Hi, everyone. Thanks for joining us today. My name is Scott Berg. I lead the enterprise software and SaaS research efforts here at Needham. Today with us, we have Five9. We have the company's CEO, Mike Burkland, and the company's CFO, Barry Zwarenstein. Gentlemen, thanks for joining us here today.
Happy to be here, Scott. Thanks for having us.
Definitely. I know before we get started, I believe Barry wants to go over the Safe Harbor Act.
Yeah, so today we've been making forward-looking comments about the industry, about the company, about different trends, including AI. These are, of course, subject to revision, and we undertake no obligation to update it. The risk factors that could cause this to differ are with our Securities and Exchange Commission filings, 10-Qs and 10-Ks, which I encourage you to reference, and we undertake no obligation to update.
Excellent. One piece of housekeeping: we will take Q&A from the audience when we're done here. If anyone wants to ask a question, feel free to put it into the question box on the webinar window or email them directly to myself at sberg@needhamco.com. I guess with that, let's go with an overview of the company. We can start with that for the maybe one or two people on here that's not familiar with Five9.
Yeah, happy to do that, Scott. So Five9 is a billion-dollar SaaS company. We provide a software platform, which is AI-powered, for some of the largest enterprises in the world to enhance and elevate their customer experience. In other words, the customer experience they're delivering to their customers, their end customers. And this is a massive market. We've been in the contact center part of CX for quite some time, replacing legacy solutions. Our AI is becoming a big part of the mix of our business. We'll talk about that further. To give you a couple of other metrics, subscription revenue growth, which is about 80% of our total revenue, grew 20% in the third quarter.
And AI is becoming a big part of our business and the mix, as evidenced by being 20% of our ACV bookings for net new logos in the last quarter, in the third quarter, that is. So exciting times in our industry, and happy to dive in wherever you like, Scott.
All right, let's start with product and/or the industry here, because I think most on the webinar here are certainly interested in what's going on with the AI side of the house. But let's cut to the big question, I guess. The debate about AI's impact in the contact center space has been well documented. Generally, why isn't AI a threat to Five9?
Yeah, AI is a tailwind for our business and a TAM expansion. We've talked about this a fair amount lately. Think about it simply as, look, as AI has come on the scene, it has enabled us to deliver technology that provides things like self-service and agent assist within the customer experience. And if you think about it, our customer base is always looking for ways to automate and provide more and more self-service. If you look back in time, websites, for example, automated a lot of interactions for large brands and large companies. Think about airlines for a second, right? And AI just provides a whole 'nother level of automation for some types of interactions. Most of our customers are attempting to deflect, if you will, some of those interactions.
In most cases, it's a fairly small fraction, marginal percentage of interactions, somewhere in the 5%-10% of interactions to self-service, whether it's a voicebot or a chatbot in the self-service use case. We provide technologies. We're the leader in AI. We acquired a company called Inference four years ago that was the leading IVA or virtual agent. We've built additional products on top of that platform and integrated it to our platform. We have 10 different AI SKUs now. It's becoming a meaningful part of our revenue mix. As I said, 20% of ACV bookings in the third quarter. 50% growth in bookings. 50% growth in bookings to our install base of those AI solutions that we sell. 40% growth in AI revenue on a year-over-year basis. We've got a number of metrics that we've shared with investors.
But in the end of the day, look, we provide software for interactions for these brands to handle customer interactions, whether there's a human agent on the other end of that interaction or there's AI powering that interaction instead of a human. And we're providing software for both of these types of interactions and oftentimes a blend of those interactions. And we monetize both of those categories, so to speak. So the more automation that occurs, the more interactions that are deflected to self-service, that's actually we're getting more revenue per interaction for AI-powered interactions versus the software we sell for human interactions. So that represents a TAM expansion. If you think about a situation where you've got, say, 15% of interactions deflected to self-service, which is a very large percentage, by the way, that would result in about a 30% TAM expansion on a net basis for us.
Okay. In the contact center space over the last couple of years, I think we're all reading a lot about new startups that are kind of more point solutions that are coming up in the space with maybe different types of AI functionality. But can you discuss how or why a platform player like Five9 wins against these point solutions in these AI-based scenarios? And then maybe over the longer term, why will or why should Five9 be an AI winner against whether it's the current ones or maybe new ones that come up over the next couple of years?
