Since we've got some lunch, I'm super happy to have the chief financial officer. I'm going to still say the relatively new chief financial officer from Appian, Serge Tanja. Serge, thank you. Welcome back to the TMT conference, but this time as the CFO of Appian.
Good to be here.
Before we get into the conversation, I'm going to go through disclosures. For important disclosures, please see the Morgan Stanley Research Disclosure website at www.morgansanley.com forward slash research disclosures. So with that, maybe to kick off the conversation, you joined Appian in the middle of 2025 after spending over a decade in MongoDB. For those new to the story in terms of Appian, can you pinpoint the problem or problems Appian is solving for customers?
So Appian is a process automation platform that focuses on mission-critical use cases, especially in highly regulated industries. And that's a mouthful of catchphrases, so maybe I thought I'd just make it real with a few examples. So, for example, one of the largest asset managers in the world, one that has a lot of people running around these holes today, is using us, has automated a process to onboard and manage their customers on Appian. And before that, they were using largely a manual process, so this was a cost-saving and revenue growth exercise. Or a top Australian bank is using us for our credit card dispute resolution, and they were replacing an internal app that was conky and it wasn't scaling. Or a medical equipment manufacturer that's using us from order to install a process for their equipment and they had a point solution from another vendor that also wasn't performing, so they placed their app in. Or since we are a big government player, large civilian agencies using us for automating a process to identify and resolve fraud, And before that, they were doing manually while pulling information from multiple systems and obviously involving a tremendous amount of person hours. If you take those examples and kind of boil it down, we work with large enterprises and the public sector to automate mission-critical processes that usually span across different silos of information inside the company or frequently involving the customer. And we usually replace either manual effort, underperforming custom-built application or any number of legacy solutions. We've been in the business for over 25 years, and we've got it to roughly $800 million of revenue, and we think it's a very exciting opportunity ahead of us.
I've been around for every single Appian quarter, and I think in your answer I heard the word process multiple times, and I think there's still some of that lingering impact from when we did the IPO 2017 before your time. around being a low-code platform. So I think there's a portion of the market that still thinks that Appian just builds low-code websites and those types of things and not necessarily tied to a process.
Let me talk about that because as a relatively recent outsider, third insider, there was a moment where it sort of dawned on me on how wrong this is. And what I mean by this is this idea that there's this low-code, no-code space that over time has become associated with calling a citizen developer who takes a few hours of training and then goes build something relatively rudimentary to help in the day-to-day job. And that was what people in this room and myself before coming up here would have probably thought about. And then we realized almost all of our software is implemented by a third party, either ourselves or any number of our partners, GSI, smaller companies, and so forth. The customers pay us from five to eight figures for these implementations. And I think the reason why that, once I kind of put that together, the reason why that realized to me is like, what we're doing is not easy. And to implement and build a software on our platform actually requires expertise that most customers do not have, all but relatively small of our largest customers have. And I think that that, I think, puts the moniker low-code into perspective because it's low-code in the sense that you're not hiring $250,000 developers and keeping them on your platform to build custom code for you. you're using a third-party that implements a composable, reusable solution. But what we do is hard and very sticky, as you can see from our numbers.
Yeah, so let's dive into, because that's not only the core debate with Appian, but core debate across software is sort of defensibility in the age of AI and the risk of AI disintermediation. When we look at, like from your customer conversations, what specific use cases or system requirements make customers conclude that they need the Appian platform rather than building AI-native automation solutions or working with one of the research labs to build their own sort of agentic autonomous solutions when it comes to automating their workflows?
