Five9, Inc. (FIVN) Earnings Call Transcript & Summary

Hi. Thank you, everybody. With us for the first company keynote of the day, we have Rowan Trollope, CEO of Five9; and Barry Zwarenstein, CFO. And before we dive into the Q&A, I just wanted to turn it over to Barry. He had a few words to say, and then we'll take it from there.

I just wanted to remind everyone that our discussion today may include forward-looking statements concerning events and trends that may affect our industry or the company and its operations. Actual results may be materially different from what we discuss here, including the risk related to the ongoing COVID-19 pandemic. Please refer to the 10-K, 10-Q and the risk factors for detailed information that could cause the results to differ materially. Thanks, Samad.

Great. Well, again, let me say thank you in advance for joining us. The company is coming off of a phenomenal quarter that you just reported. And maybe we'll kick off with the fact that contact center volumes were spiking in 2020. There is a need to rush to the cloud. Rowan, maybe kicking off with, now we're starting to see the world reopen day by day, what are you seeing in terms of engagement and the industry? Is it normalizing? Is it still elevated demand? Let's maybe kick off there.

Sure. Yes, thanks, Samad, for having us on this morning. We were already seeing an acceleration in the move to the cloud pre-pandemic. So it wasn't as stark as what you might have seen with some of the other names like Zoom or WebEx, those kinds of applications that people just needed instantly sort of download when they all have to work from home. These were large companies that were already -- in terms of our customers anyways, mostly large companies that were already sort of on that path of moving to the cloud. And we saw definitely an uptick in volumes at the beginning of the pandemic as -- I think the effect that we noticed was essentially that anyone with a retail presence or any kind of in-person physical presence, physicians or retailers and so on, needed to sort of direct all that traffic to their contact centers. So the contact center became sort of the front door for the company. And that wasn't what they had before, and that definitely drove increases. What we've heard from our customers is that they actually are enjoying the fact that they now have -- even as we start to reopen some of the sort of physical establishments, these companies are saying, we're going to continue to leverage this virtual experience. Telemedicine is a great example where folks are now starting to say, look, we're actually going to move to this hybrid mode. Now that we've sort of deployed this technology and it works, and actually, customers really like it, we're going to use that for more and more of the experiences wherever we can. So I think it's sort of like we saw an uptick, and I think we're going to be at a new normal post pandemic with more leverage of e-commerce and the sort of digital engagements with customers.

Great. And then maybe if we step back, you guys, to your point, were seeing very strong enterprise demand already. And it's, I think, a secular trend. Can you maybe speak to what the key technology differentiation is for Five9's B2C platform? I think just as we spend the day digging deeper into contact center in UC, it would be helpful to see what the company's view of its differentiation is.

Sure. Well, there's really 3 things. The first is fairly obvious. In the on-premises world, right, where most customers -- where most businesses still live and breathe, their contact centers are essentially hardware systems that are attached to their home system, which is literally sitting in a closet somewhere. And most of the phone calls are getting routed to telephones sitting on desks. I mean that is the reality of many contact centers. Now some of them have switched to sort of desktop VoIP experiences. But fundamentally, they still have hardware in buildings and they're load balancing between different PBXs and contact centers. The move to the cloud totally changes all of that. You -- all of a sudden, like your agents log into a web browser, your supervisors and your administrators all log into a web browser. The whole thing is hosted by Five9 in a multi-tenant infrastructure that's running on public cloud deployed globally. So you instantly get this incredible boost in reliability and scale and the ability to sort of pay as you go, scale up, scale down. So those are sort of the first set of benefits. They're fairly common, candidly, to most transitions to the cloud for enterprise software. The second part, though, is what happens when you get there? We have a very digital sort of digital-first approach to engagement with your customers. So part of the whole story here is when you get to the cloud, you get all this additional flexibility. You can engage with your customers over WhatsApp or SMS or messaging or e-mail. And if you need to seamlessly escalate from one of those channels into a live engagement with a human being, we make that really, really easy. So that's the second thing. And then the third thing is actually all about automation. So the new trend in the contact center is AI-powered automation. And what we've seen over the last few years is AI has evolved to a point of maturity where we can leverage it to do more and more of what the human agents are doing, which, by the way, in the contact center world, tends to be very sort of high-volume, low-value tasks that we believe are relatively easy to automate. And by doing that, we think we can save contact centers lots and lots of money. So there's a huge ROI story that's emerged around AI-powered automation where Five9 has taken a lead in that market. So those are the 3 things.

