Q2 Holdings, Inc. ($QTWO)

All right. Good morning, everyone. Thank you so much for being here today. I'm delighted to have Matt Flake, the CEO of Q2; and Jonathan Price, the CFO of Q2. Thank you both for being here.

Thank you.

Perfect. So I think I'll start off with some questions for Matt, then for Jonathan, and than a handful of others. So Matt, for you first, when Q2 wins against other leading providers for digital banking, what is the single most cited reason? And you said on the recent earnings call that win rates are holding steady and that average selling prices are up. What is driving the ASP improvement?

Yes. So I think on the winning side of things, it's really a formula of -- it's not one thing typically. It's -- the single platform drives a better user experience, operating efficiency. We can get them code faster and then with AI and data, all on that single platform, we get a much better view of who the customer is. So there's a lot we can do with that with AI, plus we have -- we focus on customer experience and how we treat our customers. We always think about -- the customer has a customer of theirs on the other end. And so we're very responsive in customer support, and that's differentiated in this space. Also our experience in the space around whether it's commercial banking, fraud, retail, the 27, 28 years, a lot of people have been in this space building these products. We have deep domain knowledge on whether it's conversions, how other systems work, how you get off of them, how you help people utilize our products. So we kind of bring all of that to bear, and it pays off. You want to cover ASPs?

Yes. I mean, from an ASP perspective, a combination of the deal, the asset size of the average institution that we're selling to is getting larger. We've had a lot of discussion in the deals recently around Tier 1s in enterprise and you're just seeing more volume up market. And then you're also seeing other product lines like our fraud product line win big deals that are akin to large digital banking deals. So the combination of those things is leading to a higher average ASP when you look at the portfolio overall.

Perfect. And you've said a vast majority of banks and credit unions are still on legacy core processing systems for digital banking. And Matt, I believe you've said in the past that we're in the beginning of the fourth inning. And you've also said that the top of the funnel activity has increased quite significantly in the first quarter. Is the replacement cycle actually accelerating? Or is it your share that's actually growing?

I think it may be a little bit of both. I don't like baseball analogies because I'm not a baseball guy. So maybe it's the NBA and it's the second quarter, and there could be 2 over times after that. But no, I think that what's happening is we're beginning to distance ourselves with our platform, our functionality, our customer referenceability. And so the deals that come up, like, for instance, in the first quarter, we were -- we didn't lose a single Tier 1 deal we were competing in. We were about 100%. I'm not saying that's going to be for the future every quarter, but we are winning. ASPs are going up. Our customers are happy. We're continuing to innovate on the platform and it's differentiating because a lot of these A lot of the customers we go after, the regional community, financial institutions are seeing a lot of opportunity in the business banking side of things because Bank of America -- wells, I'll leave one bank out, Citi have they're moving more upmarket and the small to midsized companies are looking for banking relationships where they can have somebody that's local, understands the economy, understands the business they're in. So they need products to be able to replace those systems which are robust at those -- at the larger financial institutions. And we seem to have the one that does -- can help them gain those deposits and deepen the relationships with our customers

Perfect. And maybe moving on to fraud. So one thing that's very impressive to me about Q2 is that you obviously have this large digital banking opportunity and the competitive market is quite consolidated. But fraud is emerging as this massive opportunity for you. It's relatively small in your existing business and somewhat opaque for investors at times. I believe you've described fraud is no longer episodic or confide to a single channel. It's continuous cross-channel embedded in nearly every digital transaction. Could you walk us through what the full fraud deployment looks like at a large FI and why fragmented point solutions don't work well anymore?