Yeah, so let's dive right in. So, you know, I've been in and around Wall Street for close to 25 years, and I say that Wall Street time and Main Street time, the clock ticks differently and sometimes completely differently and unrelated, and I can't remember a time when it was a bigger dichotomy between investor conversations and actually what we hear from our customers. So in rooms like this and in the room that I've been all day and going back to after this, there's questions about AI becoming self-sufficient and obviating the need for software, including Appian. There's a question about agents running other agents, reporting to sort of across different silos and enterprises, going to some things called control towers, New competitors are emerging to displace people like us that have been in the business for a long time, and that's what I hear in investor meetings and kind of find myself discussing. Customers are in a completely different place. Customers are still looking, for the most part, for the first successful use case of AI in production. So I don't mean give your employees a co-pilot or a tool that makes them productive, makes them write better emails. I'm actually talking about at scale with high accuracy inside a process that runs thousands, if not millions of times a year. And the reason why customers are struggling with that, because AI is a, and I know you've learned this term now, but I'm going to repeat it anyway, a probabilistic technology that needs to be fit inside of the deterministic system to produce the outcomes that is needed when you're doing something mission critical like customer onboarding, like procurement, like budgeting, like the kind of stuff that Appian gets involved with. So with them, the conversation is different. The conversation is, I want to see value. I believe in your vision in terms of delivering that value, meaning it's AI as a node in the process as opposed to some replacement of the process. You guys have the credibility to do that because I've worked with you and my peers have worked with you for a long time. So let's partner together and do that. And so I'll give you an example. A North American insurance company who's approached us about doing the first in production case of our product called Doc Center, which is AI-enabled document extraction. And, again, there are plenty of tools in the market, low accuracy. What we're doing inside a particular process is capable of getting to high 90s or better accuracy. So we partner with them, first use case, 400,000 documents per year. It took us a couple of months roughly to implement it because, again, you want to tune in such that it's accurate enough, and they're over the moon. And we're talking about the second one. The second one is 1.2 million documents per year. And so most customers are still before that first use case. When they are engaged and when they're ready to talk about AI adoption, we see our win rates being meaningfully higher than they are normally, and we're happy with our win rates as they are, which, again, speaks to our vision of AI, a process as an essential enabler of AI, is really resonating with the customer market. And then what we're going to see over time is more use cases, ability to upsell customers, and we generally think it's a great tailwind for our business.
When we think about one of the aspects about Appian as a business is 80% of your revenue, roughly 80% of your income, comes from highly regulated industries where customers value compliance, audibility, reliability. When you think about what's going to keep Appian defensible for the next several years, is it the governance framework? Is it your implementation domain expertise? Is it the support? Or is there something more fundamental to how your platform is architected?
Yes, you're correct. We operate in the most demanding, most highly regulated, most risk-averse industries that are out there. So 80% of our business comes from government, financial services, insurance, and health care. And so we have, again, a long history of subject matter expertise and individual solutions provided in that space. As you think about sort of our framework around process generally, is that you want to deploy a best tool at every node of the situation. So historically, those were business rules engines, and there were RPA bots, there was process mining. AI is another worker to drop into the process at the right moment under the right circumstances. you certainly don't want to use AI indiscriminately simply because it's not the best tool for their job. You wouldn't make AI do math, for example. And so we bring that framework to our customers. And then on top of that, we provide them with incremental functions like security, like auditability, like compliance and certifications, which frankly you would not ask AI to create. And that comes all in the context of complex workloads that need to have higher level of accuracy. And so those are the things that we think are particularly difficult for AI to ever replace, not just in any particular near-term moment in time, but generally speaking. So let me kind of take that to an extra credit level of answer. And so one of the things that I've heard in my meetings today and generally speaking is some flavor of, okay, I get it that this is a near-term positive for you guys. I get that AI fits in the process, but what gives you confidence that's going to be true five, ten years down the road? And that's always a difficult question to answer. Because it's hard to disprove a negative, particularly when the market seems to be as bearish and as scared as it is right now. But I'll offer you two arguments. The first one is, I think implied in that question is some sort of capability of AI to become self-sufficient. So no need. It's going to self-govern. And fundamentally, as a probabilistic technology, it's just very, very difficult to imagine a world in which that happens. So it always needs a set of guardrails and protections around it in order to deliver the outcomes that the enterprise wants. So then the second question in this sort of infinite bearish sort of scenario is, okay, fine. So AI needs that, but why can't another player, a new player, a new breed provide it? And then we're just talking about competition.