That's great. Maybe we'll start with the last one first as far as pulling the thread. I know the company did the Inference acquisition, and that's one where you're leveraging, I think, a lot of automation and virtual agents. So how do you maybe think about that and the impact of solutions like that to spending in the contact center? And how is Five9 positioned to benefit from that?

Sure. So our strategy has been, number one, listen to your customers and ask them what do they need and what do they want. And what you find very quickly in the contact center -- by the way, this is not new. So this is sort of -- this has been there for a while, is look at the share of wallet in the contact center or the wallet spend, the spend in the contact center. What you immediately see is about 10% of the spend is on technology and associated costs for technology, and 90% of the spend is on labor. So as they think about how to improve the service that they deliver to their customers, technology is just 1/10 of the overall picture. And what's emerged for -- frankly, over the last 30 years, what you've seen is a trend to say, how do we optimize the labor spend? How do we say -- how do we make our managed agents more efficient, more educated? How do we help customers self-serve? How do we leverage other ways to sort of make the labor as efficient as possible? And unfortunately, over that period of decades, it's been a challenge. And the technologies have sort of come and gone that have served businesses in the sense of sort of diverting customers away from actually human engagement, but the customer experience has suffered, right? You've all experienced that push 1 for sales, push 2 for this, and you end up with that crazy menu tree, and you just want to hit 0 or let me talk to an agent, get out of this crazy system. That's an example of sort of technology that was well intention to solve a problem but ultimately created a worse customer experience. The modern version of that is chatbots, right? We all -- many companies implemented web chat. You've probably all seen that. And if you've tried it lately or even in the last couple of years, you go in and you start chatting, and what you'll find very quickly is it's sort of the modern equivalent of that IVR. These chatbots tend to be not very well done. They're hard to communicate to. And eventually, you would just want to kind of get to a human being because the technology is not great. But all that has changed. Essentially, what Inference has done -- and then the last sort of 2 years, I think, is when this has become possible. Deep learning has advanced to the place where automatic speech recognition, natural language processing and conversation management, those 3 areas, have really gotten to a place where computers can hold sort of reasonably good conversations with human beings in such a way as to sort of take the same conversation that a human being might be having from an agent and say, hey, we're going to encode that into -- we're going to train a machine learning model to do that. And so you can get answers to your basic questions. You can talk to a chatbot or you can talk to a voice bot essentially and get the answers to your questions or do simple tasks, address changes, check on an order, shipping status and so on. And what Inference has done so well is divide those across industry verticals. So when you come to the Five9 platform, what our customers find is, hey, if you're in retail, here's a set of tasks that we've prebuilt for retailers, the kinds of things that retailers need, RMAs, order status requests, shipping -- checking on shipping and so on. For health care, we have a set of sort of vertical tasks for health care, checking on claims. Insurance, banking and finance. So we have all these industry verticals with these tasks. And one of the ways that we've done that and that we continue to -- or one of the ways that we think we can differentiate, I guess you could say, and maintain a lead in this space is by using the massive amount of data that we have. So we at Five9 have access to 7 billion minutes of recorded conversations with customers every year, and that's both phone calls, literally agents talking to customers, as well as digital transaction, SMS chats, e-mail chats and so on. So all that information can be taken on an industry-by-industry basis. And you can look at what are the most common call drivers, what are the most common customer complaints and issues. And then you can sort of look at those conversations and architect a digital engagement where the computer can sort of take over for the human being. The ROI story for the customer is, look, we can take that 90% of your spend that's spent on labor and shrink it to 80% by automating 10%, let's say, of the tasks that human beings do. Now we then sell them a digital agent or a virtual agent to do that work on behalf of the human being. And that's how that works. And so for us, we make more money by selling more technology, and the contact center saves money by having fewer contact center agents and less labor spend. So it's a very compelling story. Most of our large enterprise deals right now are leading with this technology. And we could be looking at easily -- the $24 billion TAM that is out there today for technology spend, this could be at least double that in terms of TAM expansion, if not more, over the next 5 to 10 years, and that's probably a conservative estimate.

