Okay. Perfect. Hey, thanks for coming to our next session. I'm really happy to have Serge here from Appian. I need to remember that. Maybe let's start actually kind of with that kind of topic a little bit. So, Serge, when you got to Appian, kind of what was the excitement for you?
Yes. Thank you. First of all, thank you for having me. Thank you all for being here. When I got the call about the Appian CFO job, a few things were appealing. Number one was just even if you glance at the financial, it looks like a good business, right? High retention rates. If you read, high gross margins. If you read, a little bit about or talk to a few customers, you hear, you see that they're very satisfied and want to do more with us. So that was step number one. And then step number two was, as I got to understand a little bit of, the AI story, I thought that Appian had a very compelling value proposition in AI.
I've been, you know, honestly, ever since AI became a part of our vernacular, and we learned names like ChatGPT and Claude and so forth, it seemed to me like it's going to take a while for enterprises to really adopt it, and there's going to be a number of hurdles.
As it does with every new technology, particularly this one, because it comes with some risks that we haven't seen before. Like, AI is non-deterministic. AI can produce outcomes that are bad, and so it wasn't a surprise to me that, you know, the initial hype was going to take some amount of time to transfer into actual enterprise adoption, and where Appian comes in is, when it comes to AI, exactly at that moment of enterprise adoption in a way that works for enterprises, and what I mean by that is we insert Appian into processes new or existing. Sorry, we insert AI into processes new and existing, and we arm it with data across the enterprise.
Which are the two things that AI needs to be successful. AI needs guardrails and security and auditability, in order to perform its task the way that, you know, the enterprise wants it to with the level of accuracy that actually an enterprise, you know, mission-critical application requires. And also, obviously, to become better over time, we need access to all the data, which we do through our Data Fabric offering. So to me, Appian AI pitch was like, this makes total sense, and then the more I've talked to the customers, the more I felt that that's actually true.
As I got to know the company better, the third element of excitement came in, which is that we've done a lot of changes on the go-to-market side to increase our focus at the top end and really focus on high-value use cases, six, seven, hopefully over time eight figures that where we align with the C-suite, where we align with priorities because we are the high-quality product in the market. We're the best-in-class product. And that transition, I would say, began roughly 18 months ago that accelerated last summer. I'm sorry, I guess began two years ago at this point, accelerated in the summer of 2024.
And as I was meeting our sales team and seeing how they're doing things differently and frankly some of the things they're going to implement, it seemed like there's a great opportunity to improve productivity and then hopefully over time also grow the sales arc, so you know at a very simple level best-in-class product, happy customers, strong AI value proposition, and improved sort of sales and marketing effectiveness seemed like an interesting proposition, and since I joined the company I would say that all of those elements I feel the same or better about, and the thing that it was also just been lovely to see is just how good the culture is at the company, and how focused we all are across all the departments in terms of getting better, getting more efficient, frankly using AI internally, you know eating our own cooking. It's a fun place to be about.
The follow-on question I had is like you obviously worked at a kind of fast-growing. Kind of slightly scaled up, slightly larger organization. Like if you walk in there now, what was kind of what can you bring to the table? What was the team expecting from you in a way to?
Yes. So, I think that there's two primary things. One is I've seen kind of roughly basically exactly this journey, back in my days in MongoDB and what I've on some level a lot of what I found at Appian today reminds me of what I found in MongoDB, you know, six or seven years ago, and what I mean by that, it's a company that's kind of not startup but not quite enterprise grade.
