to find us and we wouldn't be anywhere near that. And what we've learned is customers don't actually want to come to your office, but they want to engage virtually. But the confidence that you're within a 10-mile radius of their household gives them confidence to go with you. Well, we didn't show up. We addressed all of that for this year. So we have 600 local service centers now, with a couple of flagship stores. And that's all in our guidance, right? It's all tech-led. So it's like it didn't even move the needle. So we're expanding margins in this context. And that's all to show up locally. So what's different this year is we -- and by the way, there's no destination. We're going to get a lot better. But we show up locally. We've removed a lot of the friction upfront so that when you engage with us, we can connect you to an expert. And then our experiences are elite in terms of the experience that we deliver, hooking you up to an expert that's right for you, not a general expert. Best price. Access to fast money through Credit Karma. So those are all the things that's in place, which is, by the way, why -- there's only 6 weeks left in tax season. And last week, when we had earnings, I said, hey, we're halfway through tax season, and I love what we're seeing.

Got it. So at the end of last year, it was actually at your Analyst Day, I was talking to Sandeep, the CFO of Intuit, about those last couple of weeks of the tax season. And he was saying, when you guys got it right, when the local search was really hitting, the conversion was 5x what you'd been seeing earlier. So when we look at what you've been seeing thus far in tax season, overall filings down 5% for the small part of the tax season that was in your quarter, but TurboTax growing 12%. Is that like 5x better conversion starting to play through more broadly through the entirety of this tax season?

Yes, and. So really, all the things that was accelerating the last 2 weeks of tax season, because a lot of the things that we worked all year on and the learnings that we had were in place the last 2 weeks of last tax season. And by the way, we had 6, 7 months to improve upon that. That's what's in play now. And of course, our results were through January 31, which is we were up 12%, but our commentary was beyond January 31. And so a couple of stats that I would share is, based on the fact that people can now find us generally if they Google us or if you're in any of these apps and you want to Google us, we had 5.1 million visitors to our local landing pages or to our stores through February 6. 5.1 million. All of last year, it was 4.2 million. So not only are we getting much more demand because we're actually showing up, most of these customers are prior assisted customers and conversion has improved. So long answer to your question is yes.

Got it. So the near term sounds solid, right? You guys, entering a much larger market opportunity, 7x the size of the DIY market opportunity. The go-to-market motion seems to be getting a lot crisper. You guys are seeing good uplift. But let's be honest, this isn't what the market is worried about, right? I mean the market is concerned about terminal multiple risks. They're concerned about new entrants into the market. Let's hit that head-on. The pushback that I hear all the time is, can't ChatGPT do my taxes? So what's the -- you guys have all said it to me, so don't giggle, right?

The very people that have told you this are laughing at you now.

And it's going to get worse, right, as these solutions get better at tool use and they get better at interfacing with other websites. Like what is the fundamental moats for TurboTax versus a like agentic computing vision of where the agents are just going to do my taxes for me?