And the market's always been competitive.
And what we do is exceptionally hard, which is why you see very, very few companies have successfully succeeded and scaled versus many who have tried. So if another competitor comes and needs to build that enterprise readiness, support, confidence from the customer, particularly in our verticals, I would flip the question. I would ask you, why wouldn't you logically avail yourself of all these tools, all of which are available, and we partner with all of them, to implement properly inside the process in a way that we've done it for a long time and generate value that way? Why reinvent the wheel?
That makes a ton of sense. Let's talk a little bit about the AI monetization story at Appian. And I understand that's still pretty early. But on the last earnings call, you noted that AI usage on the platform is up 14x year over year. From a bigger picture's perspective, what are the AI capabilities available to customers today, and how is that monetized?
Yeah, so I'll start with the framework, and then I'll walk you kind of through the progression. So for us, even before Gen.AI became popular and became usable, we were implementing earlier versions of AI and ML as a node in our process. And then, obviously, when the Gen.AI opportunity became clear to the general public, we sort of rolled out a series of features with increased complexity to effectively deploy AI capabilities in the right way. So first came a set of things that we called AI skills, where you could effectively call an LLM inside the process to produce the exact output that you wanted to produce. And that was very popular as sort of like the early use case for customers. Then came DocCenter, which we already talked about a little bit, which is perhaps the most horizontally applicable use case of AI, and there's generally a feeling like this is easy and can be done out of the box and nothing could be further from the truth, especially in enterprises that use decades of old documents to actually distract value from it. And so we've launched Doc Center in late 2024 in a number of successful cases in production across industries last year. And we're really pushing that as something that is broadly applicable and should be a driver of more adoption of our advanced tier in 2026. And then more recently, we've announced Agent Studio, which is a more comprehensive agentic offering to provide more autonomy and more use cases. and we're seeing first customers come to production, and we hope to tell you more of those stories as we go through the year at Appian World at Investor Day. And then the final step is, you know, what the product called Composer, but more generally modernization. You've heard about it, talked about it as well. AI offers the promise of modernizing legacy technology that called decades-old portion of a software stack that just kind of sit there and deploy resources and are very inflexible and difficult to manage. that now with AI can, at a lower risk, be transformed into a modern platform like Appian. Very, very early days. We're partnering with customers and we're seeing some early traction. But that's kind of the Appian journey with AI from the past all the way into the future.
And when you think about those capabilities that you laid out across those different dimensions, that's mostly available in the advanced subscription tier today. So for customers wanting to consume these AI capabilities, there's an upgrade or land potential on the advanced subscription tier. And so the question is, you know, there's also a premium tier. So how do we think about the roadmap of the premium tier, what's going to be offered in premium versus what's offered in premium?
So one thing that I would argue is perhaps different for us versus many of the other companies who are claiming the AI mantle is that we charge you explicitly for it from the outset. If you want to put a use case in production, you need to pay us. You need to pay us at 25%. That's the average realized price of what customers are paying us to go from standard to advanced tier. And then you get to deploy in production, then you get to get incremental use cases. We said two quarters ago that a quarter of our customers are paying us for the advanced tier, as evidence that we, in fact, are monetizing. We're past the product market state of our AI modernization story. But for the time being, the game is still adoption. We want to demonstrate success, and we want to be the trusted vendor that the customers do their first, second, third AI use case with. We can sell more of the advanced tier to our customers who already have some licenses on the advanced tier. Obviously, we can drive that number of 25% higher. And so that's still the medium term, if you will, goal. But you're right, we have a premium tier, which is another 25% to 35% uplift. We actually have a relatively limited number of features in there. Surprisingly, we do have a handful of customers who are already paying us this. But as we achieve seeding the adoption and move into more modernization, we will put more features into that tier, and then we'll repeat the game. That's what I like about the playbook in that we know how the game is played over a period of multiple years, and we're well along the way of demonstrating that the first step of that process is working.
