All right. Thank you everyone for joining the Appian session today at William Blair's 46th Annual William Blair Growth Stock Conference. I am Pat Mcllwee, and I'm a research analyst in the software group here at William Blair. As a part of which I cover automation companies, including Appian. Just quickly, I'm required to inform you that a complete list of disclosures and potential conflicts of interest are available at our website, williamblair.com. With that, we're very excited to have the Appian team back at our conference this year. With me today, we have Appian CFO, Serge Tanjga, who just last week celebrated his one-year anniversary at Appian. That said, it is not Serge's first time at the conference, having attended in prior years at MongoDB.
For those not as familiar, Appian provides a low-code automation platform that helps enterprises build secure workflow applications and orchestrate complex business processes across people, systems, and AI agents. The business serves highly regulated industries, including financial services, insurance, healthcare, and a growing federal business, which now represents roughly a third of the company's total revenue. That's my brief two-liner on the business, but it would be helpful to hear it from your perspective, Serge. Can you just start us off by describing Appian's product and service offering as you see it, what that toolkit includes, and how the platform has developed and continues to develop over time?
Sure. First of all, thanks for having me, Pat, and thank you everybody for joining us here in the room. Let me kind of repeat a little bit of what Pat said and then maybe make it real through a couple of examples. Appian is a process automation platform. We've been in the business for more than 25 years, and we say that we automate complex cross-functional mission-critical processes for regulated industries. That's a mouthful. Let me break it out in little pieces. First and foremost, 80% of our ARR comes from four verticals. Those are financial services, insurance, life sciences, and public sector. Then let me give you a couple of examples of what we mean by complex cross-functional and mission-critical. One of our seven-figure customers is a financial institution that uses Appian to monitor fraud.
Before using Appian, they were relying on six disparate systems and a tremendous amount of manual labor to accomplish the goal, resulting in a long time to investigate fraud and significant financial risk. They implemented Appian to replace those six systems, to have a single unified workflow. In the process, they reduced the time to investigate fraud by 98% and reduced the risk of actual fraud happening by three-quarters. Just to use another example, but this time from the public space, a branch of the U.S. Military is using Appian to manage their supply chain for ammunition, so as mission-critical as it gets. Before using us, there was a legacy platform that just couldn't handle the one million transactions a month that they had to do. They deployed Appian in a highly secure manner.
It's called IL5, which is one of the highest designations the Department of Defense has. They now have near real-time visibility into their workflow, which obviously has significant benefits when it comes to the performance that they're expecting. We've actually just had an Investor Day three weeks ago, and on our IR site, you can find our presentation where we have a couple of dozen examples that kind of try to demonstrate in more detail as much as possible what is it that we do for our customers. Hopefully those couple of examples help for now.
That's great. With that more than three-quarters of your business that are these highly regulated industries, you touched on it a little bit with the reliability and the security of the platform. Can you talk about why those customers rely on Appian? Is it the trust you've built over the last 30 years, the security? Is it the services component where you actually help them roll out the solutions? What's the secret sauce?
Yeah, I would say that it comes to both the product and the trust that we've built over the last 25 years. First, when it comes to the product, when you're automating processes, you want to have a comprehensive suite of products that allows you to attach the right problem to the right solution. What that means is, first and foremost, multiple automation tools. It's not just business rules, it's integrations, it's bots, increasingly AI. I'm sure we'll talk about that more. Data capabilities. We have a Data Fabric that allows customers to access data in other systems outside of Appian without having to move it and create complex and expensive processes. We provide Process Mining in order to improve processes over time. Just most importantly perhaps, everything that we do in Appian is highly compliant, secure.