Yes. I love the question. Let's sort of start at the top because this infatuation with LLM I think has superseded just thinking through things logically. And we're not on the outside looking in when it comes to LLM usage. So let's start, one, with the category. It's very important. We are a category of one because of customers. And so consumers and businesses make high-stakes financial decisions. And the decisions you all make every year, if you have a side business or your own taxes, those are high-stakes financial decisions. And the liability of getting those wrong is very, very high. And therefore, it's why customers demand a -- they don't demand software. They demand a service that comes with confidence and certainty to help them with those high-stakes decisions. And the stat I always like to remind us of is we actually are seeing Gen Z be the bigger consumer of these expert services. They may not want to see you on video, but they're actually -- because they don't want to be liable for these decisions. What are these liabilities? It's the, if you get payroll tax wrong, if you get your quarterly business tax wrong, if you get your consumer tax wrong. Somebody has to stand before those that are penalizing you, like the IRS, and stand behind that liability. Well, an LLM is not going to do that. So one is just from a category perspective. The second is, to specifically answer your question, our advantage is the combination of the data that we have and the longitudinal behavioral data that we have. Two, it's our AI platform. And our AI platform, the majority of it is machine learning and knowledge engineering, the lesser part is LLMs. Knowledge engineering is about correctness, it's about accuracy, it's about compliance. And then the third is what I mentioned earlier, it's HI. So it's a combination of technology and HI. By the way, I'm answering your question around terminal value and perpetuity. I'm not answering your question about the here and now, just to be really clear. So this is a question about what we are doing and how we view things. And therefore, that's our advantage. And I think the thing that's important to understand is we've built -- the brains of our AI platform, we've built Intuit financial large language models. It's our models that get trained by the data, the customer data doesn't leave our 4 walls. And we use open source, we use OpenAI, we use Anthropic, we use Llama, for things that are context to us, not core. And the reason I'm walking you through this detail is that LLMs have significantly improved the speed of coding. They have significantly improved our ability to ingest data even faster. They help with agents talking to agents and actually what we do, which is our agents need to also be able to hand things off to a human expert. At the end of the day though, LLMs need to know where the source of truth is. They need to know what is the source of truth, where is the source of truth, how decisions get governed and what actions are allowed. And in our space, that's everything because it's about accuracy, it's about compliance and it's about correctness. And so therefore, this notion of, "Well, LLM can do everything" is actually absolutely incorrect because the correctness and the accuracy it cannot do. Last thing, if I could end with this because I think it's very relevant to your question. So that's really -- that's our advantage. That's why we win. It's our scale. It's our trust. It's what we've built. Is you have to look at market structure. We've never won because of code. And just because you can do more code faster and there can be more things that are free, that already exists today. Across everything that we do, the market structure is such that there's a bunch of players and a lot of things are free. At some point, you have to make money, I believe, that's the way earnings work. Is that right? Is that still the case?

Once you become public, yes.

Yes. And that's actually why our relationships that we have announced with an Anthropic and OpenAI, which we can get into in a moment, are so important, because they wanted to partner with us, and we wanted to partner with them, because they don't want to deal with the liability of these high-stakes financial decisions. And we are playing chess, not checkers, which is we want to be everywhere customers want to be, if they choose to want to be there.

Okay. Can I add 2 things, just to get your view on them, and then shore up my argument? In addition to the data, right, and I think data is a foundational platform for making use of any of these agentic community solutions, is critically important. And it has to be, to your point, a source of truth, has to be the correct data, has to be well-governed, right? So you have the correction protections around it. So that's a big part of it. But there's also integrations that you guys have put together over decades, into every financial services partner, every payroll company to get the data in. And you guys have some pretty impressive stats in terms of how much of the filing gets done for you just because of those integrations. And then further on down the line, it's the -- I mean, you also have to submit, right? So you need the underlying rails to submit the taxes, but also the ability to get your refund faster, right? How important is that for your end customers, like the ingestion of data, making it easier to do your taxes from the get-go, but also the ability to get a refund faster? How much of a of value proposition is that for the end customer of TurboTax?

Yes. It's actually a great question. Let me say 3 things. One is everything that you just articulated, all the integrations that we have to every state, to every county, to the IRS, with all the financial institutions, to do all the things that you just articulated, that takes years to replicate. But let's say, for the sake of the argument, it can be replicated overnight. It can't, but let's just say it can. Our real advantage is what I said a moment ago. It's about we deliver accuracy, compliance and correctness for high-stakes financial decisions with the combination of technology and HI, which is what our customers demand. And even if those integrations can be replicated overnight, which they cannot be, but let's say they can, what customers care about is, if I'm a business, it's my cash flow, it's my books being right so my taxes are done right because there's huge penalties. And if I'm a consumer, it's actually all about accuracy and compliance because they don't want to deal with the government. And did I get my maximum refund, which is guaranteed by you, Intuit, and/or did I pay the least, following the law, of course? That's what the customer cares about. And that's really ultimately our advantages, that you cannot replicate with LLMs. And I would just say that we're on the cutting-edge of not only the LLMs that we've built within Intuit because it's core around accuracy, but our partnerships, right, whether it's open source or Anthropic and OpenAI -- and part of those partnerships is about getting access to the frontier models. And so we're on the cutting-edge of what's possible. And what I would tell you is I love LLMs, but they're also commodity. They're all going to keep improving. But what's not commodity is accuracy, compliance and correctness.

