Yeah, exactly. Hopefully my voice can hold up. All right, I guess we are ready to start. Thanks, everyone, for joining us today. My name is Scott Berg. I lead our enterprise software and SaaS research efforts here at Needham. Today we have one of my longest covering companies with us. We have Five9. For those that don't know Five9 very well, we have Chairman and CEO, but outgoing CEO, Mike Burkland. We have, I believe it's the President, right?
President.
Andy Dignan. We actually haven't met that much in person, so.
Not in person, yeah.
It's a little bit different, so this is great. And then still kind of relatively new CFO, Bryan Lee. But Bryan has been with the company for as long as I've known it, so I think to most investors, Bryan's really well known. But all right, I guess brief overview. Who wants to give the, this is who Five9 is for the few people that might not be familiar with the company?
I'm happy to do that, Scott. We've been in the software for CX market in the cloud, went public back in 2014. I started actually as CEO in 2008 when we were $10 million in revenue. And as you know, we're about $1.1-$1.2 billion in revenue today. It's been quite a journey. We've been helping some of the largest brands in the world redefine their customer experience. And for years and years, we've been replacing on-premise contact center equipment and solutions like Avaya, Cisco, and Genesys with our cloud contact center solution. And then over the last, I'd say, three plus years, we've been really becoming an AI company on top of that.
So we think about our business really as providing software for large brands to manage their customer interactions, whether those are traditional voice and digital interactions that go in and out of a contact center, or those that are powered by AI in terms of AI agents. We've got our own AI solutions and a portfolio of products in AI. And AI has now become a meaningful part of our business and is growing very, very rapidly. But our core business is also growing. And that's what I think some folks are probably not clear on. But our traditional CCaaS business is growing, and our AI business is growing, but much, much faster.
We will get into all that.
Yes, we will.
Fantastic. Last time I checked, $10 million-$1.1 billion. That's pretty good growth over a few years, so.
Look, it's a team effort, Scott. Team effort.
All right, so Mike, we're going to start with you. You're stepping down as CEO. You're staying on as Executive Chair, so you're not going anywhere, which means I could still bug you for a little while.
Yes, you can.
This is your last investor event. The company, as I just mentioned, had an amazing run under your tutelage. For me, it's been personally fun to work with you and the team the last 10 plus years. It's been awesome. You guys have been great from an access perspective and just from being very open to what the business is. So thank you for all that. But as you look back now, as you permanently step away from an operating role, I guess what are you most proud of over the, we'll call it about 18 years plus or minus? And what is one thing that you would change if you could?
Yeah, thanks, Scott. And it's been a privilege, my privilege to work with you as well. Look, the growth in our business is obviously, we've already talked about it. That's number one in terms of what I'm most proud of. But look, there's a couple other things too. What we're doing for customers, we're helping them, we say, bring joy to CX. Customer experience, in a lot of large brands' contact center environments, is not a joyful experience. It's a horrific experience. We're all consumers, right? We've all interacted with large brands and their contact centers, and it's not always joyful. We've helped a lot of large brands deliver a true joyful experience when it comes to customer experience.
It's not just in the form of better customer satisfaction and CSAT scores and NPS scores, but it's also tangible business results for them, whether it's revenue growth or whether it's efficiency gains. We've been helping those large brands deliver real business outcomes. Then I'd say third is what I'm really proud of is just the opportunity we've given to the, I call the Five9ers, the close to 3,000 folks that work at Five9, giving them career opportunities and career paths. We have so many people that literally started with us right out of college or even during college as interns and are now part of our leadership team. It's been an exciting journey. That's also very rewarding.
One thing I would change. It's a little bit off topic, but it's something that when I reflect on the last 18 years, I would just, the one thing that I would change is I would have paid more attention to my health in those post-IPO years. And we got a lot of folks in this room that work their tails off every day. Just don't forget about your health. It's really important. Get tested. And when you get a test result, don't take it, don't be cavalier. We're all vulnerable to health issues, and that's the one thing that I would change if I had something to change.
Understood. Very much so. With the change in the CEO role, the company announced Amit. I always say his last name wrong.
Let try this.
Mathradas. I've known him for eight or nine years now because I covered Avalara. He came to Avalara through an acquisition there. And he was President, I believe, at the time there for a long time. But looking forward to working with him again, obviously. And I have chatted with him a couple of times at Nintex since he's been there because I've known Nintex for a long time. But what is his background, particularly around AI and automation, that's interesting to you? And how does that shape maybe the product strategy going forward?