Yes. So the fraudsters are using technology and evolving, and it's coming in every direction, whether it's a consumer or a small business, the bank, the network of the small business, the network of the consumer, the e-mail accounts, account takeover, all those things. And so there's really no vector that they won't go to if they can move money out. And they're sophisticated, and so the tools have to be sophisticated. To be clear, for us, we began building machine learning for fraud -- our fraud products in 2008, we began to hire people that use machine learning to look at the behaviors of people. So it's not new to us, but we have expanded that offering, and then we're really excited about our AI products that are coming out on fraud here very soon that are in early adopter stage right now. And I think for a full deployment at a customer, you're looking at everything from check fraud to ACH fraud to account takeover to the network security to identifying the user logs in. It's just everything that we do. And so the value we have at the platform, especially when it's we're using our products and then integrating through innovation studio or APIs, other products is we have all of the data of the person, whether it's their behaviors, who they pay, when they pay, how much they pay, the Fed district. If you think about a business, it's much like a consumer. You pay the same people about the same amount of money the same time of a month, whether it's employees or suppliers to you, and so we began to use that data. We see where you log in from and determine who it is. And so the $1 of fraud costs $4 to $5 for the bank. And if you talk to a bank, if you have a provable solution that's going to stop fraud, I don't want to say they'll pay anything, but it's worth it to them because the money loss is only the beginning because after the fraud occurs, you've got to go deal with the customer, and it could be completely on the businesses fault that it happens, but it doesn't usually play out that way when they talk to the financial institution. So it puts a financial institution in a bad case. So just stopping the fraud, salvages the relationship. They'd rather pay the money to stop it, then to have to go try to collect it again and deal with the customer. So it's a huge opportunity for us. We continue to take advantage of it. I think we've announced big deals almost every quarter for the last 3 or 4 quarters on the fraud side, and I anticipate we'll continue to see some of that momentum.

Perfect. And I believe in the last 2 quarters, you've signed the largest ever fraud tech contracts within your existing FIS. Is it exactly that, that they are signing on to larger suite of solutions? Or does it have to do more with the breadth and scale of their organizations?

Maybe a little bit of both. I think one of the things that happens, and we saw this play out in 2013 and '14 is when you start to get bigger customers, other big customers start to look at it, nobody really wants to be your first $200 billion bank to use a product. And we have a lot of those customers that are expansion sales for us. A lot of them are existing customers. But as you begin to get it installed at $100 billion bank, it's easier to do it at another $100 billion bank or a $50 billion bank or $150 million. But these products are utilized by all customers of all sizes because they all have this threat. So nobody is immune to it. So I think it's somewhat the maturity of the products, the data we can provide them on the success rates. And I think you'll -- that is what's driving a lot of this plus the expansion opportunities within the existing customer.

Perfect. Thank you. And as AI makes cogeneration cheaper, how do you think about the value of your context or contraction regulatory infrastructure and your 25 years of operating scale? I believe your CTO has said that if it takes 2 things to make a product and one just got a lot cheaper, the value of the other thing goes way up.

Yes. I think writing code is we believe it's going to be less expensive to do and faster and those other things, but there's a whole layer of work that goes in after the codes written. It's the delivery of it, the support of it and then converting people from one system to another system educating people on how to use it, fixing things when they break. You have organizations around all that. You have a delivery organization, a hosting organization, a support organization and integration group that has to do that. And so the software is just one piece, and we believe there's a lot of value in AI there, but we are also trying to use AI to make those other roles more efficient as well. But right now, it takes a lot of domain knowledge and experience to work in the regulatory framework we're in. We are examined by the FFIC, OCC, FDIC and FRB. They come in and look at our systems. It's not a startup friendly environment. And so, for instance, on AI now, what we learned over the last 2.5 years, we've been experimenting with it is, you got to start with the compliance and regulatory group. And that's not usually where start-ups want to start or other people want to jump into a space. And then you got to explain the utilization where the data is, how you're securing it, who has access. And so all of that is stuff that we've built those muscles, and I think it takes a long time to do that, and you have to have some domain expertise in that area. So we're bullish on AI, and we continue to think it's going to be an advantage for us.

Makes a lot of sense. And as a follow-up to that, as we know, the majority of your addressable market is on legacy technology and that tech is provided by core processors. There seems to be a fear among investors that Frontier AI providers will join forces with those core providers and offer an AI-enhanced digital banking solution. Why or why not do you think that's a credible threat?

First of all, I'm excited to know that Anthropic can't do anything, that they can't do everything, that they have to partner with people. So there's a mission there that they need to go partner with people. I think if you look at FIS, they announced a deal with Anthropic for anti-money laundering and -- what was the other thing that we did?

Large financial crimes thing.

