All right. Good morning, everybody. Thanks for sticking around for one of our last sessions, but certainly not the least. We have Amit Mathradas, CEO of Five9, and Bryan Lee, CFO. Thank you guys for joining us this morning.
Thank you.
Amit, let's start with you. Five9 describes itself as a voice-led enterprise platform for customer experience. For those newer to the story, how does the platform fit into the contact center today, and where do you see the company's strongest right to win as the market evolves?
Yeah, really good question. Look, Five9 has been around for over two decades, building a cloud-native, voice-first platform that focused on the contact center. Over the years, we've evolved that to plug into digital, to AI, and now actually bringing all facets of that together and serving that up with visibility tools like Workforce Engagement Management, analytics, and other things that are required to monitor, watch, and grow your contact center. The other big piece of it is we deliver this through an open ecosystem, plugging into integrated systems, plugging into complex systems of record where you drive the outcomes. To your question, where do we have the right to win?
I think in today's world, when you are looking at a deployment where humans, agents, and tools need to come together to drive complex use cases, particularly in regulated industries, I think is where we do really well.
Okay, great. It's a great level set. Let's talk about just sort of the market bifurcation between on-prem and cloud. Roughly 60% of the contact center market is still on legacy on-premise systems. How is AI changing the urgency around cloud migration, and are customers increasingly looking to modernize their contact center and deploy AI as part of that same decision?
Yeah, look, as you said, 60% of our space, 50%, 60% of the space is still on-prem. I always joke that if you haven't gone to the cloud, till now, you've really not wanted to go. That has been shifting with AI. As AI is becoming more pervasive in the contact center, particularly Voice AI.
I think what customers are starting to realize is that the architecture for on-prem really does not support the best outcomes for things like the new generation of Voice AI and other capabilities. That is, in a lot of cases, forcing customers to lift their heads and say, is it time for us to move? Is it time to consider moving to the cloud? In a number of cases, they come to us and test our AI capabilities first, followed with the CCaaS component and the connection points that come beyond it. We are starting to see that driven with the shift in AI.
Okay, great. Maybe a bigger picture one. One concern around AI is greater automation will reduce the number of human agent seats and pressure traditional CCaaS spending. You've described the future as Humantic CX. What role do you expect human agents to play as AI handles more interactions, and how does that combination expand the value Five9 can deliver?
Thank you for that question, and I think it's core to how we see the world and where we are going. I always start with, think about this massive conundrum of companies, including ours, spend millions and millions of dollars of marketing money and others to attract prospects to become customers. When they become customers, we tend to then spend as little as we can engaging with them because it moves to an OpEx center. What you are seeing, what is happening with AI is, it is changing the economics of the contact center. A lot of the base use cases, the repeatable motions are now going to AI, Voice AI, other toolkits. What that does is it is leaving humans to go drive and deliver more of the complex use cases.
The ones that in some cases in regulated industries, it is mandated that you have to still engage with humans. For us and where we see our role in that, what we are seeing with customers is in a number of cases, they are using this economic shift in the contact center not to remove dollars, but to start reaching Tier 2, Tier 3 customers that they could have never afforded to engage with before. So interactions are dramatically going up. This symbiotic relationship between AI and humans is where we see the world, and that is how we see it, which is it is not humans, it is not agentic, it is Humantic.
You are not going to see a world where it is binary. AI is going to flow to humans, back to AI, and that cycle will continue. That is the evolution we are seeing in our space. I think the data that humans produce and the data that AI produces is going to make each of these pieces better. So that is the way the world is going.
Bryan, how do you think about that in terms of the future seat growth outlook? Because arguably, if interactions are going up, but agents are handling more of the initial call volume, maybe you are routing to fewer agents in an end state. Just talk us through how you think about that vision from more of a-
Yeah.
agency lens.
Today, our human agency count continues to grow at a healthy rate. If you look at our second quarter results, our CCaaS subscription revenue grew 7% year-over-year, and the human agency count was right in line with that. But, for our customers, if they look ahead three, four, five years, they too actually don't know exactly what that optimal mix will be between human agents and AI Agents, which is why we actually introduced this revenue commit model.
It basically gives them the flexibility to kind of shift the mix between the two buckets within bounds, and so it gives them that flexibility, it gives them the visibility, and it gives us the predictability for revenue as well. It's really a win-win situation. This has been resonating really well with customers, and perfect examples during the second quarter was we had a $100 million TCV win that was on this five-year revenue commit contract. We actually have one of our largest renewals ever, where an existing customer went from month-to-month contract to a five-year revenue contract with expansion.