Yeah, well, first and foremost, Scott, we partner with a lot of these point solutions because they want to ride on top of our end-to-end platform. AI is only as powerful as the contextual data that it has access to. And that's really important. If you think about a point solution that's not tied into a platform like ours, then think about it as tunnel vision, right? It doesn't have the contextual data such as interaction history, right, that consumer has had most recently, real-time interaction data, customer information data, which might be in a back-end system. All of that's available in our end-to-end platform. But it's also access to all the channels of communication, like a simple chatbot.
If it's standalone, if it cannot escalate to other channels like voice in particular, you've got to be essentially integrated to our platform or using our AI in order to deliver that truly personalized AI experience. So again, a lot of these point solutions are coming onto the scene. A lot of them are partners of ours. They're trying to plug in and they want to plug in. We've integrated many of them to our platform. We obviously have our own AI solutions that we position. But again, our approach is to do whatever is best for the customer. So if we're working with a large brand and they're trying to enhance their customer experience leveraging AI, we're going to be a trusted advisor to them and we'll help them deploy whatever the best solution is to drive that automated customer experience and deliver accuracy and efficacy with that AI.
Okay, so I was out at the company's headquarters a month and a half ago. I sat with both of you, got a chance to look at your new agentic platform that you released during your European conference. I think the question that's come up the most since then is, why would a customer prefer to go with Five9 in these particular areas versus a Salesforce's Agentforce strategy that's out there today?
Yeah. Well, first off, agentic and AI agents are the next generation of what I've been talking about before, which is virtual agents, right? Whether it's a voicebot or a chatbot, AI agents are really the next generation. They're powered by generative AI. And that allows our AI agents, if you will, to be more autonomous, to do more reasoning and decision-making. And that just opens up the aperture for the complexity of interactions that we can now handle with self-service through these AI agents. And it's a major leap forward for us in our AI technology. But when it comes to partners like Salesforce, as well as some of the point solutions we just talked about, in the end of the day, as I said, look, Agentforce is front and center for Salesforce. They are leaning into our partnership. We've been partners for many, many years.
We have something like 1,200 common integrated customers with Salesforce. Salesforce is very anxious to partner with us so that in the end of the day, just like I said with these point solutions, their AI agents, once they're mature, because they are still early in this game, can plug into our platform. And they do that so that they have access to all that contextual data that's in our platform, that interaction data, whether it's real-time or historical. But it's also an end-to-end path for that consumer. So if they're in a chatbot and they need to escalate to a voice call, they do that through integration to our platform. And the way they integrate to our platform is actually via our APIs, which we call VoiceStream and TranscriptStream. And we monetize those APIs on a volume basis.
That's a very significant revenue opportunity for us. In the end of the day, look, customers have to run a customer service organization. They have to have a contact center. They have to have a CCaaS platform, and they have to have a CRM system. You just can't get by without both of those critical components. When it comes to the AI portion, in some use cases, eventually, when Agentforce is mature, that may be the better AI solution for a given use case. In other cases, it's going to be Five9's AI and our AI agents. In the end of the day, we partner with Salesforce very successfully. If you think about it, the greatest analogy here, Scott, is the digital channels. We've been partners with Salesforce forever.
We both offer digital channels, and we partner very effectively in accounts to make sure that whatever works best for the customer is what works best for us. And quite frankly, if it's their AI, we still win because we monetize those APIs at a very significant price point. It typically equates to about $40-$50 per AI agent that we generate in revenue to us for that integration. So we either monetize it through our AI solutions or we monetize it through those APIs to integrate to their AI solutions.
Very helpful there. I think with the advance of reasoning in some of these technologies, it's giving customers the ability to automate more through these technologies. But when I was out there a month ago, trust was a big component of our conversation, right? How do you get a customer to build further trust in the technology and extract the full value of such a solution? You gave kind of the dial analogy, but what does that look like really in action?
Yeah, I mean, we call it the dial of trust. But essentially, this is where technology and people come into the equation. Our experts are very important in this equation, our implementation teams and experts, and our technology. We'll talk about GenAI Studio for a second. That is our design studio. It's a tool. It's a technology that we deliver as part of our platform for our customers to use. But our implementation folks can help customers do this as well. And it allows for testing of a voicebot or a chatbot or other AI solutions as you're deploying them or before you actually put them into production. And in the end of the day, trust is built in several ways. But one of the most important parts of trust is, and when it comes to AI, is accuracy, right?