Large financial crimes things. So those are areas that we don't necessarily play in. That's mostly for really large financial institutions. But for me, I think it's exciting to see that they're reaching out to these areas that are vertical with deep regulatory frameworks to partner with people to go do that. And I think we are using those tools as well. It doesn't seem as if there's any exclusivity to them. I don't think that's how those OpenAI or Google or Anthropic are working. So to me, it just states the case that they're not going to go build these things on their own. And we have the access to the tools as well. So we are spending a lot of time figuring out what the next generation or what agentic banking looks like. And if you look at our track record, we're usually first to market on this stuff.

That makes a lot of sense. And the way I see it that it seems at times the market is missing that you can't drop in AI layer in a bank and have it magically integrated with all the systems, and I believe you integrate with over 1,000 different providers. So do you think that's fair and something that also solidifies your advantage in an AI world?

Yes. I want to be careful, they tell me I'm not allowed to use word moats anymore with AI. But I think those are things that are hard and they take a lot of time to go build and experience and building them. And those are things you can't get wrong. You can't get somebody's balance wrong. You can't get a payment wrong. You can't get identifying somebody the wrong way. So it's going to be a while before you see that. And if you talk to our customers and prospects like we do, they are trying to sort it all out and they want to be thoughtful about it. And we're working with all the customers that we meet with about what is your strategy, what would you like for us to do? And what they've come back with is we want to become more efficient. We want to be able to use your platform to be able to write things that are unique to us, and then we want to help with fraud. And if you look at our AI products Q2 Code, Q2 Assistant, our UAM product, that's about account takeover. That's about helping their back office become more efficient. It's about allowing them to go innovate on the platform rapidly with a small group or forward deployed engineers or whoever they want to go do that. So that's the space we're going at, and that's where we're seeing a lot of interest from our customers. And so we're going to lean into that.

Perfect. And my next set of questions has to do with monetizing AI. So in particular, with Q2 Code, you've described Q2 Code as a discretely monetizable product SKU with a hybrid pricing model. So it's the subscription-based model that your customers know and love, plus credit-based overages for token usage. How are early adopter conversations going on pricing acceptance? And when do you expect enough data to set a scalable pricing model?

Yes. I mean the early conversations so far are good. I would say our financial institutions historically have not consumed from their major tech providers in anything, but the last many years in a subscription model and prior to that license maintenance software. So I think what we're clearly going to have to do here is more of what I'd call a crawl-walk-run approach where you have to create an analog for them that they measure their business on and to price it along that dimension. So take fraud is -- probably an easier example whereas Q2 Code is trickier because the value to the FI may differ if their willingness to customize their own instance is lower than it is for, let's say, a larger institution or an institution with more technical aptitude. So fraud, I think, is an easier one. Matt already laid the case out for what 1$ of fraud cost institution. If you can deliver products and price them that have to do with the outcome that you drive with that product on that fraud dollar, I think it's much more clear, and we're seeing that already in these early discussions. So from a pricing and packaging perspective, I think it's going to depend on the product and the value prop as to our customers, the financial institutions' willingness to move away from that historical pure subscription model. And so as we talked about on the earnings call, I think, at least on the fraud example, it seems to resonate at least with the early adopters that if you can bundle a credit that includes -- they don't care about our token costs, they don't care about our gross margin. All they care about is the value that we can deliver from these products. But if you can translate our infrastructure, our tokens and the value that we're delivering to a pricing scheme that they resonate with, then we think it can work. And so long-winded way of getting to, I feel like there is clarity in some of the products, more than others, and we're going to have to figure it out because the underlying cost structure of these products to us is different. What you can't do is price a product, let's say, Q2 Code at $5,000 a month and then just hope their utilization doesn't put you upside down. That's not a scalable way to grow a product or a business either. So it is a journey we're on from a time line perspective, for any of these products, to the extent they make it to GA, and we've talked about products in the past over the last 2.5 years that didn't make GA for a variety of reasons, including the customer's willingness to pay. We would expect that they would be there by the end of the year if they get there. And so -- and on the fraud side, we're pretty confident in the early signals. And then Q2 Code and Q2 Assistant, I think they have a good chance as well. It's just there's more of a gate on who it's applicable to and when because of like an example I talked about where they may not have a small credit union, may not have a propensity to want to build on top of their instance and just take it out of the box, Q2 Code may not be relevant.