How does the actual contract work? Are you licensed for a certain number of seats and then you have a capacity commitment on top?
There is a commitment level in terms of, you start out with a certain level of seats, certain level of AI Agents, but there is certain bands within which they can actually flex and still just commit to the certain level of spend with us.
Okay, got it. Amit, back to you. Five9, you mentioned this, just in terms of regulated industries, you have been concentrating resources on complex enterprises, particularly in regulated industries, financial services, healthcare. What makes those environments especially attractive, and which elements of the Five9 platform matter most when customers deploy AI at scale?
Look, when you think about regulated industries, the way we talk about it is a lot of them are dealing with the valuable and the vulnerable. What that really implies is that human in the loop is a critical component for them. Two, these industries have a very high bar for regulatory needs, governance, security, and we have been building alongside them for 20+ years delivering all those capabilities. When you do that, and you build along these customers for 20 years, you also start understanding their use cases. What are they using their orchestration for? What are the workflows that they are building? What are the partnerships and capabilities and connection points that you need to have to build into your product? When we are looking at that, these are things that we do really well.
For us, it's really about how we help these sorts of customers with the complexity on the human side, on the agent side. What brings this all to life, and I think the second part of your question was, what part of your offerings help with that? It's all of this. It is bringing together these products that are driven by governance, security, regulations, understanding their workflows, and one that connects humans, digital, and agents all together, serving that up.
Okay, great. Bryan, let's jump back to you. AI revenue is now growing much faster than the core CCaaS business and represents a growing portion of subscription revenue. As the mix shifts towards consumption-based AI, how should we think about the implications for revenue durability, customer expansion, and really just the overall economics of the model?
Yeah, absolutely. Just to level set, in the second quarter, our AI revenue generated more than $150 million in terms of annualized run rate revenue, grew 78% year-over-year. That was an acceleration from 68% that we reported in the first quarter. Now it makes up 15% of subscription revenue versus 9% a year ago. So a lot of strong momentum across the board. This is with new logos as well as existing customers. Our million-plus new logo wins, virtually all of them have AI attached, and that trend has been continuing for many quarters now. On the existing customer side, we've instituted what we call an AI Blueprint program.
What we do there is essentially we go into a customer environment, we turn on Five9 AI Insights, which is our product that goes and clusters and analyzes every interaction that's happening in the contact center, and it identifies areas of inefficiency. We can go in and recommend other AI products that we have to help solve those issues and to increase ROI. So that's been resonating really well with our existing customers in terms of driving adoption of our AI products. From a pricing model perspective, AI is all consumption-based, all of our AI products, and it's usually done in a block of minimum committed minutes or whatever the units may be with any kind of overage charged at a higher rate.
Of course, when they do go over, typically they'll come back to us and increase the minimum commit with us very quickly. But as I mentioned earlier, the revenue commit model is what we're actually standardizing toward now, so that'll give us even more durability and visibility into revenue going forward.
Perfect. Amit, let's talk about competition. AI native voice vendors are moving quickly and in some cases, delivering pretty compelling results. So as enterprises evaluate those offerings alongside broader CX platforms, which parts of the market do you see as most vulnerable to point solutions, and where can Five9 sustain a differentiated role?
Yeah, look, I think one thing about point solutions in any technology shift, point solutions always first to market, they will bring out the new capabilities fastest. I think what we are starting to see is, particularly in the industries that we serve and the customers we serve, if you believe that you are going to have humans in your contact center solving complex problems, they are coming to us and saying, I need your AI to work alongside. If you think about the contact center and a customer or a patient is calling in, the first thing you're doing is you're calling on your phone line. So the baseline, even on what AI runs, has to be great telephony, great routing. The infrastructure has to be solid even for your AI to work.
So that's one capability that we bring to market in droves. Two, the ownership of routing, and I'll give you a very small example, becomes critical. So think about the fact that you have called in and you have a very complex billing situation. And you've called the company maybe two times, you've been on the website, and you're irate. When you call in and you talk to a third-party AI solution, what it recognizes is complex billing situation, customer may be irate, let's send them to the billing queue. When you call into a platform like Five9, which has our agentic routing coupled with Agentic Quality Management, and what that means is I have gone and looked at all your human agents and scored them, human and AI agents, scored them. Who scores high on empathy?