Yeah, great question, Scott. Look, we've been, the board and I have been in this mode of the CEO search since July when I announced my retirement. It's close to a six-month process. We're just absolutely thrilled to have found Amit and have agreed to make him our next CEO. He brings a combination, it's a rare combination of strategy and execution at the highest level. Again, I do think you tend to see CEOs that are really good at strategy or really good at execution, but I think he's one of those rare folks that has both. He has a track record, a proven track record in AI at Nintex, as you know, a proven track record of growth at scale at Avalara. I think he's a strategic thinker.
And I do believe that the winners in our market. This is a very rapidly evolving market with AI coming onto the scene. I've been here 18 years. In the last three years, there have been more changes than there were in the first 15 for sure. And so I do think the winners in this market are going to be those that are able to be nimble enough and agile enough to form the right strategies. He's got a proven track record of making the right strategic moves in both of those companies as a senior leader. And I'd say lastly, culturally, he's such a great fit for Five9. We have a group of people at Five9 that are. We take a lot of pride in being honest, caring about our customers, our employees, our partners, our investors. We're genuine, and we do it the right way.
Amit's one of those types of people. He's just a great human being. I've spent a lot of time with him over the last few weeks. He starts February 2nd, and we're thrilled.
Good. All right, let's talk about the business a little bit. When you think about the long-term growth drivers of the business, Mike, how do you think about the contribution from the on-prem to cloud migrations versus AI growth? You talked about before you have a strong growing kind of core CCaaS business still, but how do we think about maybe the growth drivers between all that?
Yeah, I think the best example of this is we talked about the Gartner data last earnings call, the Q3 earnings call. Gartner published a report that really broke this market, the CX market, into a couple of different components. And they're projecting traditional CCaaS to grow at a 9% CAGR over the next several years. And that's because there's still 60% of the human agents in contact centers running on-premise. And so there's a huge cloud migration that still has to occur. And that's a durable 9% TAM growth opportunity, if you will, that we're in. The good news is we've got a second leg of growth here, which is this AI-driven CX. And Gartner's projecting that to grow at a 34% CAGR. As you know, we disclosed our, in the most recent quarter, third quarter results, our AI revenue grew 41%.
So I think we have an opportunity to win more than our fair share in both the traditional CCaaS market as well as the AI opportunity. And these worlds are coming together. They're not separate. And we'll talk about this in a little bit. One of our biggest competitive differentiators and moats is the fact that we've got an end-to-end platform that has AI embedded in it as opposed to being a point solution for AI or just a CCaaS solution.
Okay. Andy, we were just talking about the on-prem to cloud conversions just a little bit and how it remains a sizable opportunity. I guess to what extent is AI now acting as a catalyst that may start to push those customers to make that transition or maybe even accelerate what that process is? Is that something that you all are starting to see?
Yeah, I mean, if you look at the contact center TAM, right, we talked about it being $24 billion. It's about 40% penetrated. So that's our core business, still 60% of that left to go after. About a year and a half ago, so like mid-2024, we used the term AI fog. We were seeing customers starting to think about, all right, how do I deploy AI? Do I not migrate to the cloud? What's that look like? We've kind of come through to the other side of that, right? And we're seeing a lot of success. Of that 60%, the largest is the large enterprise customers, right? These are the ones that are sort of very complex. We talk about it being almost like a heart surgery or a heart transplant to move from on-prem to cloud.
And we've done a lot of things to de-risk that from a services perspective and tooling and things like that. But we are seeing some customers want to start with what you could call AI first, where they say, look, let's deploy AI first. Let's keep our existing Cisco Genesys on-prem solution. That's kind of in the sales process. As they go through, it still makes a lot more sense for them to move to the cloud at the same time to be able to deliver on the full capabilities of AI. And so I think it's a combination of customers not being ready. They have to move off these on-prem solutions. Some of them have been end-of-lifed, right, by some competitors of ours. And then you also have the Cisco's and Avaya's of the world.
They're not investing in these on-prem solutions to be able to get to that point. So AI is definitely going to be a catalyst, I believe, for that last 60%.
So I guess we've got a product end-of-life. There's some new technology. Sounds like there's a lot of reasons to make the move. What's the hesitation then for some customers? Or is it just being, I don't know, scared about the change in business process or something else? I guess what's that hesitancy?