If you've got AI and it's not accurate, it's not going to be trusted by the consumer for sure. And it's definitely then not going to be. You're not going to build trust with our customer, the brand. So we have the technology to do that testing upfront, make sure things are accurate, fine-tune the AI if necessary to make sure accuracy levels and containment rates are high. And again, part of that is the expertise of our people that allow enterprises to achieve that level of trust.
Last question for me on product in the industry at least is talking about pricing kind of in the space, right? We've seen pricing start to evolve over the last year, but how do you all anticipate pricing to evolve further, especially within the AI functionality? I mean, we've got seats, we've got some transactional components, something else consumption-based. I guess, what are you seeing today and how does that maybe morph more in the next three to five years?
Yeah, great question, Scott, and it is changing. It's a morphing marketplace when it comes to pricing schemes, and look, we give ultimate flexibility to our customers. All of our AI solutions are available on a consumption-based pricing model. Usually, those are pre-committed packs of consumption or volume capacity, if you will, but in the end of the day, I think the AI world especially is not everything's tied to a seat. If you think about it, a lot of the products we're selling, these 10 AI SKUs, most are not actually tied to a human seat. Agent Assist is, but it's actually priced on a consumption basis, volume basis in terms of number of interactions or number of minutes that we're assisting a live human on.
But if you look at some of the other products like AI agents that are self-service, right, those are software solutions that are not tied to a human seat. So I think seats, again, will continue to kind of fade to the background in a lot of the models, if you will, and a lot of our key metrics for our business. And subscription revenue is the key. That is the indication of how many customers are coming onto our platform and how many products they're utilizing from us. And if you look at our subscription revenue, again, 80% of our revenue mix, 20% year-over-year growth in Q3. That is the key metric that we want folks to really pay attention to because it's the most meaningful metric around our business.
Right. Moving to go-to-market because you certainly had some changes to the business over the last couple of years. The first thing I wanted to touch on was at your CX Summit in Barcelona, the recent one here, when you announced the new Five9 Genius AI Suite, you also discussed a four-step kind of strategic process to deliver AI business value to your customers. Can you maybe touch on what this approach actually entails?
Yeah, absolutely, Scott. Again, not only do we have this Genius AI Suite of AI products, right? That's the 10 products that we provide. We've also got a Genius AI four-step process to help, essentially, to help these brands navigate and give them a blueprint. We call it the AI Blueprint for how to deploy our AI solutions to enhance customer experience, self-service, and drive efficiency as well. So it's about customer experience, but it's also about driving efficiencies and cost savings, and the four steps are straightforward. And we actually, it's not just a process. It's also tools and technology. For example, we'll go into a customer, a prospect, a large brand, and actually turn on some of our technology. We turn on VoiceStream, and we turn on AI Insights. And we can run it for one week to four weeks, usually.
We'll basically have a record of all of their customer interactions during that time period. And with AI insights, then we can cluster those interactions to analyze and basically identify areas for using AI to drive automation, for example, right? Whether it's voice self-service or chatbot. And then we also leverage our process from there to implement and test. We talked about GenAI Studio, but there's a phase where we're essentially implementing the solutions for them. And that's professional services plus technology. And then eventually, we're applying those and putting them into production. So it's a pretty defined process, but our customers love the fact that we're advising them as the AI experts to do this.
I guess, what impact do these newer AI technologies have on kind of sales cycles right now? Obviously, you have a big perpetual installed base out there that's ultimately got to move to the cloud. But are you seeing any impacts maybe over the last couple of quarters on those sales cycles, either faster or slower? Because my guess is I think.
Yeah, I think it's safe to say, Scott, back in the first half of this year, I think there was a little bit of AI fog, as I call it. Folks were just trying to figure out AI. There was a learning curve that most of these companies had to come up on. And it was a bit of a distraction. And it really peaked in Q2. We're starting to see a very different behavior when it comes to companies understanding AI, understanding the use cases, partially because we're helping them understand it, right? That's our job as trusted advisors. But I think for the most part, companies are realizing that to really leverage AI to deliver hyper-personalized, automated customer experience, for example, a shift to the cloud is important. It's hard to do this in the legacy world.