Who scores high on billing situations? At that point, our agent, because it is connected on the same platform, can say, I am routing this specific question or this specific customer to this specific agent that scores high on empathy and does really well on billing situations, not just to a billing queue. That's the differential. When I think about point solutions, the way I describe it is, they've already come out and shown companies that you can get to 30%, 40% containment. They are like the hammer, right? We are like the entire toolbox. How do you get to 32? How do you get to 42? The next tranche of containment of resolution, that's where platform shine, that's where architecture shines, that's where owning routing, telco, all these pieces really matter.
Okay, very clear. I want to talk about this $100 million landmark agreement with the Fortune 100 financial services company that you landed. I think you worked closely with Google on this. What did Five9 prove through that process, and what needs to happen for that joint go-to-market motion with Google to become a repeatable source of large enterprise wins?
Yeah, I think what it's proven is we've done this before. What it's proven is, this is the second or third very large Fortune 100 customer that we continue to win within these industries. To me, it's proven that, hey, we have the right to win. We have the best of these products. Customers of this size are choosing us for the end-to-end solution. That's one. Two, I think in partnership with Google, look, this was not just a, Hey, let's sign a piece of paper and do a partnership and go in. This was like hands-on-keyboard.
Integrating, building, delivering. From an ongoing basis, from a go-to-market perspective, what we are working with Google on is how do we continue to get closer on the integration? How do we continue to present as one partner, not two? Feedback from this customer to me was, hey, one of the reasons we went with you is because you're presented as one solution, and we obviously have trust in you. We have trust in Google, and this really made sense based on where we wanted to go. By the way, this was one of those on-prem to cloud migrations that we were talking about, where it was finally kind of being unshackled for the reasons we spoke about.
Excellent. Awesome. If anyone has a question in the audience, just feel free to raise your hand. I am happy to flag it. Bryan, the current growth outlook is supported by a sizable backlog of signed business moving into production. As you look beyond the conversion of that backlog, what gives you the confidence in the durability of growth, and which parts of the business do you expect to drive the next leg of expansion?
Yeah. As you said, coming into 2026, we had a very strong backlog that was comprised of new logos, install base, as well as CCaaS and AI. The acceleration throughout this year, the shape of that really has been driven by the schedule of the ramps of each of these customers that are in backlog converting to revenue. I think looking ahead, the best way to think about it is this, and we touched on this a little bit earlier. If you look at the CCaaS market, it is $24 billion. That is 40% penetrated. The remaining 60% is mostly upmarket and international, where are key areas that we are focused on. Then if you look at Gartner estimates for AI and CX, that estimate in the next three to four years essentially is another $24 billion plus.
So more than doubling our market opportunity going forward. We will continue to execute and invest in ways where our goal is to drive stable revenue growth on the CCaaS side. AI will remain the fastest growing part of our product portfolio, but we are expecting that to grow in multiples higher than where CCaaS is at. So we believe there is a really long runway for durable growth ahead of us.
One thing I will just add to that, just the size of the market, and I think this goes back to your earlier question around humans and AI. The one thing for us is, look, over time, will there be compression in human seats? Sure. We are well positioned to handle that because, as Bryan mentioned, $25 billion CCaaS, $25 billion AI, we play in both parts of that spectrum. The dollars don't leave the contact center. All they're doing is maybe moving from one bucket to the other. If the TAM has really doubled, and you play in both pieces, that's really where we're focused on.
Bryan, there's been a lot of discussion this week just around the use of frontier models for, let's say, bleeding edge intelligence tasks relative to maybe open source being sort of more of the workhorse. I guess, how are you guys managing your own sort of model orchestration internally today, and how does that kind of change in the future?
Yeah. I'll start, and please chime in, Amit. We are engine agnostic, so we have a team that's actually benchmarking all the different LLMs out there that's supporting different use cases. We've gotten really good at optimizing the different engines for whether it's a certain use case or in terms of optimizing cost. So what has happened in terms of the AI, how does this impact our AI gross margin at the end of the day? If you look at our AI revenue portfolio, the largest portion is made up of AI Agent revenue. If you look at the gross margin of that, it's in the high 70s. From an LLM cost perspective, we typically get into a one-year renewal with the provider, and the cost trend has been stepping down each year.
We believe that as AI becomes a much bigger portion of our overall revenue, that's going to be accretive to total gross margin.
Anything to add there, Amit?
No, I think kudos to Bryan and team who are managing this. We are, as you mentioned, engine agnostic, and as we see costs come down, we are taking advantage of that by shifting back-end workloads. So far so good.
All right. Excellent. Amit, you've sharpened the company's focus around becoming a voice-led enterprise CX platform. We've talked about this, and you've also made several organizational and leadership changes to support that strategy. Where do you see the biggest opportunities to improve execution, and what evidence should we look for that those changes are having the intended impact?