Yeah, I mean, if you look at all the tech enterprise technologies out there, contact center is kind of one of the last frontiers of being all the way moved to the cloud. And I think it's because this is how brands connect with their end customers, right? And so I think it's less about the technology, and it's more about the change management to be able to make these moves. And this is why I think you see sometimes it takes time for customers, these very large brands, to move from on-prem to cloud is because ultimately they do kind of business unit by business unit. But I think back to the previous question, I think AI does allow that for them to get more efficiencies and deliver a better customer experience, deliver that joyful customer experience.
Having both AI and your contact center in the cloud is the fastest way to deliver that.
Okay. And Andy, you just talked about kind of this large deal environment that's out there. And you all have won some of these large deals, Wells Fargo deal, largest deal in the space. You've talked about CVS and FedEx amongst other customers, UnitedHealthcare versus UnitedHealth Group. I live there, and they've got like six different names. I always forget which one it is. Anyways, but can you touch on maybe why you're winning a couple of these deals in particular? What's made Five9 the platform of choice for some of these large deployments that historically maybe you weren't involved with?
Yeah, I mean, I think if you look at the platform, I mean, this is the, if you look at a contact center, brands just cannot, they don't allow downtime, right? You need to have a very scalable, reliable platform, and that takes demonstrated over a long period of time. And so I think that's one key part of the reason. The other part is ultimately these are complex solutions. And so having a services team, having a sales team, having really a company that cares about the success and the durability of making these large migrations, like I told this story a couple of times today, one of the largest brands that we won, we get to the end of the sales cycle and I asked them, well, hey, you had us or were the other competitor, why did you choose us?
And you said, we had the most belief that you would deliver on this three- to four-year journey of getting us moved off of the platform. And so, and again, obviously you have to have a great product to be able to do that and be scalable and reliable. But all those other intangibles I think are what's allowing us to continue to win these big deals.
Okay. Now, Mike, you were kind of generically hinting at some of the AI differentiation out there. And the debate whether AI is a threat to contact center or the CCaaS space in general has been well documented. I think most here are certainly more than familiar with it. But as AI adoption matures and is embedded across the platform, where does AI create durable differentiation for Five9 versus becoming maybe table stakes across the market?
Yeah, I mean, we've talked about this a little bit. There's such an advantage of having this end-to-end platform to orchestrate interactions, consumer-to-brand interactions, whether they're voice or digital, whether they're handled by AI or by a human. To have a single platform to orchestrate those interactions is so critical. Point solutions, for example, we'll get into point solutions in a second, but that is such a powerful combination to have that end-to-end platform. We also embed AI throughout a lot of the components of our platform. For example, AI routing is now part of the core CCaaS offering, even though we don't consider that AI as our revenue mix. But again, I think you're going to see these worlds kind of come together where AI just enables even our core platform to do a better job. AQM is another one of those areas, Agentic Quality Monitoring and Management.
That's a whole nother opportunity we just announced. So again, I think as we continue to innovate, expand our AI portfolio beyond just the traditional use cases that people know about, which is self-service and agent assist and deliver some of these other AI-powered feature functions. It just allows us to do an even better job of delivering that customer experience that these large brands are looking for.
Okay. So staying on the AI front a little bit, you all were early in this kind of, we'll call it generic agentic theme with your IVA technology, right? But how has the IVA technology evolved into agentic agents that today are capable of whether it's reasoning and all sorts of different functionality? I guess in practice, which use cases are driving the most tangible customer value today from what was traditionally seen as IVA to now something that's more agentic?
Yeah, there are many use cases that our customers are deriving tangible ROI from, whether it's self-service, right, and our AI agents that can, again, do a lot more than just a simple voice bot. And it results in value propositions like a better customer experience, better CSAT, but it's also a cost savings. And in the long run, a lot of these business cases that our customers are writing to justify the purchase of our platform is surrounding labor arbitrage. And so there is an opportunity here to grow your headcount at a slower pace or hold it flat for a period of time and have more and more interactions handled by AI and self-service. So again, agent assist is another big area for us. It's our number two product from a revenue perspective.
A lot of the tangible ROI that comes from that is it's efficiency gains, it's productivity gains, and it's also CSAT because those agents, the human agents that are leveraging Agent Assist can actually do a better job of solving problems for customers.
