Why AI Will Transform Customer Experience: Cresta CEO Ping Wu and Sequoia’s Doug Leone

[00:00] Today, if you think about the business, they feel like multiple personality to the customer. So in the sales phase, [00:07] They call you very, very aggressively. [00:10] And once you sign up and become a customer, [00:13] you're dealing with entirely different personalities, right? And you're dealing with service departments. I feel like these are really disconnected, right? And I do feel like AI agents can make this entire experience a continuous, long-going conversation throughout the entire customer journey. And LRM is a perfect tool to do that. And that will really bring the level of personalization, the level of customer experience that wasn't possible before. [00:43] Bye. [00:55] - Hi, and welcome to Training Data. [00:56] Today, we're joined by Cresta CEO Ping Wu and Sequoia's Doug Leone, who sits on the Cresta board. [01:02] Today's episode dives into the gnarly world of the contact center, a giant legacy industry filled with slow-moving incumbents that is responsible for driving the vast majority of company-customer conversations. [01:13] Ping understands this world deeply, having first built Google's contact center business before becoming product leader and then CEO of Cresta. [01:22] Peng joins us to talk about the different waves of technology that have hit the call center, [01:25] how he sees the future of customer experience evolving with LLMs towards an abundance future, and why his playbook is to meet customers where they are, blending human agent assist with autonomous digital agents.

[01:37] Doug Leone also shares his perspectives from several decades of investing in company building and his hot takes on whether we're in an AI bubble. He also shares where he believes the value will accrue in AI. [01:48] Hint, it's in the application layer in this gnarly last mile. [01:52] Enjoy the show. [01:53] ping welcome to the show [01:57] And thank you for bringing me, bringing along our special guest, Doug Leone, as well on your board. My pleasure. Thank you. Thank you both for joining. Ping, I want to start by asking a big part of the AI thesis is that AI is going to replace labor globally and that the TAM is in the tens of trillions of dollars. Obviously, the contact center, the call center is a big pool of labor that, you know, is just begging to be automated. If you had to guess how much of call center labor spend will actually be automated fully by AI? [02:25] Yeah, so we internally, we've spent a lot of time debating about this. The reality is I don't think anyone knows for sure. And if you ask, it depends on really what they're selling. And you ask different people and they give you different answers. And some people will say that 100% human will be gone in contact center. And some gardeners research actually shows that none of the Fortune 500 over the next five years will have contact center connections. [02:53] gone entirely human-less. So, you know, it's also probably falling into somewhere in the middle. And in fact, we got asked this question two years ago when GPT-4 first came out. And a lot of people will say that maybe in two or three years, there will no longer be humans in the context. So at that time, our belief is that probably the transformation, especially for existing Fortune [redacted address] longer than a lot of people think.

[03:21] Mm-hmm. [03:22] Doug, what do you think? What's your bet? [03:24] At the limit? [03:26] It's a hundred percent. [03:29] But I'm mindful that there are still IBM mainframes and Cobalt being used in America in the banking system. So to me, it's not really... [03:41] What percent to me is the speed of which this is going to happen. Is it going to happen within 10, 20, 25 years, 30 years? Because whether the answer is 30% or 60%, if it happens in 50 years, that means one thing for companies like Crest. If it happens in three years, it means something else. So the end number is not the relevant metric for me. To me, it's the speed of adoption. [04:04] Great distinction. [04:05] Yeah. [04:06] Ping, you've been working in the contact center AI space for well over a decade. Prior to becoming CEO at Cresta, you ran the equivalent function over at Google. And so maybe for those of us in the audience that don't know the contact center market, can you tell us a little bit about what it is, how big it is, and how technology has served it so far? Yeah. When we first talk about contact center, a lot of people will naturally think about call centers. A lot of humans sitting there listening, answering calls. [04:36] The contact center really is a broader category that including the OMI channel interactions from emails to digital chats and, you know, on website and in apps and also including calls, of course. And the overall market is quite big. And there are.