All right. Well, I think that everybody in the audience definitely likes that idea, especially because it bodes well for your long-term growth. But if I maybe ask a follow-up to that, we get asked a lot about what we think will happen in agent headcount as a result of more AI and automation. And I guess, how do you think about -- because you mentioned something earlier that ties into this, which is the contact center is no longer just a cost center, right? It's actually maybe the tip of the spear for engaging with customers or the end customer. So how do you think about maybe headcount inside of the contact center? Or does that not matter so much as your model continues to evolve?

Yes. For us, less headcount replaced by digital technology provided by us is upside to us. But because we monetize a digital seat that replaces a human seat, that digital seat will monetize at somewhere between $400 to $600 per digital agent per month, whereas the human agents, we're monetizing at an ARPU of about $204 a month, $205 a month. So it's more than double the spend. Our sort of mid-tier offer in this category for digital agents is $450. So we now -- we don't have enough experience yet to give you exact ARPUs and so on. We're getting started. Inference had really the best technology in the market in this space and -- but was just getting started in terms of their revenue run rate and customer penetration. We've just seen incredible interest in this technology from our customers. And yes, so I would say, some companies will probably actually go with less agents. It's very compelling to them. It's like -- it's a better -- in some cases, a better experience for their end users or their customers. So I've talked to some customers who are like, look, if I can take my 1,000-agent headcount down to 900, that's a good thing. And they'll take those savings and redeploy them partially to technology that obviously offset the -- to replace the labor, but they may deploy that technology in other ways as well. They may start to deploy those dollars in other ways as well. For some companies, they're just looking to reallocate. They're saying, look, it's hard enough for us to keep up with our growth, for growing companies, and just to find new people given the turnover in the contact center. So this technology kind of gives them a solution that say, okay, well, we have 1,000 headcount. Now we were planning to go to 1,200, but maybe we can go to -- maybe we can hold it at 1,000 for a while by implementing these automation technologies. And so ultimately, it is cost savings to the business. How they choose to use it is going to vary by industry.

Understood. And maybe sticking on the technology side of things. You mentioned messaging and now we're hearing increasingly about omnichannel, right, where consumers can engage. It used to be maybe just phone 15 years ago or desktop web chat. Now I can text you. I can WhatsApp you. I can use Facebook. By the way, I don't use any of these things. I'm a research analyst. I'm basically just the nerd that calls people. But let's say, for the people that do use all of these different channels, how does that -- how do you think about Five9 in terms of omnichannel capabilities given that voice is kind of the original genesis of what the company was solving for?