When you operate in these industries, very frequently, you have to be certified by a number of governing bodies. Appian has all those certifications. Those are sort of the requirements to play at the level that we're playing and for the processes that we are automating. The other piece that you talked about is that we've been at this for a while. We've been at this for 25 years. 70%-plus of our business comes from customers who spend over $1 million with us. That's where we're deeply embedded with the core of their platforms. They frequently have development teams that are developing on Appian, and they are standardizing in terms of bringing incremental workloads on our platform. It's both the platform itself and sort of the longevity that we have with our customer base.
Okay. That's great. I think it's a good segue into my next question because I think it's still not completely understood in the investment community what Appian does, providing these reliable, deterministic automation workflows. What foundational probabilistic models like Claude do, what the distinction is, where there's similarities and overlap, where there might be some harmony between the two. I think you had a Venn diagram at your Investor Day. Can you kind of just talk us through that?
Yes. First thing I would ask you to do is to think about a process. A process is a series of steps through which work gets done. Appian has been automating processes. With thousands and thousands of nodes, meaning steps in the process for 25 years plus. When we started off, most of the work in the process at every branch, at every node of the process was done by humans. We were automating the flow of information and governing the accuracy of that process. Over time, other workers, other than a human, were introduced in a process. Those would be rules-based engines, those would be integrations, those more recently were RPA bots. Again, sort of a number of different workers to accomplish the task more effectively. Now comes AI.
When seen from a view of a process, AI is just another type of a worker, and it comes with some strengths and some weaknesses. The strengths of AI is that it's very powerful, that it can reason the way that some of the other digital workers cannot. Doesn't, of course, approach the reasoning of a human, but can do a lot of when it comes to extraction of information, checking of that information, reasoning in face of ambiguity. That's the good part. The negative or the weakness, if you will, is the accuracy. The accuracy is not there, at least not in the verticals that we operate, and at least not at the level of accuracy that our processes require. You all are users of AI in your daily life, so you know that it's frequently wrong.
Our position to customers has always been, figure out how to use AI in processes that actually drive your business as opposed to the side, and figure out how to do it in a way that manages or sort of leverages its benefits while managing against its weaknesses. That's how we implement it at Appian in our processes. What that means is we implement it as a worker to do a specific task. We check its work. We frequently have multiple models working to confirm the work. If we're still not sure of the outcome, we will kick it to a human to handle the exceptions. That way, we leverage the benefits of AI, which is speed and cost, especially if it's vis-a-vis human labor, while at the same time maintaining accuracy.
This has resonated with our customers, because what our customers have seen over the last couple of years, and all enterprises have, is that AI promises don't usually live up to reality when it comes to implementation and scale. By contrast, our message resonates because we're able to show you a proof of concept. We're able to demonstrate performance and reliability. We're able to deliver business value, and then we can actually turn it in production at a much larger scale and still maintain all those features.
Okay, that's great. Just to continue on the topic, I think it's clear that there's somewhat of a fear of disruption associated with this technology. When attending your customer conference, your user conference, you talk to these customers and they seem to be leaning more into these trusted platforms like Appian, rather than they're trying to move away from them or replace them. Can you just talk about what you're seeing at the ground level in those customer conversations, what those dynamics look like?
Yeah. By the way, thank you for coming to the conference.
Of course.
I think that the best way to get to know a company like Appian is to actually talk to our customers or visit one of our events. We have a series called Appian Around the World. The biggest event was a month ago in Orlando. We'll be in your neighborhood before too long, if you will. When you talk to our customers, it's interesting. You mentioned that I've just hit my one-year mark at Appian. I went to last year's Appian World, the big conference. Of course I went this year again. Our message has been actually quite consistent, right? AI is most effective in process. You want to deploy it inside the guardrails of your existing efforts. You want to make sure that you leverage its strengths while controlling for its weaknesses.
We've been pretty consistent with that message for a couple of years at this point. A year ago, our customers were sitting and listening, but still sort of feeling like they were absorbing what we're saying and how's that different from some of the other messages in the market, particularly sort of, let's call it agents everywhere story that some of the other vendors were selling, which is, "I'll have my agents, you'll have your agents. My agents will manage yours. There'll be a control tower." All these very hyperbolic, if you will, features as opposed to actual performance on the ground. Customers were trying to resolve what's truth versus fiction. Fast-forward to this year, our message is still the same. Obviously, we have more features in the market and more traction in terms of actual references.