And so there's a lot of opportunity for us to scale our processes to automate internal work, in fact, using Appian frequently as our own solution, and building cross-functional efforts such that we become faster at executing as we grow as opposed to slow down, which is what happens to a lot of companies. And I was fortunate enough to sort of see that process at Mongo and be a part of it. So I feel like I bring a decent amount of, you know, relevant experience to that part of the story. And then the second one is capital discipline, which first of all, I got to credit Matt and the team, kind of right about that time that they decided to focus up market. They also decided that, you know, being unprofitable is no longer okay. And you've seen the results. We went from negative 12% EBITDA margin in 2023 to our latest guidance is roughly 10% for this year. So that's that. That largely predates me. [crosstalk]
You take a credit here. I take a smidge of credit, but not the entire 2,000 basis points. But that capital discipline will go forward. When you return to growth, it looks a little bit differently 'cause now we're going to return to moderate growth when it comes to investments. And then you need to be very explicit about, you know, the outcomes that you're trying to drive, the ROI, staying, making sure everybody's accountable, being willing to pull the investment when it doesn't work. And that's also another cycle that I kind of lived for a while. So I believe I can help on that side as well.
Yeah. Okay. And then you mentioned go-to-market already. Like I think because Appian was a little bit under the radar over the last two, three years. A lot of people, when I talk to investors, miss that a little bit. So can you speak a little bit like what Matt did there? A while ago, you know, it looks like we are coming out the other end, so like it should be exciting from here. Yeah.
Fingers crossed. So, Appian was spreading itself too thin, sort of, across the application landscape. And I would say at the low end are very simple use cases, that are, you know, barely any automation on top of what already exists in the place, all the way to exceptionally difficult, highly mission-critical. One of my favorite stories is. We have our top customers come to the headquarters every once in a while, and then, you know, all of us execs, if we're there, we meet them. So I met with a part of the. It's a governmental institution, but one not based in D.C. And they sent like 20 people to the Appian headquarters. And they were there implementing this application and with great support and partnership with us. The way they describe it is like, "If this doesn't work, U.S. financial system is in jeopardy. So their words might not.
That's important. Yeah.
Yeah. So my point is that that's on the other side.
Yeah.
So if you spread yourself across that entire spectrum, you kind of aren't clear with yourself or with your customer what you're best at And we're clearly best at the high end when it's complex, when scalability, high-quality performance, and significant value can be generated versus some of the stuff that is at the lower end, where there's also more competition and where we don't mind their win rates, but those never really turn into, you know, seven- and hopefully over time eight-figure deals. So really the focus was, let's go where the data and our product quality tells us we should go. And that frankly was two pieces. One piece was we reduced our sales org i n the summer of 2024.
And, you know, for those of you who are students of basic math, you know that if you eliminate the least productive part of your sales arc, like the rest is by definition more productive without really doing anything. And you're saving money, and that's important. What we've done really since then is started instituting leadership processes and discipline around how we approach our larger customers that has led to further improvements in sales productivity. And that to me is the hard part. That to me is the thing that's exciting. And again, it's been happening this year and latter half of last year, you know. So that's the momentum that we now need to sustain going forward while at the same time expanding the org.
Yeah, and then so there, there's the stuff that you can do internally, but then obviously you have externally, like what's going on in the world in terms of like things. Can you kind of speak towards kind of macro and, don't go federal yet because that's the next question, but like what are you seeing like at the moment in the normal world and then, like a follow-up?
Yeah. So, I frequently see dichotomy between the world of Wall Street and the world of, you know, corporate.
Yeah. Real world.
Either way, good or bad, right? And so what's happened throughout this year was a tremendous amount of headline volatility, tremendous amount of market volatility. We have tariffs. We don't have tariffs. This peace agreement, this not, and on and on and on that obviously is very, you know, as a person who reads newspapers, can be very unsettling at times. And obviously we've seen significant gyrations in the market this year as well. However, if you look at our business, we have not seen that maybe yet, but we certainly haven't seen it yet translate into hesitance of customers to engage. Hesitance of customers to pursue their IT and enterprise objectives. We haven't seen any changes in the deal cycles or win rates or really anything around that. We talked a little bit about AI.
There's an element, some combination of excitement, frustration, and a little bit of concern that all surrounds AI, but the conversations are good. We've had conversations with almost all of our customers around like our AI offering and sort of presenting at a high level. We told you that roughly 25% of them are paying us for AI already. So that's supposed to show you that like at least so far, you know, my macro crystal ball is non-existent. It hasn't impacted how the business works.