It can be done, but it's not going to be as effective, and so AI is a catalyst for continued migration in our core CCaaS market off of the legacy on-prem systems and into the cloud with companies like Five9, and yes, it's one more component in the sales cycle now, but that fog that we talked about in the first half of the year is lifting, and it's really not delaying sales cycles. It's just becoming a component within them now.
Okay. I know if you go back to last year, I believe it was on the first quarter call, you all had kind of formally announced winning what I believe was the largest deal ever in the contact center space, right? Now, you've had a couple transactions that have been in a similar zip code at least, right? $35-$40 million. So maybe that part of it's not new. But maybe can you help everyone understand why are you winning these deals now? And what else do you have to do to continue winning deals of this size? Because if you went back two years ago, this is new territory for you all.
Yeah. Well, the market has opened up, Scott, and these very large enterprises, and we've talked about several of these mega wins, as we call them. That one in Q1 was an over $50 million in ARR win with a large financial institution with, I think, $70 million-plus customers themselves. It's a market that is really exciting to be in. We're winning because of things like scalability and reliability of our platform, our leadership in AI, and frankly, our people. We are the trusted experts. Our implementation teams are second to none, and the NPS scores reflect that, so in the end of the day, it comes down to trust, as you said. These large brands are trusting us because they know that we're the best partner to help them achieve their goals when it comes to customer experience, and so it's not just technology. It's also people.
And it's obviously AI is front and center in that too. Those are all reasons why we're winning. And I would say, look, to continue to win up market, we're in very good position because we have highly referenceable large customers, right? And the more you have the proof points, the more you're going to be able to win the next opportunity. We're obviously always enhancing and extending our leadership position. When it comes to our platform and AI, we're investing heavily. And that will also contribute to our ability to keep winning market. And I'll leave it at that.
Okay. I know over the last couple of years, the company continues to get more excited about how you engage with partners in these cycles. And I think it was when I was on site there last month, you all had talked about some excitement around your large European partner, specifically in Europe, kind of momentum getting behind all of you. But I guess maybe help us understand how important these are to your business today. Maybe how does that evolve? And are your partners really equipped to help with this AI journey that you're on? Because part of it's new for them as well.
Yeah. That's a great question, Scott. So yes, partners are absolutely critical to our go-to-market strategy, especially overseas and internationally. And Europe is a big part of our go-to-market motion. We've signed on some major strategic channel partners like companies like BT, for example. And they're leaning in. They're excited about this market for many reasons. And AI is actually a really great opportunity for all of our partners too. They want to be the trusted AI experts for their customers as well. So our joint customers look to us as their platform vendor, if you will, and we're an AI expert. But those channel partners also want to become AI experts. So they see this as an opportunity. They want to partner with the company, the platform that has the best AI. And Five9 is well known for that. So it's a win-win.
AI is not just a catalyst for moving large brands to the cloud, but it's also a catalyst for some of these partnerships that we're entering into and continuing to sign up in large numbers.
I guess lastly, on some of the go-to-market stuff here at the moment, you also noted that you anticipate FedRAMP approval at some point relatively in the near future here. But can you talk about some of the incremental benefits that you've already received from starting this process? And what type of maybe TAM opportunity does the federal government unlock for you all?
Yeah. This is a big initiative. It's a very expensive initiative. And again, it's a journey. We estimate the TAM to be about $2.5 billion in incremental opportunity in the federal space. And again, it also opens up some of these large commercial brands also do business with the federal government, obviously. And they also, in some cases, require FedRAMP approval. So the TAM expansion is actually much larger than that $2.5 billion direct TAM expansion. But it's a big opportunity for us, and that's why we're investing in it.
We go back to the third quarter results after a tough second quarter bookings environment, right? You all had talked about it. You certainly discussed it here briefly earlier. You had mentioned business commentary was better in Q3, that you're pleased with some of the traction there. I guess kind of a two-part question there is, one, how do we get comfort with that trajectory continuing to improve further from here? And then secondly, what really improved? Was it really just comfort in some of these AI technologies and understanding what they can do? Or is there maybe something else in the end market that you're seeing customers need?