It was actually, you were there, so you saw a number of household names on the stage. Pfizer was there presenting about their AI use cases. CIBC Mellon was as well, a number of the other customers. What we're seeing this year is customers leaning in. What we're seeing is they're nodding their heads as they're looking at our story of process plus AI. We hear them asking what I would describe as kind of late-stage funnel questions about pricing and packaging, about change management, about implementation. All to which to indicate that our messaging is, the market and the customers are kind of coming around to seeing our way when it comes to how to best deploy AI.
Mm-hmm. Okay, great. I think another one of the kind of thematic concerns with the software sector right now is pricing models. Appian's pricing model is a little unique. Can you talk us through, just kind of clarify what it is exactly?
The first thing that I would say before I kind of break into the pricing models is that ultimately, our customer or our persona that we're selling to is the business process owner or a business unit owner inside a large enterprise, whether it be a financial institution, a government institution, or a healthcare or manufacturing or whoever. What they're buying from anybody, including Appian, is first and foremost impact. It is either cost savings or revenue acceleration or efficiency, some form or fashion. Our job as a vendor is to demonstrate the benefit that you're getting from us and then charge a fair price for the benefit, such that the customer gets the ROI while we get the revenue, of course. The pricing model itself is a way to derive that value.
Once you convince the customer that you're delivering that value, the pricing conversation is secondary. At our Investor Day, our Chief Revenue Officer gave a few examples of how that process works with Appian. Usually, when we get access to the senior decision-makers and our customers, and once we demonstrate the product, they get the value, and the rest is, if you will, a point of negotiation as opposed to whether the sale happens or not. To your question, we meet customers where they are in terms of their preferred way of selling. Whether that's user-based, whether that's per application, which is quite common in the public sector. We have something called Appian Success Plans, which is sort of an unlimited way to buy our licenses in order to accelerate sort of the process of becoming standardized on our platform.
We have consumption both in a hybrid way where we sell AI usage on top of user-based model as well as pure consumption. Again, customers want value. Customers also want budget visibility and clarity, and we want, obviously, to maximize our revenue. We've also been raising prices pretty consistently year-over-year for multiple years now. We think we have the tools at our disposals to really capture this opportunity ahead of us, and I would argue that our performance over the last few quarters demonstrates that.
Absolutely. Yep. You touched on this, I'm going to jump to this question. Appian hired a new Chief Revenue Officer mid-2024, Mark Dorsey, I believe. Since then, you've made a lot of tangible progress in the sales productivity with a stated focus on selling on value. Can you talk a little bit about the steps that Mark has taken over the last couple of years, and how it's manifested into this level of performance?
Yeah. Mark predates me by roughly six months at Appian, and he's picked up on a change or championed the change that has begun even before he joined, which is the focus on the high end of market, focus on large strategic deals, and focus with leading on value. The first thing he needed to do, frankly, was upgrade his management team in order to bring people with more experience of doing these large strategic deals, seven and eight-figure deals increasingly. He's done that. More than 50% of our sales leadership is new, started early and middle of last year, and we brought the kind of pedigree and the kind of discipline that frankly wasn't there to begin with.
Second was reviewing all the basic cycles or motions, if you will, that come into sales, or whether that's forecasting, whether that is deal qualification, meeting with the customer in person and developing those relationships. Pricing as well, and sort of the discounting philosophy. Tremendous amount of progress has been made on that front. Finally is just continuous drumbeat of value, value. Demonstrate value and then charge for that value. Bring along your value consultants to demonstrate that value. Make sure the customer signs on to however many tens of millions of dollars you're saving them, because if you're saving them or generating for them tens of millions of dollars, then you get to charge seven figures for it.