From a pricing strategy perspective, the market's been concerned on seat-based pricing models. In your government business, you actually don't price per seat, you guys press wrap, but in the commercial opportunity, there's still significant exposure to seat-based pricing. So as we look at pricing over the next 12 to 24 months, how do you think pricing is going to evolve and what's the timeline for the company to potentially see consumption or utility revenue
start to hit the income statement yeah so i think of our pricing tools as sort of a matrix and what i mean by that is on one axis you have all the different ways in which we charge and those are per user per app we have enterprise level agreements that are sort of unlimited in nature we have consumption both as a overage to other models as well as individual ones and then you have sort of ways within each of those models to drive incremental pricing so those tiers It's pure price increases. We increase pricing every year. So we have multiple tools at our disposal to kind of drive the customer we want them to go. The one thing I will say, though, is, and this is what's different about Appian today than would have been the case two years ago, and it doesn't get discussed as much as I think it should be, which is our go-to-market transformation. We've always had a good product. It's always very sticky. Our customers rely on us to solve the most difficult problems, and that's generally just quite remarkable to see when you sit in the room with them. where we haven't been as strong consistently is on our go-to-market execution. And what we've done, and you know this because you've been with us for a long time, but roughly two years ago, we began to more aggressively focus on the upmarket. We focused our efforts there. We actually reduced our sales order about 18 months ago in order to just focus on the top end. And what that really means is selling value. So, for example, in the fourth quarter, we talked about a customer who signed a seven-figure deal with us. It's an aerospace manufacturer, and in the process of designing the solution with them, prototyping it, if you will, we concluded that we can save them $60 million, and they agreed. When you have an agreement that you're saving somebody $60 million, then the question isn't, oh, it's this many users at this price. It's that I'm going to do this. I can do this for you. We both agree that I'm uniquely positioned to do this, so I deserve a portion of that. And, you know, P times Q might change, but if you're selling value, and that's the marching order number one for our sales organ in 2025 and 2026, then mindset shift to sell value. And if we sell value, like, you know, the units will resolve themselves.
Yeah, that totally makes sense. Let's talk a little bit about some of the growth opportunities in specific parts of the business and specific verticals. Let's start with federal. So last year there was a big concern about Doge and the impact of what Doge could have on software spending overall. I mean, you guys did fantastically well last year when it came to U.S. Fed growing well above the growth rates in the business. I have it at sort of mid-20s growth in 2025 and accounting for 25% of total revenue. So as we go from 25 to 26 post-Doge, what do you see as the prospects of the federal business going into next year?
Yeah, so Doge was an unequivocal positive for us. And, you know, I imagine if I had been here a year ago in this seat that I would be receiving quite a bit of skepticism on that point. And, but what it did is focus the government on efficiency, particularly when it comes to their technology spend, working directly with the vendors as opposed to the intermediaries, and really beating the drumbeat of automation. Automate or die, that's the world around easy when it comes to software these days. And the reason why, and obviously that plays to our core strength, the efficiency to streamlining, to eliminating manual processes, consolidating legacy platforms, legacy solutions onto a single modern platform. And we've seen that demonstrate itself. There was a little bit of disruption in the first quarter where we weren't sure, like, who's who. But since then, we've just executed really well. And I will also point out in the fourth quarter, much of our revenue B was driven by the federal space where we exceeded our expectations despite the fact the government was closed for half the quarter. And so as we think about it going forward, I should say one more thing. Particular achievement from our perspective, which didn't help the numbers in the fourth quarter, but is an indicator of the journey that we've made in the government, but arguably more broadly was the framework agreement with the Army. So we've issued a press release that we have a framework agreement of up to $500 million in spending with the Army over the next 10 years. And that's really a hunting license to go and find new use cases and more quickly pursue ability to get more demand onto our platform. So to me, that's an indication of, A, our success with one of our best customers, meaning the Army specifically. B, some of the changing sort of tailwinds, which I think are structural. And we think all of those are positioning us well for growth next year and in the future in the federal space. And, you know, the pipeline is looking very strong into next year.
which kind of goes to, like, when we think about the overall growth of the company, you know, what potentially hopefully gets better is, like, the commercial business. And you mentioned, I think, in Q4 was one of your best commercial bookings quarters. So through the lens of, like, what you're seeing from the sales productivity side, is there a potential for the commercial business to start to get on a similar growth path as the federal business?