And then, now, the federal part. So there was a, you guys have probably from just a history of, as well, like more exposure to federal than others. That was incredibly volatile this year with DOGE, the shutdown, etc. Like how is it playing out for you? And there's a negative aspect, but it's also there should be a positive one because if you want to get more productive as a government. More efficient, like software and Appian should be in theory kind of be right in the middle of that.
Amen, so I'll divide that into kind of. The near-term blocking and tackling, which has been, you know, a fun ride, and then, you know, what that means in the long term and what we firmly believe it does, so when the new administration came and DOGE rolled into town, things were disruptive in Q1, and people didn't know if they had jobs. People didn't know who reports to who, who approves what, and despite that, we actually had a very, very good first quarter in federal. By second quarter and in third quarter, it felt more like business is stable. We had always fears, like, in enterprise this is true no matter what your customer is, but enterprise software, you're always concerned that, like, do you know who the last signer is?
And when you get to the last signer and you find out that's not the last signer, then you realize the deal's not going to close, but frequently you don't find that out until that very end, so we were always worried about like, okay, well, this all looks good. The demand is good. The conversations are healthy, but is somebody going to pull the rug underneath us? And that did not happen. We had a good Q2, a good Q3. Overall, if you look at the entire federal fiscal year, that business grew faster than total totality of Appian, which like if I told you that, you know, on January 20th, you'd be no way, right, so I credit that to the execution.
That is our highest performing part of our sales arc and also just the relationship that we've built across the government 'cause we are based just out of DC and some of our largest early successes as a company are actually there, then the shutdown came. Nobody did anything for six weeks. We talked a little bit about our guidance, how we do expect that to result in some amount of disruption because people come back and it's honestly, you know, you come back, you got to look at your inbox, you take a few days to you know, dig yourself out of it, then it's Thanksgiving, and so now we're racing to close business in September, and we're more cautiously optimistic, but there was that piece of disruption and we'll sort of have to see how in the end it plays out.
If you step back longer term, at the end of the day, the focus on efficiency seems to be a secular trend. Particularly when it comes to technology. And I mean that because, once sort of the dust settles, it seems like it's obvious, right? Why would you have, why wouldn't you push aggressively to automate your paper processes? Why would you work through intermediaries to actually deal with software companies? You know, one of the large parts of the government, their CIO issued a memo that's called Automate or Die with like 150 applications. I forget now the deadline, but the answer was like, either you modernize or you shut it down and, you know, go.
And so that kind of government on some level has greater flexibility when it sets its mind to something than an enterprise does 'cause an enterprise can break things, is worried about processes. Or I think there's more, once there's a clear motion, there's greater ability to run. And we expect to benefit from that going forward. So it's an exciting opportunity for us.
And then, I wanted to switch gear a little bit. So you mentioned your AI story as well a little bit and how exciting you are, excited you are about that. Can you just frame it a little bit for us? Because, you know, here on stage, I'm talking to all of them. Everyone is going to be big on AI, but like what does it actually mean? Like, you know, i s it going to be? Are you going to be like building agents? Are you on a platform to build agents? Are you giving data for agents? Like, how do you, you know, d inner party, you're kind of describing what AI, Appian does on AI?
At dinner parties, I just say we're an AI winner.
Yeah. Okay. Great.
But if we were to break it down, Appian automates complicated processes. That's what we do. Now think of processes as nodes where actions need to be taken and directions in which the process continues from there. And again, we do that for complicated processes frequently across enterprise, frequently involving the external customer and on and on and on. And so at those nodes in the process, and we have plenty of lovely automation online in case you guys are curious to see, frequently there was just a human, right?