We showed some examples a few weeks ago, but the point is there's more and more of that culture and that drive, and that's what's been driving the performance as of late.
Okay. Yeah, that's great. Kudos to Mark and your whole team for the results that have materialized.
Agreed.
We covered the risks associated with AI pricing, et cetera. As we think about the other side of the equation, benefits of embedding AI within this platform, can you talk about how you're leveraging it within Appian's platform, using Composer to build applications, how you're plugging agents into those applications, all of that?
Yeah. For us, we tend to think of AI in two flavors. One is AI as a worker, and then the other one is AI as an author. Far, we mostly talked about AI as a worker, meaning a node in a process. You assign a task to AI, it does it faster, cheaper, more efficient than alternative workers that we can deploy. In order to get AI in production, AI in your processes, you need to upgrade your license with Appian. We have multiple tiers, but the standard tier is the starting point, and that's where customers don't get access to our AI features in production. In order to get access in production, you got to pay us more. You upgrade to the advanced tier, that's 25%-35% uplift right there. That's the first way we monetize AI.
The second thing that happens is that advanced tier license involves a certain amount of AI usage, what I would describe a moderate amount for a medium-sized use case. As we see customers become more effective and more successful with their AI applications, that's where we kick into the next level of consumption benefit, where you over-exceed what you have there, and then you have to come back and contract with us and buy incremental AI usage bundles from us in order to be able to support the use case that you've built. It's early days, but we're starting to see some of our most successful customers get to the point where they need to have that incremental spending with us in order to maintain the functioning of their application. That's on the AI as a worker side.
By the way, 40% of our customers have access to our AI features. They're paying us some amount for our advanced tier or other ways to buy our AI. We said $100 million of ARR is at that level, so we can keep upselling those 40% of customers, plus push to 40 more. For all of those customers, we still have the consumption upside in the form of AI bundles that we'll be increasingly selling as we go forward. That's the AI as a worker side. Now, if you think about, you heard about the concept of vibe coding and ability to use natural language in order to faster develop applications. This is where our Composer feature or product comes in.
Composer allows you to use natural language elements to build more quickly and more collaboratively between the developer and the business user, and most importantly, more cheaply, applications on Appian. We GA'd Composer in December of last year. We just announced meaningful upgrades to it. What Composer does is it lowers the cost to develop an application, which further opens the aperture of stuff that can be built on Appian. If you think about it, you guys are in the investment business, so you and I would both require a business case for any new spending. Usually that comes with some upfront commitment in the form of professional services, followed by the long-term benefit. As you lower the cost of that upfront commitment, it drives the NPV, improves the NPV of your overall investment, which means that more projects are going to clear the hurdle.
That's what we're hearing with Composer. We're hearing that customers and partners are very excited about the possibility of accelerating development on Appian. The way that's going to benefit over time for us is more ARR on our platform, whether from new applications or from modernizing legacy applications, which obviously is an enormous market that's sort of been out there and elusive for a long time.
Okay. Is there any way you can frame how much faster that implementation is with Composer? Or how much lower the costs, that friction?
Some of the early modernization cases are saying that for right now, we're seeing 25% plus upside versus the normal methods. I wouldn't focus on any moment in time right now. Our goal is to continue lowering that effort through continued innovation. Wherever we are right now, we're going to be better by the end of the year.
Yep
Two years from now, of course.
Okay. Great. If we turn to the existing customer base. You've always maintained exceptional retention rates in the high 90% range, but recently you've actually seen an inflection in your net retention and your growth rates. Can you talk about what makes the Appian platform so sticky, what is helping to drive that inflection on top of that stickiness? The optimization of your sales motion, the adoption of AI, kind of all the above?