Yeah. I think I mentioned one thing earlier, Scott. I think that AI fog is lifting. And we felt a little bit of that from the outside, from the market's perspective in Q3. But we also had better execution in sales. And as you know, we made some changes in our sales organization, brought in some new leadership. And we've been focused even during the third quarter when we delivered better results. And I'll get into that in a second. But we focused on really realigning some of our segmentation to maximize coverage across the entire market, not just whale hunting, but dolphin hunting, as we call it. We're investing in enabling our sales folks, our sellers, as well as our sales consultants to become the best AI experts in the market, for example. That's an ongoing effort.
But in the end of the day, I was really pleased with the increased execution that we saw in the third quarter by our sales team. Bookings were upping incrementally from Q2 to Q3 ahead of our internal forecast, in spite of a couple of deals slipping into Q4. And what I really love to see is both of those deals got closed in the first half of October. So again, it's great to see the momentum in our bookings engine. And I'm just really, really encouraged by the energy, the culture that we have in that group. We're fired up to be going into our sales kickoff here in early February. It's going to be exciting.
I've got probably two or three financial questions here for Barry. For anyone else that would like to ask questions, feel free to input them into the questions window on the webinar page or email them directly to me. I've already received a couple, but happy to take a couple more as time permits. Barry, you had kind of an interesting NRR number in the third quarter, right? It was the same as it was in Q2. It was unchanged for the first time, quarter over quarter, after seeing some deceleration for a period of time before that. I guess what sort of factors do you need to see to actually have this kind of bottom out or start to inflect upward? And maybe it has bottomed out, but one quarter is not a trend for me. I'd like to see it for at least a couple, of course, right?
Absolutely. So thanks, Scott. The biggest factor by far is macro. We've had excellent logo retention. When transactions pick up, as they inevitably will at some point, we'll see that benefit very promptly. On the macro, the key thing we look at, as we always have, is the debit and credit card spending, JPMorgan Chase, Bank of America, whatever. And it's showing pretty much what we expected, which is that it's bumping along the bottom, even slightly negative in some cases in real terms. Looking longer ahead in terms of the dollar-based retention rate, we are pretty confident, as confident as you can be about these sorts of things, that the rate is going to go up. And the drivers, first of all, and this is key, is that our million-dollar-plus customers, which is 56% of our Q3 revenue, grew 29%. So it's growing in the third quarter.
So it's growing meaningfully faster than the overall business. And they have dollar-based retention rates that are significantly above the 108 basis points that we reported in Q3. So that is a tailwind. In addition, the AI is a new vector of growth. And additional sales that they will also help, internal sales will also help on that score as well. So we feel that we've weathered the worst of the storm in terms of the macro. And we've got these two tailwinds to look forward to.
I know in the third quarter as well, the company saw some really nice expansion in margins, 130 basis points, improving quarter over quarter in gross margins, and I believe it was 190 points on the operating margin side. Gross margins have been a focus of yours the last quarter too, talking about how you're going to get some improvement there. I guess what are your expectations for gross margins over the next, I don't know, one to two years? You're not guiding the fiscal 2025, obviously, at this point, but how should we expect them to maybe trend?
Yeah. I think we were very encouraged by the strong performance in terms of gross margin improvement, 130 basis points year over year, 120 basis points sequentially. The drivers behind that were primarily, in the first place, the stronger revenue. We always have higher gross margins when the revenue goes up. And secondarily, we also had the benefit of that RIF, that reduction in force that happened during the course of the third quarter, so better cost discipline. And then in addition to that, we also have a continual shift every year away from usage, which is this last quarter, 13% of total revenue, which has gross margins in the 50s, up into the subscription part of the business, which has got margins in the 70s.
So those tailwinds and intense focus on improving gross margins and implied in our 2025 guidance that we already talked about at a high level, not formal guidance, we expecting margins improvement there as well. So pretty encouraging. You also mentioned the operating expenses. There we had quarter over quarter, 190 basis points improvement. We should expect to see further improvements over time in the sales and marketing area, in particular, not so much in the R&D side of the equation. But certainly, overall, operating expenses as a percent of revenue will look better.
All right. With that, we will take a few questions here from the audience. The first one kind of touches on the component with Salesforce that you mentioned before. But the question is, can you talk about some of your excitement around your partnerships with Microsoft, CRM, and ServiceNow as eventual growth levers? You already talked about Salesforce. But maybe touch on your partnerships with the other two because you've certainly spoken about them positively.