Human is a part of the process. In some situations, it's possible to replace that human or some of the other sort of tools that we've had before, like an RPA or whatever with actually dropping, you know, access to GenAI. So we don't build models. We allow customers to access them one way or the other. And for the purposes of performing very specific functions inside of a pre-designed process. So what that means is that we arm AI with the right instructions, very clear instructions. We arm it with data that it needs to make decisions, and we give it risk parameters such that if it doesn't cannot make a decision, that it doesn't. And then we work iteratively with the customer, frequently using our own professional services org to actually get it to accuracy levels that are good. Because frequently these applications start at like low accuracy levels, below 50%. Like that's not good.
You know, that's fine if I'm, you know, researching my skiing holiday in Europe 'cause then I'm actually going to go to every hotel website, but it's not fine if you're going to release it on something that gets done thousands of times a day and actually runs your company. And so think of it as, and I'll get more specific about what that looks like, but think of it as AI is a node in the process used where it's advantageous to the process. For example, you don't need to ask AI to do basic math. Basic math is deterministic, 1 + 1 = 2.
You don't need AI to do. But if there's some amount of judgment and sort of the right kind of judgment, if you will, armed with the right kind of data, then AI can be very helpful and accelerate processes. So what we've seen is success so far, the first success we've seen right now, the big one is actually document processing. And that's obvious from the perspective of people in this room. It's actually much harder to implement in the outside world because documents aren't clean PDFs built last year. They're medical records, t hey're 15 years old. They're sometimes handwritten. You know, they can be in different languages. The photocopy was bad. So, like, to actually get a document processing system to work in AI is a non-trivial task.
And they say we're succeeding, w e're having great customer references, and that's a use case that's probably applicable across most of our customers. So that's going to be a big incremental push for us into next year. The second area is our product called Agent Studio, which we just GA'ed in December, which again takes the idea of autonomous action by an agent, but under the sort of definition and guardrails that are kind of germane in an Appian process. But it gives the customer a way to build an agent that's specific for them, give it the parameters that it wants to run, and then kind of help the agent get smarter over time 'cause that's the other important thing. Like where you start is not where you're supposed to end.
That's the beauty of AI, that it ought to become better over time, and so that was our most successful beta in our company history. Several of our beta customers were basically trumpeting it for us to go GA so they themselves can go into production, and honestly, just looking at the variety of use cases in that beta, I was impressed. I was impressed in terms of what customers are already trying to do with the platform. Now, as with all things, you know, when something GA's, that's really just the end of the beginning as opposed to sort of some sort of massive opportunity that we got to go in the market, engage with the customers, build proof points, and then over time that will sort of be the second kind of part of the story. And then the third part of the story is modernization. We think that legacy app modernization is a market that has always existed, but it was small because it's hard to do.
AI offers a tremendous amount of promise to make that process easier. And we think we're naturally suited to be one of the winners only because our destination platform is the best platform. And some of the products that we already built, and our strong professional services arm, will make those projects go better and faster than competition.
And then, from a CFO perspective, the one thing you need to, I guess, worry about more than the sales guys is like, okay, how do I monetize some of that? Like, what's the pricing vehicle here? Like, are we at a stage where you can talk about that already?
Yeah. So, the first step that we talked about for the last year and change is you need to pay us more to have access to our AI features. So we have most of our customers in what's called the standard tier, and then we have an advanced tier. So if you want run AI in production, you actually need to upgrade to a standard, to the advanced tier, and that's a 35% uplift. And obviously, you know, very, very high margin. So that's the first way. So when we say customers are paying us for AI, that means that they are on the advanced tier. And the reason why that matters is because they're not paying us a little bit more. Like that, these are particularly for large customers. These are significant amounts of money that they're saying, "Yeah, less.
I'm ready to go, and I know I will get the value," so that's exciting for us to see. Over time, there'll be an element of consumption in our AI modernization. We have this concept of an AI action where when you actually ask an agent, or any other sort of AI tool to perform an action for you, that obviously consumes external resources. That obviously, you know, also is where you generate the most value, so if you go above certain limits, you will then pay us incrementally for that usage as it builds, and we see that from time to time happening right now, but it's more of a, as use cases become more robust, and people get more ambitious in what they're trying to build, that will be an incremental part of the growth story.