Let's break it into two pieces. Appian's always been exceptionally sticky, and that should not be a surprise, honestly, because once you deploy us for your mission-critical process that ties to the rest of the data in your enterprise, that runs things that run your business every day, you're going to have a hard time getting off us. Now, most customers are very happy with us, so there's even no impetus to do that. Even if you wanted to, it'd be an exceptionally hard thing to do, which is why the retention rates have been in the high 90s. The net retention rates, so our ability to expand on top of that renewing business, has, in fact, improved for the last couple of quarters, and that is a combination of two things.
One, our AI story is resonating, upgrades to the advanced tier, going forward, in particular, the ability to sell incremental AI bundles on top of it, plus our improved sales execution across the board. If you think about it was 112% two quarters ago. We just did 115%. That's a function of continued improvements to sell to an already very happy and sticky customer base.
Mm-hmm. Okay. On this one, for those in the audience not as familiar, as a KPI, Appian discloses the number of customers it has with over $1 million in ARR. Last year, that number grew 22% to 140, so it was a clear inflection from growth of only 5% in the prior year. I think you just said that cohort represents roughly 3/4 of your overall revenue base.
70%, yep.
70%. Yep. On the back of that sales force optimization you just talked through, can you talk a little bit about how you've shifted the organizational focus upmarket, and if that comes at all at the expense of some of the customers downmarket, if there's any churn as a result of that downmarket?
We were spending too much. If you go back three years ago, we were spending too much time and too many cycles on small deals and low end of enterprise and mid-market. That was a mistake because it's equally as hard to sell a seven-figure deal as it is a five-figure deal, and if you had to pick which one you would do, you would, of course, do the seven-figure deal or a six-figure deal. That's where actually the retention is also better because the customer's deploying you to something mission-critical when they're writing that kind of check. That means that they like you more, and they're more likely to take another meeting and develop another workflow, and on and on and on. We made a realization that we needed to pivot.
At the time, we also reduced our sales org, this is 2+ years ago at this point. We focused on the high end of market strategic deals. That's when we made the sales transformation and the changes that we already covered. That's the journey that we've been on when it comes to go-to-market execution and still significant upside as we look ahead. Where you see it most is what you said, is the acceleration that we've seen in the $1 million customer cohort. Again, once you're selling large and strategic, you're more likely to break through the seven-figure panel.
What's interesting to me is that even though we've expanded the number of customers, the average spend per seven-figure customer also grew last year. That sort of shows you that even though usually when you grow customers rapidly, average spend declines because you're usually bringing customers in the low seven figures. That's not the case with us. We've actually been able to keep upselling our existing customers. Now we have a growing number in the eight figures as well. Over time, we'll continue to play out that trend. We showed sort of a chart of customer expansion. What's really remarkable about Appian is that we don't just grow for the first couple of years in an account. We grow in years five, six, and seven as well. That, again, is the function of the size of the market and the sort of just how broadly applicable the platform is.
Okay. as we think about large customers. Roughly 1/3 of the business is now from the government. It continues to grow as a percentage of the overall revenue base. It sounds like this is deliberate, given some of your recent commentary. Can you help investors understand how you're thinking about the opportunity within that segment and how much runway you see to continue growing within the federal space?
I would say there's two things that are incrementally interesting in the public sector compared to the overall Appian story. The first is the changes that happened in the federal government and the focus on efficiency that started at the beginning of last year with the new administration has been a clear positive for Appian. The simple reason is the government wants to deal directly with the software vendors, is more focused on the total cost of ownership, and getting faster time to value, and that obviously benefits us as opposed to working through intermediaries, which is how we sort of historically had to approach to market. The second is, as you think about our historical go-to-market execution, we've generally been always better in the public sector than we have been in the commercial.
Most of the changes that happened over the last two years were actually on the commercial side of the business. When you have a solid base of execution and a high productivity, which we already had in the government space, with the secular change in terms of the focus, in terms of how the government wants to do business, those two things combined is what brought the acceleration in the revenue on that side. We see that as a fundamental structural change as opposed to a temporary change. As we think about the pipeline of the opportunities, the ability to deal directly and be the prime seller as opposed to a subcontractor, all of that is still on the table.