Yeah. We continue to partner very closely with ServiceNow. Again, we've built enhanced integration with our solutions to leverage. Again, this gets back into what I said earlier, AI from both of us, right? Everyone is going to provide certain components when it comes to AI. And we're partnering very closely with ServiceNow to do very similar things to what we do with Salesforce, quite frankly. And Microsoft is also a unique partner. We have a lot of Dynamics customers, Dynamics CRM customers, if you will. But we also have delivered game-changing integration with Microsoft Teams where it's bidirectional presence.
So essentially, this allows a human agent or even eventually an AI agent, but I want to get ahead of ourselves, to have visibility into the back office to the Teams users to see what knowledge workers might be available to escalate a situation to or to ask a question to. So this is really important. And it's bidirectional presence. We're arguably ahead of many others to deliver this. And again, Microsoft Teams has become the UC solution of choice in many large brands.
Next question is on some of the AI, I guess, success that you're having is, how often do you think your AI products win versus other ones in sales cycles? It's probably early there. I don't know if you have any data to date in terms of what you've seen out there. But I guess the question's probably hitting at, if you had 10 deals today, how many of them do you think you're winning versus maybe some of these point solutions or other vendors?
Yeah. You know, I don't have data to share with you, Scott, but I can tell you that it's very rare that a point solution is competing with us. Oftentimes, we'll be in an account. They want it all from their CCaaS provider. They want the AI that we provide. It's best in class. But in situations, there are some corner cases where they do either already have a deployment before we arrive on the scene, a deployment of other point solutions for AI. But oftentimes, we'll run into a point solution in the account during the sales cycle. And guess what? We partner and provide the best combination of solution, sometimes even on our paper, even though it's a corner case. But again, the key is that they're integrated to our platform.
They can't really do their job effectively, the point solutions, without being integrated to our platform.
I kind of like the next question here is, the question is on usage mix. Barry, you kind of talked about what the mix was here so far in the third quarter. But does usage mix ever trend to zero? Or should it just kind of slowly bleed lower as you have more successes out in the market?
Yeah. That mix shift has been persistent over the last 10 years. And we don't see it changing much as we continue to move further and further up market.
I guess probably the last question here is, within your voice conversations today, what percentage of those conversations do you think the agents or your agentic platform can have success automating today from what you've seen?
Yeah. You know, what most of our customers are telling us, Scott, is their goal is to have AI self-serve in the voice world, especially, a fraction of those interactions that's somewhere in the 5%-10% range, as I said. Now, can we do more? I think over time, as we deliver better and better AI as an industry, I think that could go a little higher. But most of our customers are thrilled, absolutely thrilled, if they can get 3%-4% of their total voice interactions deflected away from human agents and handled in self-service. So I think there's some extreme views out there. That's not reality in terms of what our customers are trying to achieve. Again, it's not a constraint on our technology. We can do as good a job at this as anybody.
It really comes down to a balance between, look, consumers don't always want to use automation. There are certain use cases where self-service via VoiceBot or Chatbot is a good solution. There are other times when they want to talk to a human being, a human agent. The most important thing is they have an escalation path as well. With an end-to-end platform like ours, again, you can start an AI, try to achieve containment of that interaction. If you want a human, and we all have been consumers, we all realize this, sometimes you just got to talk to someone. Having that escalation path, that channel hop capability within a single platform is so important. That's why so many customers choose an end-to-end platform like Five9. Can I just add one other comment on that usage mix?
Again, just to clarify, subscription revenue, 20% growth, 80% of the revenue. Usage revenue is long-distance usage. It's 13% of our revenue mix, and again, as you said, becoming less and less of the mix because we're not attaching long-distance to a lot of our large mega wins and enterprise wins. They're purchasing their long-distance from a carrier directly instead of through a resale arrangement for us. That's actually a very good thing for our gross margins. That is a lower gross margin business, the long-distance usage resale, and so this is actually good news that it's not a growth vector for us. So I'll leave it there.
Fair enough. I think that's a great place to end. Mike and Barry, I certainly wanted to thank you for your time today and participating in the conference. I think you got a couple more here left today in terms of small group meetings. But thanks again. Good luck. And we will chat more on the fourth quarter call.
Thank you, Scott. Thanks, everyone, for joining.
Thanks.