Perfect. And moving on to another topic that I think makes Q2 very differentiated. So you were the surviving platform and 93% of M&A transactions involving a Q2 customer in 2025 and you specifically called out the Synovus Pinnacle deal in the first quarter. You said you don't model a hypothetical M&A into guidance and that most of the time you would expect there to be upside from an M&A outcome. But with M&A activity continuing to pick up, how should investors think about the magnitude of that unmodeled tailwind?

Yes. I mean it is pretty tricky because, first of all, there's different archetypes of a deal. So if you say a Q2 bank acquires another Q2 bank or credit union, you go from 2 platforms to 1. So you're actually out of the gate, while strategically keeping the customer, growing the user base, the ability to cross-sell products and typically the term of the deal elongates, those are all positives. But on day one, you're actually starting from 2 platforms going to 1. So you actually have to fight to get a hundred cents on the dollar back to that and then grow your way beyond. So the economic impact of that actually going bigger than what is your A+B today may take some time. And so it's pretty tricky to know that other than strategically from an enterprise value perspective, that's a good thing when we win those deals. A different archetype would be, let's say, a Q2 bank acquiring a non-Q2 bank there, the economic opportunity is pretty clear and much sooner in the cycle because you're typically acquiring a target bank's customers and moving them to the system quickly, and you're still negotiating terms, you're opening up the contract in every one of these transactions. So I guess long-winded way of saying, the opportunity is there in all of these. The key is that we're winning them. But then the opportunity and the timeline to monetization varies widely based on the individual deals. So I guess what I would look to is when we talk about, let's say, a Synovus Pinnacle or deals that are booked, that's when they hit ARR or subs ARR. And then when the migration occurs of the target bank onto the acquired bank is when the revenue flows. So the best indicator because you're right, we don't model the hypothetical M&A. Even though with a lot of our customers, they're reserving implementation slots months in advance for their hypothetical M&A journey. We don't know who they are. We don't know when they'll actually happen and sometimes they don't. So we're not building that into the forecast. So the best we can do is work with the known deals. And then to the extent it's a deal where again, there's not perfect data out there in the markets of whether they're a Q2 customer or not, but the different type of deal will dictate how soon they can have a revenue implication and what that might look like. Yes. There is no great way to sort of model that the future M&A outcome. But the win rates are definitely a tailwind to the business. And if you think about -- we can talk about the levers that drove outperformance in subscription revenue in '24, '25 and Q1 of '26, the big ones are M&A execution, cross-sale and renewal execution. Those are the big ones.

Also I'd say, on a modeling perspective, there was a lot of momentum on M&A in '25, and it slowed down in the first 4, 5 months of the year so far. So in modeling it can be dangerous.

In Q1 was 21 transactions in our customer base. We're on the right side of 20 of them but that sample size is lower than many of the quarters of the last 6 quarters, meaning 21 deals.

Great. Very clear. Thank you. And Jonathan, maybe moving on to some modeling questions. So you've raised the 2026 revenue guidance to $875 million to $182 million, with subscription revenue growth of at least 14% and EBITDA margins of 27%. You've given an initial 2027 view of 12.5% to 13% subscription growth with, call it, 150 to 200 bps of margin expansion. Could you please walk us through the building blocks, how much is new logos, expansion, pricing, churn improvement, et cetera?

Yes. I mean from a '26 perspective, I mean, the nature of our model is such where the existing bookings that we booked through all last year into Q1 largely gives us the visibility to guide to the revenue profile and the profitability profile for 2026. As we think through, we can look at any period, Q1 as an example, the mix of this business has been consistently shifting to more of it coming from, I'll just call it, expansion. Expansion could be selling digital banking -- retail digital banking to a commercial-only customer, commercial digital bank into a retail customer, one of the fraud deals was an existing digital banking customer, cross-selling other products, all the ones we've talked about today. So you have much greater surface area of larger institutions with bigger wallets that we're selling more product into today. So if we have this discussion about the bookings mix or the revenue contribution in terms of incremental in the period. A couple of years ago, it would have been at least 50-50 from the net new side, if not well over 50-50. What we're starting to see now is a gradual shift towards more of the bookings and the revenue mix in an incremental nature in a period coming from expansion. And so I think that's what's characterized 2026 so far year-to-date. And I think you're going to continue to see that mix on the expansion side push above 50%. In any one period, especially in a quarter, it may look different than that. But over, let's say, a year or longer, we're starting to see that, and we continue to expect that will shift more and more towards expansion. I want to make sure I answer the rest of your question.