Got it. So let's talk about those relationships a little bit. You signed a relationship with OpenAI last fall. Gives you access to all of their users, right, in terms of they're going to see your data. You signed another similar partnership with Anthropic. Can you talk to us from the Intuit side of the equation, what's the benefit that you're going to get from these partnerships? Yes, you have access to the models, you have access to their users. How does that benefit the top-of-funnel, if you will, for Intuit? On the other side of the equation, maybe address the investor concerns that I hear all the time. And it's some form or fashion of you're letting the fox into the hen house, right? They're going to get access to your data. They're going to understand who your users are. You're giving some of that key IP, some of that key assets that Intuit have built up over the decades away to these competitors that are into your backyard?

Yes. So we're doing none of that. Let me now explain. So there are 3 principles in our relationship with OpenAI and Anthropic, and by the way, our relationship and my relationship with both Sam and Dario, and our teams, I have a lot of respect for what they're doing, and these partnerships are ultimately going to be very good for customers. But let me go back to your question, there are 3 principles. One is that we own the customer experience. The second is the data, the data model, the AI capabilities, they are in our platform. So everything is through APIs and MCPs. We don't share any data. We don't share any domain-specific. There's no knowledge sharing. And by the way, they're really not interested in the knowledge sharing because they do not want to deal with the liability of these financial decisions. That was what was enticing for them, which is "We want somebody that has done this and we trust, and really understands AI and HI," which is Intuit. So principles are we own the customer experience, the data models, the AI models. They're all in our platform. So anytime you're engaging with our experiences, and you know it's us, we're not in the back office. Anytime you're dealing with a financial question, you know you're dealing with our brands. But you're in our platform through APIs and MCPs, and nothing is shared. And that's in our contract. And we have governance model around it. And third is we don't share in the economics. And so this is -- so those are the 3 principles. Now to answer the other part of your question, we see this as a big top-of-the-funnel opportunity with OpenAI. And with Claude, we see it potentially as maybe top of funnel, not really. It's more about solving the long tail of things that's context to us, particularly, like for larger businesses and mid-market. And so therefore, let me give you an example. With Claude, we think this could be really helpful in terms of accelerating verticalization across industries. So let me give you an example. First of all, the customer doesn't know what's -- the customer is on our platform. But in the case of a construction company, and every construction company has like 5% of their KPIs, they want to look at what's specific to them. In the case of a construction company that wants to look at their project time line, their lean waivers, their subcontractor payments, and then what does that mean to their cash flow. To connect those 4 or 5 things, that's a long tail of what a construction company would need to see. A roofer would need to see something different. An architect would want to see something different. So now on our platform, Cowork is actually fired by our AI brains. Because remember, at the end of the day, whatever the deliverable to the customer, it has to be accurate, it has to be compliant, it has to be correct. And so the customer can now say, "I want to see this and we'll create a KPI on the dashboard." That's context for us. We don't need to solve these long tail of things. But the real core for us is that what we deliver to them is correct, is compliant. So that's an example of our Cowork integration. The customer doesn't know. The customer is on our platform and they don't really care. But if they say, "Hey, I want to see this KPI in this way," we can now create it for them. Our brains is driving the creation of that. But then in terms of the LLM that we would use, that's on top of our LLM, in that case, it would be Cowork. Multiply that by hundreds of different industries, that's the nature of the partnership. I'll end with where I started. This is the top-of-funnel opportunity for OpenAI -- with Open AI. And for Anthropic, we see it as an opportunity to accelerate serving a lot of long tail of industries and verticals particularly being disruptive in mid-market.