Look, from the first earnings call that I did, I mentioned four areas, right? Working on our culture, improving our culture, operational efficiency, improving the core, and then really driving AI success in the future. As I think about it, a lot of this organizational change in design was really setting us up for where we are going. We've brought in a new CTO, and for the first time at the company, we've put product engineering, architecture, AI, all under one roof. This serves two purposes. One, it's a cultural purpose, like how do we all get aligned, drive, build, move things forward? Two, from an operations perspective, it's about removing roadblocks, right? Faster delivery, quicker delivery, moving, adapting to what's happening in the market, building for this unified Humantic ecosystem pretty quick.
On the sales side, we brought in a new sales leader, Rob, who is fantastic. His real back-end pedigree was driving enterprise and thinking through both mid-market and enterprise. As we shift to winning more of these larger deals, how do we reorganize and build for that? As you think about those buckets, that is where I am focused. In the core, as you just mentioned, we have this large win. How do we do that on a more repeatable basis? In AI, we are growing in multiples faster than our core CCaaS. How do we sustain that, build that? What is the next capability that comes out? I think things that you all should look to us to hold us accountable to is, one, can we sustain this level of AI growth faster than the core?
That is one. Two, how do we actually evolve our platform and move towards the end state of Humantic, where interactions are becoming far more important quicker? What features functionality are we delivering? I think that is another one. The last one I would say is, look, at the end of the day, I am a capital allocator, right? Are we moving our dollars from the right places in the business to go unleash all these things above that I mentioned? I think is another one that you should be looking at.
Okay, great perspective. Bryan, back to you. Five9 has delivered some solid margin expansion while continuing to invest in AI, the enterprise go-to-market motion, and the broader platform. How do you think about the balance between funding those opportunities and continuing to expand profitability and free cash flow?
Yeah. No, great question. As Amit mentioned, we are going through a lot of different org design change efforts as well as executing on long-term cost savings initiatives. Just to give you an example of some of those, we are increasing the mix of our offshore hiring and leveraging the global footprint that we have. We are improving our spans and layers to improve the speed of execution. We are increasing automation to improve efficiency. All of these come with long-term cost savings. Up front, though, there are a lot of temporary overlapping expenses that we are working through this year, most of which will be done this year.
Going forward, with the efficiencies that we have, we are planning to reinvest portions of that surgically into key areas like AI innovation and go to market to make sure that we continue to drive strong revenue growth. But at the end of the day, we've always been a balanced growth company. So we'll focus on driving top-line growth, but also expanding margins in the long run.
And Amit, just based on your observations, how mature is your internal AI strategy today? What types of efficiencies are you generating across those different buckets of.
Yeah. Really good question. I think like a lot of companies our size in our space, I would say there's more work to be done. I often remind people that AI within coding has been around for a few years now, but use cases in other functions are only picking up steam today. So I'm not waving the victory flag and saying, hey, everything is agentified. I think we have more work to be done, but we are definitely on our way.
Is there anything that changes in terms of hiring philosophy going forward?
Really good question. I am not going to get up and say things like, oh, I don't think I'll need as many people, because I don't know if that's true. I actually feel like what we are seeing is if I can code faster, if I can drive 30%, 40% improvement in velocity with an engineering team, then maybe I can keep all the engineers and produce 30%, 40% more code and build faster for this evolving market. There may be opportunities for optimization as well. Not saying that's off the table, but I think that will only unlock once we reach certain milestones, and I'll feel comfortable sharing that then.
Okay. Very interesting. To close out, let's pull all this together. If we look out three to five years from now, Amit, what do you think will change most about how enterprises manage customer interactions, and what role do you want Five9 to play in that future?
Yeah, look, I've been very clear around where I think the space is going, and there's a lot of moving pieces, so I'd say I hope we're right. But the interaction layer is becoming far more important. It is moving to one where customers are realizing that the connection points between humans, AI, digital orchestration is going to be critical. We clearly play a role in that, and that's one that I am seeing shifting right now, and that connection between running it all on one platform.
Two, I actually do believe that a lot of contact centers are going to move from OpEx centers to revenue generation centers because when you can start driving interactions with Tier 2, tier 3 customers, improve NPS, reach more of your base than you could never afford to before, I think it presents tremendous opportunities for our customer base and prospects. And I feel very strongly that Five9 is well positioned to take advantage of both those new trends that are coming out, and so that's what we're building for.
Fantastic. I think that's a great place to leave it. Thank you both for joining us this morning. Really appreciate it.
Thank you. Yep.
Thank you, Matt.