Okay. Now, Mike, I know you've had this question probably a thousand times, but the contact center market is very dynamic, as we've already talked about a little bit. But why does a platform like Five9 win against these third-party agents in particular, these agentic point solutions? And is there a particular scenario that you feel like you're winning best in those? I saw a private competitor of yours today talk about, I think it was $330 million worth of ARR. I get investors that quote stuff like that or others that are maybe at $100 million or $200 million, which is a little bit more than where you're at today. But why are you winning versus them? What's the right value proposition? And why do you think you'll ultimately be better positioned?
It gets back to what I said before, this end-to-end platform. And something I didn't mention that is really critical is data. So we're the system of record for conversational data. If you think about a contact center platform, all the interactions with consumers are going through our platform, whether they're digital or whether they're voice. That conversational data is absolutely critical for AI to do its job. And if we're talking about an AI point solution instead of Five9's AI agent, for example, to deliver self-service, if it's a point solution that's not connected to our platform and they don't have that conversational data, Scott, you're a consumer, you're interacting with a brand.
If you're going through a voice bot, a simple point solution voice bot, as clever and as sexy as the UI might be or the interface or the voice might be, if it doesn't have access to the conversational data that you had in that contact center a week ago, for example, with a human agent or yesterday via a chatbot, if it doesn't have that data, it cannot deliver a personalized experience to you. It's going to start from scratch. It's not going to have any history of where you've been, what your problems were, what the resolutions were or were not, and so that conversational data is a huge competitive advantage for our platform. That's why a lot of these point solutions, these AI point solutions, are integrating to Five9 and wanting to integrate to us because they need our data.
And in the end of the day, we've said this before. AI will be a team sport. There are going to be use cases where a customer might use a point solution for AI, but only if it's integrated to our platform to get access to that conversational data. And we monetize that through VoiceStream and TranscriptStream. So we're kind of, we're a winner when it's our AI, but we're also a winner even when it's a point solution, for example.
Okay. I guess when I think about to that exact point about how you still monetize that through whether it's the transcripts or the voice stream, is if I put myself in that customer's shoes, when I think about, I don't know, total cost of ownership and the cost-benefit analysis, if I still have to pay you for some of that access on top of paying for this third-party agent, why buy the third-party agent in that scenario? Because if I start adding it up, it's likely more expensive than just buying the full solution from yourself.
Yeah. Look, it is a very valid point, Scott, and it's also driven by value, right? And just like I said, without that access to our data, that customer experience isn't going to be accurate and it's not going to be personalized. And self-service without those two things is a fail. It's going to fail. And so in the end of the day, customers, the brands are going to pay that uplift to get to our data. And whether they see that as a reason to go with our AI instead of a point solution, that's just one more advantage that we have.
Okay. Mike, you've been clear that AI functionality at Five9 is primarily priced on a consumption or capacity basis today. But you've also been discussing an increasing use of kind of revenue commit-based contracts as customers start to balance their agentic use versus the core contact center use. As that model matures, how should investors think about the long-term evolution of pricing? The minimum commitments today, it's kind of interesting. It does change the viewpoint there, but I also think most of the examples that you all give on how customers adopt your agentic framework usually says the customers still pay you more than what that minimum is going to be.
They do, and so the way I see this playing out in the future in terms of pricing models is, look, today our AI solutions are consumption-based or capacity-based pricing. A lot of our historical product set has been seat-based, but we're starting to write contracts even with existing customers on renewal that are derived based on an interaction, basically a number of interactions, whether they're handled by a human or whether they're handled by an AI agent on the back end. We're in the business of providing software to help these brands process those interactions and manage those interactions effectively and successfully.
So what we're doing is we're actually entering into some contracts with customers where we'll derive some level of minimum revenue based on some assumed mix of human seats and AI agents, but we'll give them some flexibility to essentially make it fungible for them between those two types of software solutions. But as long as they pay us that minimum revenue, they're good. And so it's a way for us to think of us as generating revenue based on level of interactions. And this gets us out of that debate around humans versus AI. At the end of the day, consumers are going to reach out to brands more and more if they can reach them more easily, whether it's with AI or with the humans. So as interactions go up, which by the way, they always have and they probably always will, we'll continue to benefit from that.
My two comments on that are, one, you're essentially creating a model that incentivizes them to meet their business needs as they need, but it also incentivizes them to actually kind of try some of the new technology a little bit, which I think is interesting to remove some of the friction. But the stat I always like in this contact center environment is voice calls; the number of voice calls has gone up since the early 2000s. Even when all this digital customer engagement has increased, when people were calling for the downfall of voice calls back in 2000 or 2001, that scenario likely doesn't change anytime soon, even with all the technology.