Yes. We've made a big investment in our digital channels and -- well, we call them digital channels, not omnichannel. And so we've got -- I would say, right, we're up at the front of the pack now in terms of digital channels with -- including SMS, including web chat, including WhatsApp, you kind of name it, e-mail, all the ways you might want to engage with the business, we've got a really deeply integrated solution. What differentiates us here, there's a lot of companies out there that focus on digital channels only. But when it comes to sort of that seamless transfer from the digital channel to the live channel or even a multichannel, so the agent is literally talking -- maybe you start out in chat talking to a bot, okay, talking to an Inference IVA. Okay, you have a good experience, but something -- you're at an edge case and you need to speak to human being. The human being can come on live on the back end of that chat and start talking to you and the agent can talk and resolve your problems. When that human being comes on board, the AI -- the Inference AI that's been engaging with the customer presents all of the information that's already been gathered to the agent. So they don't have to do any of that, right? They don't have to reask you, oh, can you tell me your account number or this or that and other thing. It's all sort of seamless handoff to the human being. And if you want to escalate from there to a voice call, you're one click away. So that's the power of sort of the integrated platform. We see more and more customers going digital-first, and we see more and more customers, in some cases, going digital-only. Now this is not so common in the U.S. because I think our expectations are different. But you go to other countries, South Africa is a great example where telephony and voice tends to be very expensive. We have a bunch of customers using our IVA technology that are exclusively WhatsApp. That's the only way you can get support from them. And so I think it's going to be a mix globally in terms of what we see. I definitely see digital volume spiking and companies trying to keep those labor costs relatively flat, if not shrink them. So we should see the agent count over time, global agent count, which today stands somewhere in the teens million, 16 million is probably the best estimate we have. That probably flattens or goes down over time, but the overall technology spend in the category, as I said, probably has the potential to double or more conservatively.

Great. That's helpful. And earlier this morning, we had a speaker that mentioned the increased focus just generally from the contact center software providers on WFO and bringing it together and that Five9 acquired Virtual Observer. And so I think maybe just help us think about your philosophy. I know you also partner on the WFO side. Just what the approach is at Five9 on the WFO side and maybe how the integration of Virtual Observer is going so far.

Yes. That's gone really exceptionally well. It's been beating our internal expectations in terms of attach rate and adoption and also sort of customer success metrics like NPS. So the net -- the Virtual Observer product was a great acquisition for us. We've integrated it into the interface with Five9, and we're seeing a lot of customer success there. Now we tend -- so I guess the strategy here is portfolio expansion and capturing more share of wallet -- of the 10% of the wallet spend that's spent on technology, right, we want to get more than our fair share of that. And so we needed to have a complete portfolio, and Virtual Observer was really checking that box. It was a timely acquisition for us because in a work-from-home world, you really -- it becomes a requirement. In the old world, pre-pandemic, the way that -- if you walked into most contact centers in the world, what you would see is supervisors sitting in [ turits ] actually like literally, physically above the agents who sit in cubicles. And the supervisors would walk around and listen to agents and overhear their calls, maybe look at their screens. When all those agents went home, that form of management by walking around just vaporized overnight. You're not going to be doing that anymore. So how are they doing it? Well, Virtual Observer essentially virtualizes that exact experience. It takes that experience of walking around, looking at screens, listening to calls, maybe whispering in an agent's ear, try this or try that, and puts that into a web browser. So you can see all of your agents live. You can click and listen in on a call. You can see their screen in real time. You can pass them feedback through chat. You can whisper to them, all that stuff. And so it's just seen incredible adoption. For the high end of the market, we partner with a great company called Verint, who has sort of the market-leading technology for large, large enterprise who have very complex needs in the WFO space. So it's an area that's evolving pretty quickly. There's a lot of innovation in this category, and Virtual Observer has been going really well for us.

That's great to hear. And just -- I'll remind the audience, feel free to send questions in via the live web chat portal. I actually see a follow-up to an earlier question that I had, Rowan, where investors wondering, when you think about the different channels that Five9 can help engage with, how does that work in the context of maybe your partnerships with Zendesk and Salesforce and some of the other companies that are maybe adjacent to where you guys operate?