New logos, expansion, pricing, churn improvement?

Yes, pricing and churn improvement. I mean, Matt already talked -- we talked about ASPs at the beginning -- churn, 2026, candidly doesn't have the noise the 2025 did. 2025 had one quarter early in the year, the second quarter that had some localized sort of again, total churn was on planned throughout the year, but there was localized churn in the second quarter that created some noise and seasonality. In 2026, both from a Q1 actual perspective and then the rest of the year, we don't see any lumpiness in the churn. We don't see any sort of term fee concentration from M&A, meaning nonstandard churn, but M&A-related churn. So we're trending ahead of plan there. We feel good about churn overall for the year. And then for digital banking within it runs much lower than corporate churn. So yes, all on plan or better from that perspective.

And as a quick follow-up to that, going forward, do you expect to implement roughly the same amount of banks each year as you had in years past? Or is there any prospect of that potentially increasing?

Digital banking units are relatively predictable in a given segment. I mean at the end of the day, it's not just us that have the sort of 5-, 6-year contracts. Most of the industry operates on that. And so that dictates the cadence that these deals come up at. And then you typically have a time line that they'll go to RFP in and then plan their implementation cycle. And so I think if our win rates continue to press up, our win rates are pretty good. And Matt's talked about, certainly in the Tier 1 space when there's commercial, we're way above 50% win rates. On average, win rates are still at 50% across the portfolio. And so we're winning our fair share of better. And we're not seeing more units come up in a given period than we used to historically. I think where you're going to see more volume where you've seen it already over the last few quarters, is with these ancillary areas of the business that are becoming more strategic with fraud or innovation studio or precision lender, what we call relationship pricing now, where those -- those businesses grow at a higher clip than the digital banking business. But there's often less friction in the sales. So we talked a lot about fraud already. In the fraud products, a lot of them are banks are reselling to their commercial customers. So they're always willing to buy more. If they have a customer on the other side that they're, in most cases, mandating use it. So as they grow their commercial book, we grow our fraud products within that customer base. There's much less friction to that unit velocity than you see on digital banking, where there's a more fixed cadence.

And do you think cross-sell is a relatively new lever that Q2 is pushing? Or has this been part of the company's DNA for a longer time?

It's been part of the DNA forever. I mean...

Yes, I think we've always had -- our success team has comp and is driven to drive more or pay to drive more cross-sell in the product. I think one of the things that's happened is if you look at the accumulation of Tier 1 financial institutions, and the way that they buy things, if you're above $20 billion, you're usually buying retail or commercial or precision lender separately and then you go make a separate decision and then you get that up and running. If we treat them right, the product works, there's a logical reason to move to a single platform. So you have one administrative console. You have a lot of customers that are small commercial or consumer and commercial, so they get one log in ID for that. And so we're really exploiting that as we have 100-plus customers in that area. So -- and I think 60% of them are -- or 30% of them are only running one of the major products. So we're really leaning into it. . We probably added more hunters to that because it's a pretty competitive sale. So we go in, our win rates are higher in that area, obviously because the platform is already installed. So we're just refining it and getting better every year at it.

Perfect. And moving on to margins. You stepped up to 62% gross margins in the first quarter after completing the public cloud migration and you've guided to 60% plus for the full year -- excuse me, 60% for the full year and 65% plus by 2030. You said that there is a potential future step function in 2027 or 2028, as you optimize elasticity and costs in AWS. Can you give us a sense of the magnitude of the next leg? Are we thinking 100 bps, 200 bps or more?