It's a perfect data point, Scott. Look, our industry has seen so many opportunities for automation over the years, whether it's websites or whether it's IVRs back in the day or whether it's AI agents today, and at the end of the day, consumers will reach out to brands more if they're given more opportunity to do so in the right way and effectively in an effective way, so again, I anticipate interaction volumes to continue to go up over time, and if that happens the way it's always happened in the past and we have a pricing model that reflects interaction volume regardless of the mix, we'll end up in very good shape.
Excellent. I have two more questions myself, then happy to take any questions from the audience with the time that we have remaining. Bryan, let's talk about NRR a little bit. This quarter's trailing 12-month NRR dipped slightly to 107% from 108% in Q2. One point's really not that big of a deal, but the metric has been in decline over the last couple of years. As you look ahead, what are the most important levers that maybe need to shift for NRR to trough or maybe inflect upwards here? Because Mike talked about growth rates for the core contact center environments and some of the agentic growth rates. That should probably mean, at least my math would say that that number probably has to go up if you're going to at least maintain share a little and take share.
Yeah, absolutely, Scott. So as you said, in Q2, LTM DBRR was at 108% and then 107% in Q3, and there are really two key tough compares that we've talked about throughout the entire year, and that's one of our largest customers that was finishing a multi-year ramp throughout 2024, contributing significantly to revenue that year, and also seasonality because in 2024, we had very strong seasonality versus minimal seasonality as anticipated in Q3. So going into Q4, those dynamics still remain, so there will be more headwind than not in terms of DBRR. But going into 2026, we expect upside in DBRR because we'll be lapping those tough compares, and then on top of that, we have a backlog. Typically going into New Year, we have a backlog of new logo bookings that we've already won that we have great visibility into and converting to revenue.
But what's unique for this year is that we have install-based bookings that are also in the backlog that will be helping convert into revenue. And these are very different from the traditional install-based bookings that we used to have where we have a fully deployed customer with our platform that's simply adding seats and turning out right away into revenue. These are upsell, cross-sell software, including AI that requires deployment as well as new business units that we're winning with our existing customers that also require deployment. So it's very similar to new logo bookings in that sense. So we've been having more and more of these types of bookings that are starting to convert into revenue in Q4 2025, but more so in 2026. And that'll be a nice uplift to our expectation around DBRR having upside in the year.
Okay. And given the land-and-expand model for the company, it sounds like some of those expansion deals are more in go-live than contract signing. Is that correct?
Yes, the revenue recognition will ramp actually as the products are turned up.
Okay. Okay. Something to note, a little different than most of the software companies I cover. So last question for me is, Bryan, you've laid out some clear profitability trajectory and goals guiding to a midterm adjusted EBITDA margin of 25%-30% for calendar 2027. That's already next year. I can't believe it. I guess what portion of that margin step up so far is just structural from natural operating leverage of adding just revenue and high margin revenue every year? And what's more coming from discrete cost actions? How do we kind of think about those maybe two levers going into next year?
Absolutely. So we did a full operational review in the early part of 2025. And at that point, we came up with many different cost savings initiatives. They include things like increasing the mix of offshore hiring, delayering, spend control expansion, automation to increase efficiency, and renegotiating third-party spend. I mean, the list goes on, right? It's a broad list of cost savings levers that we continue to execute on even through 2026 and beyond as well. So that will be a nice driver of operating leverage. But in addition, if you look at our AI portfolio, that's the fastest growing category of our revenue. And the biggest portion of our AI suite is AI agents. And AI agents, if you break down the gross margin, it's actually in the high 70s%, low 80s%.
So naturally, as that becomes a bigger mix of our overall revenue portfolio, that's actually going to give us leverage on the margin side as well. And we do believe that from all of this conversion to cash flow ultimately as well. And we've been very focused on different areas to improve the conversion from EBITDA to free cash flow where, for instance, CapEx historically has been 6% of revenue, whereas in 2025, we've guided to under 3% of revenue. We expect that trend to continue, as well as areas like working capital where traditionally our business has been much more monthly invoicing based. But as we move upmarket, a lot of our larger customers are on annual invoicing, annual prepayment. And we're actually standardizing that across a larger base of our customers in the enterprise.
So we should get more efficiency in terms of working capital to help improve the conversion rate there as well.