Yes. And this is a strategy that's born of years of sort of working in big companies. I came previously from Cisco and before that from Symantec. And we don't live in a world where one size fits all. And I've seen and been part of experiences where customers suffer when big competitors -- big companies fight with each other and compete because what they do is they sacrifice interoperability. They say, look, if you want to get everything from Microsoft, we're not going to interoperate with other things because you can get it all from us. And that's typically the kind of thing that really hurts the customer experience. We have taken a very different approach, which is we're open to partnering with everyone. And if the customer wants to leverage, say, a Salesforce digital channel or a Zendesk digital channel, we will complement that solution, and we built very strong partnerships with both of them, ServiceNow is the third one I would point out, where we want to do the right thing for the customer. And so if the deal or an engagement with a customer is being led by Salesforce or Zendesk or Salesforce, we're going to fit in our solution and fill in the gaps in their product, bringing our voice, bringing our AI and automation and so on. And they really love that. So that's -- the partnerships across the boards have been going really well. ServiceNow has really been on a tear lately, I think doing very, very well with their service desk product, customer service or the service desk product. Our relationship with Zendesk has been very strong. They have a low end sort of talk capability that they've built into their product. But when they need a serious contact center, we're one of their top partners, and similar with Salesforce. So we take a sort of customer-first approach to strategic partnering, where we're willing to sort of cooperate and compete at the same time for the customer's business.

Great. And maybe staying on the theme of partnerships, an exciting one the company has is with Zoom, right? I would say it's always exciting to partner with a household name. And we've heard good things based on our own work. But I'm just curious maybe how we should think about that partnership bringing UC and CC together and maybe what the value proposition is both for Five9 and for your end customers.

Yes. The partnership has been really fantastic. Eric, the CEO and founder of Zoom, and his team are just incredible to work with. And they've done a great job getting traction with their phone products. So when they talk to us about their intention, the launch of phone product, I said, right, I want to be first as your contact center partner, and I want to be the best partner you have. Now they don't have an exclusive relationship with us, but we've worked really hard to make sure that our product interoperates really well and that we have good relationships with their partners and with their sales team. So we actually show up. Usually, in many of the deals that they are doing where a contact center referral is needed, we're at the front of the line there, usually. And so that's been going very well. Last quarter, we announced that over 30 deals in the Enterprise segment came from either Zoom or Microsoft Teams referral. And in fact, most of those deals came from Microsoft Teams. So to pivot over to Teams for a moment, they've got a very large opportunity in front of them with their Teams UC offer. And just like we integrate with Zoom, we integrate with Teams, and so we've seen really strong traction there, especially in large enterprises. The large enterprises are clearly making decisions to go all-in with the Microsoft stack. And in that case, we plug in very seamlessly with Teams UC and also with Dynamics.

Great. That's helpful. I want to maybe shift away from technology and strategic partnerships then and get back to maybe some of Five9's execution success, right? The company saw a very strong fourth quarter. Enterprise revenue, in particular, was robust. Maybe dig a little bit deeper into that successful upmarket push and just give us an update on the progress the company has been making as you continue to get to that 1,000-, 5,000-seat contact center deployment?

Sure. I think I'll start out by saying we've been very deliberate about our growth strategy. And in some cases, we've said no to large deals where we knew that was going to sort of stretch the limits or potentially sort of cause the high-speed wobbles. And so in that approach, what we've done is we very strategically and thoughtfully continued to expand the largest enterprises that we can serve. So it's partially, it's an internal service delivery capability, global coverage. We spent about 3 years rearchitecting our platform to move to the public cloud. That's now done and behind us. So we're now deploying globally on GCP, public cloud servers. So that gives us real flexibility and speed to get out into these markets where local data residency -- and frankly, with voice also, I would point out, you actually -- from a latency perspective, if you're sitting in Germany in a contact center, you can't have your voice hairpinning all the way back to the West Coast of the United States and then back to Germany. So having local presence is particularly important for real-time services as our local telco relationships. So we've been building those out as well. So that was part 1 was sort of running the train tracks in front of the -- Dan, our President and Head of go-to-market in front of his train. Second was about a 3 sort of pillar transformation strategy we put in place in our go-to-market teams, and one of the things that we did last year has been very successful. So we've created a strategic enterprise account team. So we divided up the teams into a -- we now have a commercial, mid-market, enterprise and then strategic enterprise. That's worked really, really well. And frankly, in Q4, what you saw was all 4 of those cylinders kind of hit at the same time, and you saw growth just kind of across the boards from all 4 of those teams. And that's reflected in the commentary that we made about our bookings and pipeline strength. Of course, the revenue that we saw accelerating in Q4 came from what we had been working on over the past year and years. And we've been very consistently going after those $1 million ARR deals. Of course, we announced our largest ever deal, which was a $12 million ARR per year -- $11 million, $12 million, Barry, correct me if I'm wrong. And -- but we also announced that for the full year, in 2019, we -- at the end of 2019, in Q4, we had 59 customers paying us million dollars in ARR per year. And at the close of 2020 in Q4, that number had accelerated to -- Barry, what was the final number?