It's a fair question, but we've been operating exclusively in the AWS environment now for just a few months. And we have an entire DevOps practice internally that is tasked with planning that outcome for what that could look like in '27, '28, but it's hard to know until you work in that environment. In the data center for the last 20-plus years, you're always solving for peak capacity, what the maximum utilization is. And so you never really have to worry about how do you manage the cost as utilization of the customer goes up and down. You can't -- in our space, there's no like you -- can't afford to have downtime, you can afford to have issues and money movement, et cetera. So you're learning how the system works in AWS in this instance in real time. And so it's just going to take us time, we think, a year or so to then understand what is the incremental automation tooling understanding of how to manage the elasticity from an administrative perspective in AWS to know what the quantum of that upside is. Clearly, we're comfortable that the long-term outlook of 65% is achievable, and that can come from other areas than cloud benefits, obviously. But we're going to find out here in the next year, and we feel like that is unlike a lot of the marginal uplift between here and 65%, which will come from things like subscription revenue mix continuing to press up above 85% will come from efficiencies in the COGS line items from things like AI and the benefits of AI. This one is unquantified yet, but we do think it's more of a step function to the tune of, let's say, 100-plus bps versus 10, 15 bps or the sort of sequential increases we'll see because of things like rev mix or marginal efficiency in the business.

Perfect. And if you had to name the top, call it, 2 to 3 conditions that would create the most meaningful upside versus your stated multiyear outlook, whether faster conversions, higher adjacency attach, better pricing, lower churn, AI monetization. What do you think that would be and which is most likely?

I mean for us, I mean, since Matt launched the 2026 like internal town hall was determined, we pushed all of the company, it was all in with the AI emphasize in terms of how we're going to change our operating rhythm as a company. The holy grail is accelerating top line through AI-driven products, whether they're AI native or AI value props or the value prop to the customer doesn't work in the absence of AI. That's the holy grail. So for us, that would be the one that would be top priority and the best possible outcome in this world of sorting winners and losers with regards to, can you win with your customer in the age of AI. That's certainly the focus. I think there's tons of cost opportunities when we think about efficiencies. I mean, obviously, we talked about it when we rolled out the long-term forecast, the 65-35 model does not contemplate a big uplift from efficiencies on the AI side, whether that's in support, delivery, engineering, sales and marketing, that's the model that assumes we're going to execute and manage our costs and the revenue profile according to the plan and the ultimate outcome of reimagining how we deliver software, how we support our customers, how we go to market. is not contemplated in that. So I think those are the 2 I'd call out, and I'll let Matt jump in.

No, I think you hit them all.

Perfect. And this next point is particularly interesting to me in this AI moment where investors have the perception that competition is generally increasing across the board. But you noted recently that expansion deal length terms are actually increasing versus historical averages, which is a strong signal of long-term commitment. What's driving that? Is that customers locking in better pricing or the breadth of what they're buying? And does that change backlog or revenue recognition dynamics?

I mean, you can take Q1 as an example. The average duration added in a renewal was more than 10% greater than Q1 of 2025. I mean we talk to our customers all the time, especially we've been really trying to pressure test for potential red flags that the software narrative is also the way the financial institutions are thinking about it. And anecdotally, that's definitely not the case, but they vote with their wallets. And when we look at how they're renewing in Q1 of '26, the fact that they are going longer duration, these are already 66-month contracts and renewals are often renewing with 2, 3 years left and then adding 30, 40, 50, 60 months to that. If they were concerned about the start-up disintermediation or the Anphropic example you gave or their own ability to write code faster and disintermediate the vendor by doing it themselves, you probably wouldn't sign up for another 3, 5, 7, 10 years today in the second quarter of 2026. And when you match that with their anecdotal discussion that we're having about they need partners like us to go on their AI journey with them because, a, they're not technology vendors, but even if they can write code faster, they typically want to shift the indemnification to a third party. They don't want to operate the system. They don't want to manage fraud and count takeover and all that. So even if they can theoretically write code faster even as nonengineers inside the bank, one of the big reasons they will lean on their technology vendors isn't just for the product. It's for the ability to offload that risk to a third party. AI doesn't change that. So it doesn't mean that because we're seeing Q1, Q2 of '26 longer durations in net new and renewals, really on the renewal side is what moves the backlog to your point, it doesn't mean that's going to be that way in 3 years. It's just right now, I can tell you, our customers are depending on us when it comes to their AI journey and they're committing to us in terms of the deals they're signing here in '26.

And what is the attitude from your existing or prospective customers on AI? Are they all in? Or do they have some reluctance or hesitancy?