So, on the invoicing component though, because you might have discussed that before, but it's the first time I recall it at least, is your business has a. It's not a transactional component, but it's a usage component to it, whether it's as you move up market, less on the telephony reselling side, but more about just the number of seats and agents. You talked about seasonality before. How do you kind of normalize or true up for those contracts if usage moves up more than expected?
Yeah. So there's going to be, if you look at a perfect example is AI, right? Because all of our AI portfolio is based on either capacity or consumption based, but it's actually a block of minimum committed units of whatever the unit is and then charging for any kind of overage on top of that. So that way, in terms of when you're getting the annual invoice, you already know how much of the minimum commitment that they have and then additional invoicing if there's overage there.
Will you do all that invoicing just at the end of whatever their contract term is at the end of 12 months?
So it's true up at the end. Initially, there's an annual invoicing upfront and then true up at the end.
Okay. Cool. All right. With that, we do have a couple of minutes for some additional questions. Got a hand in the back.
Yeah, thanks. I was curious just about how you guys are going about developing the models themselves, the AI agentic models. Do you guys have your own pre-training stuff, training stuff, or are you guys also working with other parties to try to adapt and then maybe enhance them as you deploy them to the customer base, and out of curiosity, if you do do that, which models are those that you're working? Is it Claude or something from Amazon or probably Gemini is doing a lot of stuff, Google's? Anyway, just curious if you want to define for them.
Yeah. Our strategy has been to be engine agnostic, as we call it, right? So we're, and we've taken that strategy from day one when it comes to AI. We plug and play the latest and greatest models and engines, if you will. And it's usually on a customer-by-customer basis and a use-case-by-use-case basis. And we work with the customer. In fact, Andy and his team are front and center on this, so I'll let you kind of give the more detailed answer. But look, there are some use cases when Deepgram makes more sense. There are other use cases where another engine makes more sense and is more accurate. And that's our job is to make sure that we're deploying technology that is going to help the brand achieve the outcomes that they're trying to deliver to their consumers.
And it's about accuracy and personalization and lack of hallucination and inaccuracy, right? So it's quite a process, but it's also something that we've strategically aligned on from the very beginning and worked with every single one of our customers to fine-tune that. And then we actually use that customer's data to kind of uplevel those models even greater so that the fact is that they're knowledgeable around that brand's solutions and FAQs and everything else.
Yeah. I guess the only thing I would add on top of that is we're not building our own large language models, right? What we're doing is we do also allow bring your own model, which we see a lot of higher-end customers or customers in healthcare, financial services, as we can plug that in and have our software sit on top. What we do though is we build a lot of very industry-specific use cases where we have the prompts built on top of whatever engine that we're using. And that way we can take those same use cases and bring them to the next customer, right? So that's kind of how we think about it.
Thank you.
Thank you.
Quick question on that as well. So you mentioned that you bring each other model and you pick the best in class. So how is the data kind of like the data model that you mentioned that you have, how does it play in that case?
Yeah. Remember when we're talking about training an LLM, it's about having that LLM be able to communicate accurately and effectively with the end consumer. But when you think about there's another piece of, I call it contextual data. And if you're the consumer on the other end of that interaction, that AI, whether it's running on Deepgram or OpenAI or Gemini, it has to have access to the conversational data that resides in our platform to know that you were talking with Andy in customer support last week and you were calling in about problem XYZ. And here's where you left off, right? So there's a difference between the conversational data and how the model's trained. They're actually very different.
Can you explain why technically the existing customer bookings and cross-sell are taking longer to roll out or what?
Yeah, I'll take that one real quick, and I know we're past time. But look, when we say taking longer, what we're really, we don't mean that. What we're saying is the type of booking is very different. We used to just add a seat and flip a switch. That's no implementation. It's instantaneously revenue. What we're doing a lot of today in our install base is upselling AI products and new divisions within a brand. And those have a deployment cycle. So they're longer than the old seat turnups, but they're not longer than these types of bookings were, say, a year ago or six months ago. We're actually getting better and better at these AI implementations. And they're actually, over time, they'll actually take less time.
So it was really a comparison between the old install-based bookings that Five9 used to talk about because those would instantly turn to revenue. Hope that helps.
With that, we are at time. I want to thank everyone for joining us. Mike, once again, it's been a blast. Andy.
Thanks, Scott.
Bryan, thanks so much for the time.
Yeah, thanks, everybody.
Thank you.