$91 million-plus ARR yields.

So really, really strong growth in the enterprise segment, and we're just -- we're seeing that the work that we've done to build out the maturity in the company is working. And again, we're just executing and kind of plotting along and continuing to make progress with these larger and larger enterprises.

I want to maybe pull on the thread of that very large deal that you mentioned and given how sizable it is compared to your average ARR per customer. But was that a deal where you guys had to go chase them? Was that an inbound? I guess are you -- and that's maybe in the spirit of a broader question, which is are you seeing more large organizations come to you because they know that now that the [indiscernible] exposed some of the challenges they have and that they need to move to the cloud? And is it shifting the way customers are either coming into you or your outbound efforts?

I wish it was as easy as just letting the customers come to us. We worked that deal for 2 years, 2 to 3 years, chasing that opportunity. One of our reps just did a sort of dog on a bone -- a junkyard dog on a bone is how I thought about him. He didn't let it go. And it was heavily competed over the last year. They brought in all the competitors and did all the sort of typical things with RFP. So I give credit to Dan and his team for just incredible sales execution. But ultimately, what won the deal, in addition to that, was product capabilities. The fact that we've expanded our portfolio, the market-leading IVA that we've added was a key component of this deal. When you add all that up, sort of a 1-2 punch, right? Best technology and best sales and field engagement with the customer and again -- and also just grit not letting go on it. So amazing deal for us. It is an international deal, so it's -- we're -- we have one of our e-staff is relocated. One of my direct reports is relocating to the U.K. to establish our -- to grow our office and our presence there, service delivery capability. With these size of deals, what comes along with them is a bunch of on -- sort of like dedicated resources to support a customer of that scale and so on. So yes, we're really taking the European expansion seriously. And this is, I guess, you could say, is one of our biggest -- but this will be one of our biggest anchor tenants in that region.

Great. We're starting to get actually a good bit of investor questions coming in. Rowan, I'll give you a chance to catch your breath and have a sip of water. Barry, we have one for you. Did the company take any different approach to their guidance for 2021? Was there any difference, you think, maybe in the way that the guidance outlook was put together given that there's still some uncertainty in the market, but there's also significant tailwinds? Or just any comments you have around the framework for this year versus last year.

Absolutely. So those of you who've been following Five9 know that we've always taken a very prudent approach to our guidance. And for the last 6 years through 2019, when we gave guidance -- I'm assuming through 2020, when we gave guidance for the full year, we'd always invariably give 16% guidance and explain at the time that, that is not an indication of the health of the business. It is simply a prudent philosophy. Now unusually, this -- when we just had our last conference call and gave guidance for 2021, we gave -- instead of the 16%, which is characteristic, we gave 20%. 2% of that 4% increase -- percentage point increase is from the fact that we have made this a very successful Inference acquisition. But the other 2% is reflective of the fact that our business is just extremely strong. And if you look at it more closely in the immediate quarter, Q1, that guidance that we gave of 29% year-over-year growth is the highest we've ever had by far for any quarter. Typically, in Q1, we give about 20%. So it's reflective of that, and I'll leave it at that.

Great. Thanks for that color. I think it's very helpful. And Rowan, I received a couple of follow-ups on your commentary around the company's ability to serve large customers. I think the first one is, how many of these large customers can you either implement or serve at once? And maybe how should we think about the company's capacity to onboard large customers as demand increases there?