Yeah. Then, if you think about it, like, how do you think AI is going to be like a growth leader that sits on top of that? It does feel more listening to you. It kind of attracts the whole. It brings everything up basically because it's kind of interwoven with that.
I think it's both. It's interwoven.
Yeah.
I think that's the right way to think about it.
Yeah. Okay, and then the other thing I wanted to talk about is like now, differently, now that you're kind of looking at the organization, you have a cloud business, you have a self-managed business. You know, like historically we, you know, like on, in my shoes, it was always like, "Oh, get the self-managed off as quickly as possible, bring it into the cloud, and then you have more control," etc, and you had it with your previous company to some degree. Where are you on that? You know, how's your picking there for the Appian story?
So, roughly 80%-90%, depending on the quota of our new business, is in the cloud. So our cloud transition on the subscription side is like, it's very mature. Like we're most of the way there. I however think that there will be an important element for self-managed, frequently on-prem deployments for a long time to come. Our self-managed business continues to grow slower than the cloud, but it does continue to grow. And ultimately, you cannot tell a customer how they want to deploy. All you want to do is align to their IT strategy and make sure that they pick your technology. And that's particularly relevant for us. Like, because in addition to the government, our big verticals are financial services, insurance, and life sciences. Those are heavily regulated industries that are always going to be laggards going into the cloud.
So we expect to, I mean, some of the greatest wins that we've had as a company this year were actually on-prem deals. So I think of that as definitely a part of the story and something that, you know, we're not in, by the way, we rarely see customers take an existing workload and move it from one to the other. You know, that I think will happen in some long future, but we haven't really seen happen very much. And so in the meantime, we're just driving adoption wherever we can. I will say one thing as the CFO, the on-prem business does give me a little bit of heartburn 'cause 75% of it is upfront. And so it's just going to be a bit more lumpy than in the ideal world you and I would've liked. But at the end of the day, it all evens out and it's not the majority of the business. And it's good business, so we're happy to take it.
Could you, well, sorry, I'm not giving the playbook of a CFO we know, but is there like a multi-year, single-year aspect there? Like.
Great question.
Yeah.
No is the answer. Thank God. At least we don't have that.
Yeah.
We actually, for our multi-year on-prem contracts, have structured them with the customer such that we only recognize the first year upfront. We got to get in, we are getting into the accounting minutiae. Like we've written the contract such that we and our auditors are comfortable that we can just take the one year. Because then if we were taking multi-year deals, of which we have many, that would create the volatility in their line to be even greater, in which case we would probably need to think about a different way to look at the business, like ARR or something. But I'm happy to say we don't have to do that.
Yeah. I know a company that you had that before. Changing gear a little bit the last few minutes. It's like, profitability. I have to say before your time, that wasn't that much of a focus at Appian. It changed a little bit before you joined. Like where are you guys on that journey?
Yeah. So again, I actually think it's changed quite dramatically before I joined. It may have just become more obvious in the last two quarters that I've been there, but look, Appian in 2022 and 2023 continued investing aggressively because, when other software companies were pulling back, right? Like I forget now when the ZIRP era ended and, you know, everybody got the memo that, you know, you got to focus on profitability more than you have in the past. Appian chose to ignore that memo, and by the way, hindsight is 20/20, but at the time, that's not an entirely rational decision. Strong product, large market, differentiation, surveil zigs, then why wouldn't you zag? But the execution just wasn't there to support it, and the losses expanded, and then the company did the right thing, which was to pivot, and one thing that's great about Appian, Appian can pivot quickly when it gets conviction about something.