[04:52] Historically, there are around 17 to 20 million agents, human agents actually work in the contact center. For the software market, it's probably in the tens of billions. And for AI markets, according to some research, it will be in the high tens of billions of dollars. [05:12] And is the use case mostly... [05:14] You know, customers calling in to complain, customer support. Is that what these contact centers are mostly used for? [05:19] Oh, so, yeah, so... [05:22] Customer call-in, there are all kinds of reasons they call in, right? Complaints or fixing the issue. But also, I think a lot of people may not realize that there are probably a quarter of the contact center, 25%, is actually revenue generating. They're including selling stuff or collecting money or retaining customers and that kind of conversations. So it's not 100% customer support. So I have a question for you that I never asked you. If you look at the contact center, and I'm old enough to date myself. [05:51] You go back 30 years, you heard names of Aya and God knows whoever else that's barely living in and out of bankruptcy. You go back 15 years, you see Genesis of the world. [06:04] What caused a bright young engineer called Peng Wu 15 years ago to be attracted with this market? And one could have said it's always been a stodgy market. It's always been of low interest. It always created these slow-growing companies. What is it that interested you now? Of course, now we understand it's a vibrant market with lots of opportunity. But turning the clock back 10 years ago, what attracted you to this market?

[06:31] First of all, 15 years ago, I didn't even realize that's a long history of slow growth of the market. Otherwise, maybe I would think differently. And second, at that time, I just do remember there's a period of time there's a lot of excitement of conversation, AI technology, and especially around consumer facing speakers. [07:01] contact center will probably be the most exciting opportunity for conversation AI to transform. And it's because it has all the issues that traditionally people get excited about, VC get excited about. It's a massive market. A lot of humans working there. And it's in the middle between businesses and customers, right? And it's all the interaction going through. And also no one's happy in contact center. So if you, you know, by no one, I mean, [07:31] not be too happy because the wait time is very long. And the agents, by the way, I think a lot of people may not realize the agent, the workforce attrition in contact centers is massive. It's our average is 35 to 40%. In some cases during COVID, some company have more than a hundred percent turnover. I mean, right. So it's very high stress and it's not a very fun job. And also the business also feel like they're always the opportunity to do more with less. It seems no

[08:01] I think that's a great opportunity for AI and technology to bring abundance. Abundance is the answer, in my opinion, to solve all these issues. [08:11] So you were working on this at Google 10 years ago. I imagine this was the small language model wave, the BERT days. Was the technology ready at that point? And maybe walk us through the different waves of technology that have hit the contact center. Yeah, so that's a great question. Even long before that, there's technology called IVR that you press one, two, three for different routes and for different call reasons. And then since then, there are innovation around the input. [08:41] natural language, and that's with the advance of natural language processing and TTS and text-to-speech generation, that experience is getting better and better. When we first started in contact center AI at Google, it's even before BERT, actually. It's before Transformers. It's mainly using AI, or at that time, using AI to do classification, intent classification and entity extraction, [09:11] is still manually crafted, right? So that's the last generation of technology. And then after that, of course, the transformer came along. Initially, it's also for classification purposes. Still, the experience is manually crafted, but then the LMS entirely changed the whole thing. Not only the conversation experience on automation side, but also just you can understand conversation in a way that never was able before.

[09:35] And what does that mean practically in terms of the rollout of this technology inside contact centers? Does it mean that, you know, customers were just extremely unhappy when it was IVR and then they were slightly less unhappy when you started to have kind of more transformers in the flow? And now customers are very happy to be talking to an LLM-based agent? Or how has the evolution of technology changed the customer experience? [09:56] Yeah, I think the way we would like to think about it is really from the first principle, right? And, you know, there are a lot of the conversations shouldn't even happen in our view. And, you know, the fact that it happens because the customer is not happy. [10:26] And then that usually reflects some process broken or website updates that freak out people or, you know, firmware update that bring down network and all that kind of stuff. So you need to fix that first. Right. And first, you know, avoid interaction if it's not necessary. Right. And, you know, beyond that, I do feel like, you know, their AI can automate a lot of interactions that no one wants to have. [10:56] interactions that should be self-served. And then on top of that, I do think that contact center AI will enable new interactions. That's the ones that you cannot afford to do that today. So all these are improving customer experience.