So I'm glad that question got asked. 3 years ago, when I came to Five9, one of the parts of the strategy that we began to implement was a channel strategy. So we hired the channel leader from Cisco who drove the collaboration side and built out over the last 3 years a very strong channel program. So that's really been what's been helping us here. In fact, in the largest deal, just to give you a specific example, we announced the European insurance company, we brought in a channel partner on that one. I don't know if we named it publicly, Barry. But we brought in a channel partner to help us deliver there. And our expansion, that scale-out in terms of leveraging channel has been going really, really well. You mentioned Zoom earlier as a UC partner. So that's sort of helping on the scaling in terms of getting leads and new deals. But we also have implementation partners. Famously -- not famously, but I guess, like for us, the systems integrator channel has always been very successful. Over the last year, we've had great success with Deloitte, and we've been expanding and diversifying away from Deloitte and getting a lot more partners in the SI front engage. So that's been very successful. Obviously, we added AT&T, if you've been paying attention to our earnings call. So that's been a big part of the channel expansion. And we have been -- CDW is another one, one of our large customers, for example. I think we mentioned them on an earnings call 3 quarters ago as one of our largest deals ever. We have CDW engaged, helping to drive more success there. So absolutely, insightful question around like how do we scale that, and it is through leveraging the partner channel. And it's a big part of our strategy going forward. It's just absolutely necessary as we scale. We won't be able to continue to hire all the people we need and deal with all of that in every industry and every large customer. So -- and the other thing I would add is, in many of these large customers, they already have a partner that they like. So it's actually about making sure we can engage and have the programs and the processes to properly engage, train and enable those partners to be part of our service delivery.

Understood. And then I got a follow-up question around the AI capabilities of the company. I know you dug quite a bit into it already, but just maybe how do you think it compares to maybe some of the other existing solutions in the market? And how can someone benchmark the quality of one platform's AI versus another's? Or I guess, how do you think about that?

Yes. Well, we think we bought the best technology in the market, and there's some industry analysts who do magic quadrant, the kind of things, and Inference shows up at the top right of some of those reports. There's only a few of them right now. But we think we bought the best technology and the best team. They took fundamentally -- so there's sort of 2 -- I would categorize this space into sort of -- in 2 ways. One, there's a sort of legacy approach. The companies that were working with company -- businesses that were working with underlying AI technologies from companies like Nuance were the sort of the legacy approach. So this idea of an IVA that you can talk to or chat bot, like natural language processing, that technology has been around for a while. And the Nuances of the world were at one time the leaders. What has changed in the last few years and the bet that Five9 made was that the hyperscalers were going to really dominate this given their investments in the consumer world around computer-human interaction -- or human-computer interaction with voice and conversation management. So Alexa, Google Home Assistant, Siri, 3 -- maybe you should use Microsoft technology. But those 3 companies are driving a huge investment in that underlying technology. So what we do that's differentiated is we actually ride on top of Google and Amazon, and we allow the customer to pick and choose what engine they want for each individual part of the process flow. So there's a complex process flow that's involved in essentially having an automated conversation with the customer. And so you can think about it as a workflow editor or a visual workflow editor. And you can essentially -- you could start with, okay, who has -- each of these vendors, the hyperscalers are competing on their automatic speech recognition technology. So some of them are better at some use cases than others. So -- or in certain languages or in certain accents. So you can pick the one that works best for you. You might choose Amazon ASR, but then you might choose Google's natural language processing. Google has got very, very -- arguably the best natural language processing out there. And then you can wire that up to a third vendor. Maybe you could use a Microsoft for the conversation and dialogue management. So we -- so the approach is sort of standing on the shoulders of giants because we said, look, you're never going to keep ahead of the hyperscalers. And my thesis is that this technology is commoditizing. So if you're out there as a contact center vendor investing in AI to do ASR, text-to-speech, natural language processing, that's a waste of money, in my opinion, because that technology is -- there's so much investment being driven by these large companies. And in the next -- because this is the new engagement model with computing in general, right, we're talking to our computers and every computer is becoming voice-enabled, it's just going to become part of the operating system. So it would be like investing in keyboard technology in the '60s, like that's not a good idea, or mouse technologies in the '90s, like, hey, that's just going to become free and part of -- and this is just the next evolution of human-computer interaction. So that's a bit like sort of wonky, but that's my view. And so we've -- Inference took that view. They make it really easy for the customer to sort of pick and choose underlying technologies, and they don't -- we have sort of essentially wholesale engagement contracts with those vendors. So you could almost think about this as the analogy might be Snowflake, right? We ride on the hyperscalers and we leverage their technologies. We do the hard part that sits on top of actually wiring them together. We have the data, that's the key. The data is in our hands, so we can leverage that data to train those models in different ways. And so they've taken a similar approach. Most of the other companies are not doing that. Most of the other competitors in this space are -- either don't have the level of sophistication that we do or they've sort of chosen double down on their own AI technology. And if I was to go invest in one of those companies, I'd be asking that critical question, like how are you guys going to stay ahead of the hyperscalers and do you think that technology is going to really be differentiated over time?