That, I mean, I give Matt a tremendous amount of credit for that. And so we've had reductions in the size of our workforce. We became extremely scrutinizing when it comes to any incremental headcount. And that's the world that Appian has lived for the last two years. And that's why margins have expanded the way that they have and operating expenses basically didn't grow. And so the trick for us is now to find that middle state, right? The middle state between feast and famine and focus on moderate growth and sort of investments that will drive top line while also leaving us room to, you know, improve profitability over time. I think we can do that.
And again, the attitude of the company is like, we all understand sort of how important profitability is. And when I have discussions around, you know, where are we going to draw that line on the list of proposals that we're going to fund for next year's budget, right? There's no debate where that line is. It just, the debate is what's above versus below it.
Yeah. Okay. So the fear we always had, like, it's not that Appian didn't get the memo. We just kind of, you know, decided to ignore the memo. and that kind of was kind of what scared a lot of investors is like, well, there was a decision and it was, it felt that, it felt on our side it wasn't rational.
Right.
Like, you know, obviously you joined the team now. Like how much of a kind of how real do you think that drive is? Or do you know, it's a little bit with Salesforce, with this kind of Salesforce there that you think like what's coming next? Is it, are we, you know, what's your thinking there now, like being part of the organization?
I can tell you that that's not a conversation that is happening. You know, when I said we need to expand margins, there's just nodding in the room.
Okay.
So that's what happens. Now I just do want to throw it in there. Given how much we've expanded margins this last couple of years and frankly the revenue outperformance was greater and there's a speed to reinvest that is always a little bit lagging, we expect a relatively modest improvement in margin next year. But over time, like, we, we have to do both. We have to do both. We have to grow and, and we have to, we have to expand margins. And, you know, a company with our unit economics should be perfectly capable of doing that.
Yeah. Okay. That's a clear answer. Okay. Then last couple of moments. Capital allocation, like, how, you know, when you kind of came into the business, looked at The situation there, like what's your thinking in terms of like where you want to go from here?
This is the thing that doesn't get talked about as much, even with investors who know us, which is that this will be the first year that Appian has generated meaningful cash flow. We needed to fund the growth when we were unprofitable. We did that in the early years through equity issuance. More recently, we've issued some bank loans, but now we're in a moment in which like we're generating cash and we are approaching a moment where we're like net debt zero. Like we still our net debt is still a slightly positive number. I would prefer for it to be a negative number. We have more cash than that simply because that's a healthier way to run a tech company. Also, occasionally that ends up being a conversation with our customers just so that we can tell them like, "Look, we're, you're doing better in the next 10 years l ike, w e have the financial strength to support you.
And so we're approaching that moment. And then look, I don't think it's hard to see what we think about capital allocation 'cause we've done it. We did a $50 million buyback last year. We funded with that, and then we did two $10 million buybacks this year to offset dilution. So, you know, I'm not suggesting any sort of significant changes over time, but at the end, Appian will return capital to shareholders, especially since, in 26 years, we've done two tuck-ins of the company. So M&A is not a core muscle that we have or need.
I mean, is there an argument to keep some powder dry on that one though, given how quickly AI is moving? I mean, we see it in other parts of the industry where it's like, "Oh shoot, there's an AI startup that could be interesting." That kind of accelerates M&A.
Sure. And it's not that we don't look at anything. We have a small corporate development team that is constantly tinkering. The reality is just it's going to be hard. Right? You know, we are of a certain size and a certain valuation, particularly in the world of AI, you get to large valuations with, without anything resembling revenue. So, that's always going to be hard for us to stomach. But, you know, could I tell you for 100% certainty there isn't some world in which like a perfect, you know, round peg shows up for a perfect-
Hole.
-round hole, but I wouldn't even know what that hole is right now.
Yeah.
So in other words, we'll be opportunistic, but it seems unlikely to us that we would do anything with size.
Yeah. Perfect. Hey, that's a good closing statement as well, Serge. I really enjoyed our conversation.
Excellent.
Thank you.
Thank you, Raymond.
Good to see you again.
Thank you.
Thank you.
Have a great week.
Thank you.