Got it. That's a great segue into the Global Business Solutions side of the equation. One of the opportunities that we're most excited about is Intuit Enterprise Suite, right? The ability to not just go upmarket into the mid-market, but also create more verticalized solutions and do more for those customers, right? And if we think about that through the lens of this AI discussion that we've been having, I think most investors would agree that small businesses aren't looking to buy code themselves a general ledger, right, or a core accounting system. That's not what they're trying to do. What they are looking to do is automate the white space around that, the processes of what does my working capital look like, or what do my KPIs look like? So it sounds to me that the relationship with something like Anthropic is about you utilizing those tools, those modules as a part of your system to be able to attack that white space. So can you talk to us about the white space that you guys see for QuickBooks particularly going upmarket into verticals? And how do you monetize that over time?

Yes. So let me start with it's a $90 billion total addressable market. So it's 1/3 of our $300 billion in TAM. And by the way, that's businesses sort of up to $100 million. We don't plan on stopping there. We will probably go up to $1 billion in size. We don't want to serve enterprise, but this TAM will become far bigger over time. The value -- and we have thousands and thousands of customers now in Intuit Enterprise Suite. And just as a reminder of what we shared last week at earnings, mid-market is growing 40%, our contracts quarter-over-quarter up 50%. New to the franchise is meaningful. And accountants are starting to drive a meaningful part of that. That's up 10 points quarter-over-quarter. And we're actually now expanding our sales force another 30%, just in terms of the results that we're seeing. Back to your question, the opportunity for us is we win on experience, total cost of ownership and price. And so for customers, it's actually not about the white space that you were asking about. For customers, the biggest thing for them is they're saving a ton of time. And because now they're seeing everything in one place, it's actually accelerating their revenue growth. In fact, we had a third-party study done by Forrester. They're seeing a 300% ROI: 1/3 of it revenue, 1/3 of it's cost synergies, 1/3 of it is everything is in one place, so the profitability is going up. It's becoming a competitive advantage for them because now they see everything in one place and they can make better decisions. The Anthropic example I used was that becomes an ability for us to be able to help them with KPIs they need to see. That's the long tail. But by doing that, they can actually compete and win and make better capital allocation decisions. And I think that back to your question around monetization, not only are we upgrading from our base, we probably have 700,000 up to 800,000 in our base that are fit for either QBO Advanced, which is mid-market, and Enterprise Suite, but there's also a meaningful nonconsumption opportunity that's new, which is starting to accelerate. So both are a huge opportunity for us.

Let's come back to the AI-assisted platform for experts. Where does Live touch this? And how has Live adoption been in QBO, because you have Live versions there? And then when you bring accountants into the conversation, accountants have always been part of the top of funnel and pushing people towards QuickBooks. Are they now more of a part of the conversation of how are you utilizing it? Like what services are going around it? And does that improve attach rates of those online services?

Yes. So I love the question because it goes back to QB Live is HI, human intelligence. And as a reminder, it went up 50% quarter-over-quarter -- or sorry, last quarter it was up 50% over last year, which means that we have figured out the demand side and how to deliver against it on the business platform, which is leading to something that you all should be expecting from us in the quarters to come, which is our biggest surprise in the last quarter was we -- QB Live, HI, and the Business platform have sort of been attached. And based on the work we do with customers digitally, they'll say, "Hey, can I have a human help me?" We ran a test, and the test was, well, what if just like TurboTax Live, it's actually just part of the platform? And that there are certain jobs that we would solve as part of the platform. For new customers, get them onboarded and get them their first benefit. And for existing customers, there are certain things we would do to make sure their books are clean and accurate, because there's a lot of help customers need to make sure they're compliant. We were pleasantly surprised by the test results. And so therefore, to answer your question, what you can expect from us as you look ahead is a potential change in our lineup where it's the digital platform and HI that's included for certain bodies of work. The reason this is really important is, one, it will lead to a subscription price change for the larger customers. But two, any time a human expert is attached to one of our customers, attach is up 22 points. Payroll and payments attach is up 22 points. That's significant. And that's all because of the benefit of the platform. So what you can imagine zooming out is our AI expert platform truly at work where it's become a service, it delivers confidence and certainty, and it's unlocking our TAM, it's unlocking ARPC. And because it's all digitally driven and technically driven, it expands our margins.