Thank you for the really thoughtful answer. And I think we have time for probably one more, maybe we can squeeze in 2. But a question we get asked a lot is how we think about Five9's long-term growth algorithm, right? So you have a hyper-growth enterprise business, slightly slower non-enterprise business, and then just this early days in terms of moving to the cloud. How do you think about it between ARPU growth, the types of customers you're adding? Just if you could dig into -- without even having to give specific percentages, how you think about that growth algorithm for those that are interested over that 3-, 5-year horizon.

Yes. Well, I think, again, it's in 2 components. The first component is the business that we're essentially in, the proven $24 billion category of contact center technology spend that's out there today. And on that front, really, it's about sales execution, global expansion and scale-out with channels. And we've been doing that really well. We're very consistent. You've seen our growth accelerate as we move up into large enterprise and as we just kind of do what Five9 is known for, which is consistent execution. So if I had to bet on one company that consistently deliver results in that expansion in the existing category, in the existing opportunity, five9 is a solid bet. The second part of it is the TAM expansion over -- and this is -- think about this as the decade-long opportunity, probably 2 decades, but I'll just focus on a 10-year horizon, which is how I'm thinking about it. So what is the big new opportunity in this space. We see it as arbitraging labor for technology spend as going to customers and saying, you're having human beings doing a lot of work that's not very efficient, that they don't really even like the agents. It's very high volume and very low value. Technology is going to be able to do that better than those human beings. And so we want to be the leader in that emerging category. That's why we bought Inference. We started with a partnership with them. And we saw very quickly -- we saw how quickly the market was moving in terms of customer demand and acceptance of this technology, and we said, okay, let's double down in this space and make a bet to win in this emerging category. And how I think about that is, again, earlier, I said, that adds up the labor spend in the -- if the tech spend in the contact center today is $24 billion, the labor spend is $224 something billion. It's $234 billion, I think, is the number we came up with. So you've got $234 billion of labor spend that's not very efficient, that we think technology can optimize. How much of that over 10 years? Well, conservatively, if you said -- I think McKinsey did a report that said 14% of that labor spend could be automated in 10 years. Now recently, Gartner predicted that 16% of it would be automated in 3 years. So the estimates are all over the map. But -- so I don't really know where the number is. But just being conservative, take 10% of the number. 10% of the labor spend is equivalent to $26 billion of spend. And so if you can capture some percentage of that and deliver some and return back to the company in terms of savings, there's a huge opportunity to grow the company over a long period of time. So it's a do what we're doing today. That gives us the base platform from which to operate. It gives us the data and the scale. And then adding on AI gives us that sort of potentially doubling of the TAM or more over a 10-year period. And we think we've got a running start.

Well, we're approaching the end of our time, but we really appreciate you joining us today. Five9 is in a really exciting time period right now, and we're excited to have recently launched coverage and get to view it more closely. And thanks again for joining us, and we look forward to staying in touch.

Thanks, Samad. Thanks for having us today, and thanks to everyone for joining.

Thanks a lot. Appreciate it. Thanks.