Yeah, so we've had better performance in federal over the last few years compared to the commercial, which isn't a function of the end market. It's a function of our execution. And as we think about all the changes in the go-to market that we've done, that's where more of the changes have been focused on the commercial. So when we call that commercial North America, The reason for that was because that's the first commercial theater where we made meaningful changes in terms of leadership, in terms of process. That was done at the beginning of 2025. And Q4 is an A quarter, but it's the largest quarter, and the performance was significant. We said best growth in commercial North America in more than three years. And so that's an indication of when you sharpen your execution, when you focus on selling by value, what's possible. So as we look at it going forward, whether it's federal, whether it's North America, whether it's EMEA, whether it's APAC, we think we have the ingredients in place, product, which has always been there, and then improved go-to-market execution, which is going to carry the growth going forward.
Yeah, and so said it another way, that commercial momentum that you saw in Q4 wasn't because of some product release. It's basically the multi-quarter effort around go-to-market focus and sales productivity. As I mentioned before, about 80% of subscription revenue comes from government, financial services, insurance, and life sciences. What's the runway in these four industries? You know, I get a lot of questions like, can Appian kind of be the pseudo-vertical company? And in terms of meeting the growth and profit expectations that you guys hope to deliver, can we just focus on these four opportunities? What's your sort of perspective on that?
Say we have multiple levels of growth. First, I would say, is plenty of room for penetration inside of the existing customers and inside of the existing verticals. And that's in the context of just the amount of processes that still need to be automated, new or legacy, that we can go after and that we are very well positioned to go after, point number one. Point number two is there's plenty of new logos in that space as well. So we sell with some of the largest players in that space, but there's plenty of white space, if you will, in terms of ability to acquire new logos. And then the other thing that I would say is, as you think about AI as a node in the process, it sort of increases the TAM of automation. And as we get past this early adoption stage where people are still concerned about actually getting value, we think that it will turn from fear to greed in terms of everything that could be done, and we're very well positioned about that.
And then the final thing I would say is,
there's nothing magical about these four verticals if you fully expand the period of time. We have success in manufacturing. We have success in retail. We focused our go-to-market investments where we are seeing the best productivity over time, but as we build our execution muscles, that aperture will also expand that will further be additive to growth over time.
Yeah, I think you mentioned when we were having the AI discussion, things like DocCentric is like a horizontal play that can drive a penetration outside of just those core four verticals. Let's move the conversation to profitability and capital allocation. 2025 was a pretty big year for margin expansion. You guys have been very clear that you guys want to get to a pace of – a moderate pace of investment in sales count and engineering capacity. We think over a multi-year timeframe, how should investors think about the operating margin trajectory, and how should investors think about the pace of margin expansion beyond 2026?