Well, you're talking about probably arguably the most conservative group of business people in the world or at least North America. So they are -- there's a range of early to -- we have some of the larger customers that have technologists in the bank that are wanting to use our tools to begin to drive, some of their own innovation, maybe unique to them. They may have a certain segment of customers they're going after. But largely, they are trying to figure out from -- one of the things we learned in '24 when we began to roll out an AI product was that we skipped the compliance and regulatory group and we went through that, and then we came through -- and then when we got further along, they jumped in and had a lot of questions. And so now we've kind of started with a lot of our customers were talking about how we are approaching this from a regulatory perspective, privacy. And so we're educating them on this, and there's some very curious and interested customers around some of these things. Most of them are looking at driving call center down, cost down on it as opposed to more using it strategically to go drive products, but I think they're working with us. And I think Jonathan's point on the extension of these contracts, banks are trying to sort out who the vendors they're going to get there. And I think we tell our AI story and they say -- I don't know if they're going to do it or not, but mathematically, they have -- or statistically, they have delivered innovation for years and years and years, and we can show it in the road map as opposed to some of the other players in the space have been a little more maintenance oriented. And so I think they're making bets that they think we're going to be one of the winners in this and that shows up in the contract extensions, the win rates because they may not be buying the products, but they're listening and learning from us. And I think there's a huge opportunity for us to take advantage of that relationship, that trust that incumbency we have with those customers to put products in play. And we're taking chances, some have worked or some are beginning to work and some have failed. But if you're not failing, you're not trying.

Perfect. And on capital allocation, you have almost $400 million of liquidity. You have your 2026 converts maturing very soon and you've already executed $102 million of your $150 million buyback. And at the same time, your free cash flow conversion has increased quite a bit, delivering 93% in 2025 against an original target of 70%. And how do you think about share buybacks versus M&A in this moment?

Yes. I mean I think we're in a position where we have maximum optionality to do what's best for the shareholders. And I think the Board has clearly signaled the willingness to participate and take advantage of low what we believe to be undervalued stock price to repurchase shares. The beauty, I think, of being debt-free is we have, a, all the market is available to us, but b, we have the capacity to go do things, whether it's investing in the business, incremental buybacks or M&A but we don't have to go finance that -- prefinance that and dictate the capital structure without knowing what that use of proceeds is. So I think we're going to be open to and willing to continue to repurchase shares if they hang around these levels or were at levels where we think it's undervalued. I think we want to continue to preserve optionality for strategic M&A if and when the right asset comes up. I mean, I think the key here is M&A is not one big thing. It's what's the type of deal? What's the quality of the asset? How much are you paying? How does it strategically move the needle for us? And the bar for M&A today is way higher than it was when we did the PrecisionLender deal 6 years ago, almost 7 years ago now or any of the big acquisitions because our open marketplace model gives us access to all the innovation so there better be a really compelling reason why we need to own it, both from a valuation perspective, from a financial perspective and then also strategically because in our ecosystem you have 2 problems with M&A that some of our peers are living right now is when you build that product through the channel, meaning us, [indiscernible] FIS, Fiserv JAK and one of the competitors buys the asset the channel behaves differently after the deal. So the entire business that they built, you now have to have conviction that you can go operate it and scale it and execute the business case under a different paradigm where you don't have the benefit of folks like us in the case of like a mantle behaving like we did before Alchemy bought them. And that's true of any of these businesses. So you have to have a lot of conviction in your ability to execute. And then the other side of it is our financial criteria is way different now in terms of our expectation of what we would want from an asset in terms of margin accretion, underlying unit economics, revenue growth. And so it's just a way higher bar now. So we're open to M&A. We think we're a great strategic acquirer in the space. We clearly have the balance sheet and the free cash flow now to do what makes sense. But we're not going to do M&A for M&A's sake.

Perfect. Matt and Jonathan, we're out of time, but any closing thoughts?

No, we're just excited about the opportunity ahead of us. We are all in on AI. And I think we're moving as fast and safely as possible. So we're excited about what's ahead of us. And we think with the pipeline where it is and our close rates, we think we're going to have a really strong back half of the year.

Perfect. Matt, Jonathan, thank you so much for joining us this morning.

Thank you.

Thank you all for joining. Bye.