One example I'm sure you know is that we signed a framework deal with the US Army for $500 million over the next 10 years, and that's a combination of a couple of things. One, the secular change that we're seeing in how the government wants to do business, plus the strength of our platform, and particularly the AI features that allows the Army to make that kind of a statement of commitment, if you will, going forward.
Awesome. Okay. One more I want to make sure we hit on. One of the knocks on Appian for a long time has been the trends in the bottom line. Profitability seems to have kind of taken a back seat to growth over time. Already since you've joined you've made material progress. We've seen a dramatic improvement in the operating margins. Can you help us understand what levers you've pulled to achieve that recent margin expansion and how you plan to balance profitability with growth going forward?
Yeah. Can't take all the credit. In fact, the journey began well before I joined. Right about the time that Appian decided to focus on the high end of the market and improving sales productivity, Appian also focused on company-wide efficiency. The results have been remarkable. We went from - 8% EBITDA margin in 2023 to + 11% in 2025. Almost 20 percentage points improvement over the last two years. This year, we're showing a 100 basis points plus improvement in margin at the midpoint as of our last guide. We talked a little bit about this at the Investor Day, but we see opportunity for incremental leverage sort of across the board. R&D will continue showing operating leverage because of our move to hire more in low-cost location, and increasingly because of AI efficiencies in the software development process.
We're going to continue improving our sales and marketing efficiency, both as a percent of revenue, but to me, more importantly, in the context of what new revenue we get for our sales and marketing investment. Our G&A spend is already quite lean, but we believe, especially with AI, we can drive continued leverage there as well. If you add all that up, we talk about sort of a multilevel growth story at Appian, and let me just kind of walk you through all the pieces using this year's guidance. We spent much of this time talking about the revenue opportunity as we should have, right? Whether that is growth from existing customers, upselling AI, new logos, federal, you name it. We're guiding to 13% revenue growth at the midpoint of the range. For this year, subscription is growing a little bit faster than that.
That's on the revenue side. On the EBITDA side, we're guiding to 31% this year at the midpoint. That's the revenue growth plus the 100 basis points plus of margin, and we'll be doing 12% margin this year and plenty of opportunity to keep that growing for years to come. If you look at the EPS line, the non-GAAP EPS line, our guidance is for $1 a share. That's 60% growth year-over-year. That's the 31% EBITDA growth, plus faster growth of net income because of the leverage between those two lines. On top of that, we're buying shares and we're shrinking our share count.
Our share buyback that we've just upgraded at our last guidance of $100 million will mean that this is the first year that we get to consistently shrink our share count, and we expect to do that going forward. Revenue is sustainable, margin expansion is sustainable, leverage of net income plus our share buybacks and shrinking of the share counts are sustainable. You think about that, not in any given period, but over a multi-year period, which is how we think about it internally. That's how we're going to drive profitability per share.
Okay. I think we just have one more minute, so if I could tag on to that. How do you think about equity compensation? I think there's been some more scrutiny around stock-based comp recently.
Absolutely. Appian has always been exceptionally careful about dilution. You can see this in the numbers. Stock-based comp as a percent of revenue is 6% for the last three years, versus companies our size, 12%, and even larger software companies, $1 billion-$2 billion in revenue, the median is 10%. We're well above peers, well below peers, sorry, in terms of stock-based compensation. Which is the reason why when you hear our buyback size of X, actually a good chunk of that, almost half of that, is going to shrink the share count as opposed to just offset dilution, and that ratio should improve over time as free cash flow increases.
Okay. Awesome. Well, thank you so much, Serge, for being here. Thanks everyone for coming. Appreciate it.
Thank you, guys. Appreciate it.
The breakout will be upstairs in Jenny A.