Yeah, so let me talk about history first, and then maybe a little bit about the present and the future. So Appian, and this predates me so I don't get to take much credit for this, did a tremendous turnaround when it comes to this focus on profitability. So right around the time when we decided to focus up market, we generally decided to prove our focus and efficiency across the company. And one thing about Appian is that when we choose to move, we move rapidly. And you've seen this. We've gone from negative 8% EBITDA margin to positive 11%. And even in my time there, I think my first guidance was for 7% at the midpoint, and we ended up closing the year at 11. And we basically kept OPEX flat. However, what's also happened under the surface is that our productivity, particularly in our go-to-market org, has improved to the point where I think our sales and marketing payback on our sales and marketing dollars has become acceptable. Now, our LTV to CAC has always been strong, but our sales and marketing payback wasn't great, which was always an impediment to growth. It's improved so sufficiently that we've earned the right to growth. That was my sort of internal grumpy when I showed up and I saw the numbers. I said, if we can hit these numbers, then we've earned the right to grow. Moderately, which is what we're doing. We're investing in go-to-market. We're investing in overseas R&D. We're still expanding margins after two years of dramatic expanding margins. But, again, because of the moment in time that we find ourselves in, this moderate base of investments while still expanding margin is very, very important. We obviously just got it for 2026, so we're not going to expand beyond that. What I will say is that we see the opportunity to do both. We see the opportunity to drive healthy revenue growth while continuously driving margin expansion. It's really the combination of two that we think is very important.
I'm giving you a lot of questions on the topic of capital allocation, and particularly on a couple of different topics. This is kind of a cross-software, so I wanted to get the Opium perspective. One, the importance of share buybacks as share prices and software, including with Appian, have come down. What level of share dilution should investors expect going forward? And how important is it for the company to get to gap profitability? Fun fact, we were gap profitable last year.
It's $1.2 million, but hey, it's in the green. So let's start there. This is exceptionally important to us. And we've always been very cognizant and careful about dilution. Our stock-based compensation has a percent of revenue in less than half of the average company our size. And that actually matters as you think about sort of compounding growth over time. So I know some people try to think about it as like free cash flow minus SPC. So you can use that framework. It's the same answer. I'm more comfortable just thinking about it on a non-gap basis. So we've just issued a $15 million buyback. That buyback is not a reaction to our stock price being where it is. It's a reaction to where we've come as a company in terms of improvement in profitability. Last year was the first year in which we generated meaningful cash flow. And what we said is like, yes, it's a $50 million buyback. That's important, but think of it as the beginning of consistent capital return policy, which we're now in a position to communicate to our investor. The other interesting thing, because we dilute so relatively little, that $50 million buyback essentially offsets dilution for us. So if you then take a step back and think on a multi-year time horizon, which is how we're running the business, We think we have like four ways to deliver value. One is continued revenue growth. And I'll just use numbers from this year as an example. 11% at the midpoint. 15% cloud, sorry, 16% cloud, 11% at the midpoint for the total. EBITDA is going to grow faster than revenue. So that 100 basis points of margin expansion means low 20s growth in EBITDA. And you have net income that's going to grow faster than EBITDA, pro forma net income, because we don't need capital to grow. So we will delever in absolute and relative terms, and that will mean that net income grows faster. And then finally, as we buy back shares, this year we're roughly offsetting dilution, but over time free cash flow is going to grow more than dilution, so we're going to start shrinking the share account, which is hard for most software companies to do. So as a result, Proforma EPS is going to grow faster than net income. And this year we're guiding to the midpoint of the range, Proforma EPS growing 46%. So as long as we can deliver on all four of those metrics over a period of time, we think we can really compound value and deliver an interesting return, obviously to our customers, to our innovations, to our employees, as well as to our shareholders.
So this last question for you goes back to the guide. I think we exited Q4 at about 16% constant currency cloud growth. I think the guidance assumes a similar level of growth. What gives you the confidence that cloud growth sustains throughout 2026?
Yeah, I'd say a few things. Number one is we actually had a very good Q4 in terms of new business, but it was somewhere back in Lotus, so you'll see more of that. It hit us in 2026, then it helped us in 2024. So that 16% is a little misleading that way. Second of all, there's a little bit of benefit of currency in Q1. So as you think about the two 16s are not comparable, so like one is constant currency, the other one is total. But then fundamentally, it's about execution and our confidence to go and do it out there in the market. We have a pipeline. We have a sales org, which has done a remarkable job of delivering a good year while rebuilding. And so now the goal is to grow and continue improving productivity, and we think we can do it. Thank you so much for giving us a update on the Appian story. Great.