Got it. I need to touch on the part of the equation that's been a little bit more frustrating, which is Mailchimp. The industrial logic of Mailchimp makes a ton of sense, right? You're do everything in the back office. Like you're saying, small businesses want to see everything in one platform. They want perspective on spending on marketing, what is it resulting in terms of the bottom line. What have been the rough spots on Mailchimp? I know it's more so the low end of the marketplace, but you guys do a lot in the low end of the marketplace. And are you still as convicted that you guys could turn it around? Because it's like the target date for turning around seems to have kicked out a little bit into next fiscal year.

Yes. So I'm more frustrated than you are, and so is our team. So 2 things I would say. One, we are absolutely in love with the customer problem. And we have a lot of conviction, especially for smaller businesses. For mid-market, we want to be and are the AI-native ERP platform that ultimately will eat up payments and payroll. In the large mid-market, we don't have to help customers with growth because there's other platforms that can integrate into our platform. But for SMBs, we have a lot of conviction that we want to solve this growth problem. And we're not in love with Mailchimp. So we're very focused on the upcoming milestones. And really comes down to -- we're actually getting traction with larger customers. It comes down to just a leaky bucket with smaller customers. And particularly, it's customers that pay us less than $200. That's where our attrition is too high. And the reason we pushed out the growth rate of what that segment could be into next year is we want to make sure we get this right. But I'll end with where I started, we're in love with the problem. We're never as a company in love with our solutions and are willing to do whatever it takes to solve the customer problem and accelerate growth.

Got it. And I want to end on sort of taking the AI conversation and turning internally. Intuit has done a great job of expanding margins over time. You've done some layoffs like over the past 2 years to garner more -- really not garner more efficiency, but get the right people in place. When I talk to Sandeep, he feels like there's a lot more that can be done. Can you talk to us about the internal use of generative AI, the internal use of AI, which you guys have been doing for a while? What does that mean in terms of the potential for further efficiency and further leverage in the model? But also the pace of innovation, how quickly are you guys moving being able to utilize these tools?

Yes. I love the question. So the way I would answer the question in terms of what our focus is, we are a company that is all about execution and all about discipline on what matters most to customers. How do we deliver undisputed benefits with high velocity? And how do we ensure we're productive in the way we do it? So a lot of our margin expansion hasn't been -- I think everybody uses AI as the reason or the excuse for margin expansion. A lot of what we've done is just purely services, APIs, automation and, yes, AI is an element of it. And that's what's led to our margin expansion, while, by the way, we've got 3 growth vectors, right? Where AI and HI is a growth vector, money is a growth vector and mid-market is a growth vector. And 2 things I would leave you with. One, we are entirely focused on growth, and I truly do not worry about margin expansion. We're going to continue to expand margins, just because of the way we built the company internally and the discipline that we have. Two, I mean, our coders are delivering 40% more code now versus even a year ago. But we have declared internally that we actually want to 3x our velocity from where we are today because of all the foundational elements that we built. That means we're going to take 10,000 builders and turn them into 30,000 builders, right? It's hard to compete with that when you've got a company of our scale that's now, from where we are, 3x-ing velocity. That will not only lead to undisputed customer innovation and benefit and growth, but it will expand our margins, because we can do a lot more with a lot less. And that's what we've built for internally the last 7 years. So you can expect continued margin expansion. When you always hear that confidence from us, it's because of the internal discipline that we have.

Outstanding. Unfortunately, that takes us to the end of the time. But great conversation, Sasan. Thank you so much for joining. Thank you.

Thank you for having us.