[11:09] Do you think end customers will ever prefer talking to an AI agent over a human agent? And have we reached that point yet? [11:17] So, look, I mean, that's a really interesting question. So I've been thinking about this, you know, on my way here. So I never met anyone to have this experience of talking to a customer support agent on the phone and go, I'm really frustrated. Send me your AI, please. And we never had that experience. And in fact, that I would encourage people to look up some of the, you know, some of the companies in a search for their customer service. [11:47] and Google will surface, what is the most popular question? The first question is always, how do I talk to a live person for this type of, you know, custom service? So I think that that time probably hasn't arrived fully yet. It depends on what kind of interactions again. [12:03] Thank you. [12:04] So I'm maybe too techno-optimistic or AGI-pilled here, but I feel like I've seen some recordings now where the AI can be emotionally intelligent. It has infinite patience, right? It's not trying to hit some metric on time to resolution. And so, for example, if somebody calls in and they're having a really bad day, [12:22] For example, your AI can be a lot more patient and empathetic than a human agent even could. And so I'm sort of optimistic on the side of the bots here. Well, I agree. There's the human component of patience or the subtleties of humanity, but there's also the training of the agent versus the training of the AI. Three years from now, who's going to be much more equipped to answer a question?

[12:52] Bitcoin. Some of the analogy came to my mind as you said that. It is clear that Bitcoin is going to win. It is clear that Bitcoin is going to be worth more than gold. Not investment advice. Not investment advice. But it is clear that the agents, by definition, and a lot of which don't even reside in America, there's a language component. I'm not saying anything bad about the agents, but there's a language component. There's a training component. There's the human [13:22] In all those dimensions, I think AI is going to win in the next two to three years. Bitcoin as digital gold is a really interesting analogy to the agent, the digital agent versus the human agent question. [13:35] Yeah. [13:35] From our perspective, we really want to meet customers where they are today. [13:39] So unlike self-driving cars, you really have to automate the entire thing. [13:44] 100% of time. Otherwise, you do not have the economic impact. For contact center, what we find is very unique is that the work is very divisible. So first, the conversation is, you know, those are every conversation's independent unit. And you can automate X percent of conversation that's ready to be automated. And for a lot of reason, we can get into details. And then for the remaining ones, you can still use AI to assist humans and to, you know, take away the initial maybe 10% [14:14] interactions like authentication or intake or lead qualification and then take away all the after call work. And also we have AI agents to help humans in the middle of the conversation to do knowledge retrieval, to do data entries, all that stuff. So that's not mutually exclusive. And as long as we feel like customer not ready to say that we just need to turn on our call center today and then go full AI, we feel like there is a long, you know, depending again, what kind of

[14:44] infrastructure. So I think the journey will probably take a different time frame. But our goal is really meet the customer where they are. So Cross is in an interesting position because you both have the agent assist product that helps make existing contact center agents more productive. And then you have the actual AI agent product that is directly customer facing, you know, autonomous agent. Where do you think most customers are today? Are they ready to go full force, just, you know, put the agent on my website, let it go crazy? [15:11] Are they experimenting with that? Where is the customer today? It depends on the customer. If you and I start an e-bike store today on Shopify, and we can automate 100%, I'm sure, because it really depends on how complex is your product. It can be ordered or magnitude difference between a simple product like e-bike or versus a real world touching many different countries and then millions of tens of millions of people. So it's very different. [15:41] by the contact center. And then the other part is the IT infrastructure. A lot of people may actually realize that before you actually enter the contact center, you will feel like, oh, this should be easily, very easy to automate. The reality is a lot of those things that humans do in the contact center today is optimized for humans. So those system record or the system action ticketing system, these has been around for decades. A lot of them just simply do not [16:11] that to make changes is through a graphic user interface that optimizes for humans. And without a real-time API, just, you know, again, these are not AI problems. And we believe that, you know, these are the

[16:23] opportunity that we work with our customer to develop those real-time APIs. And so that's why we feel like those transformation depends on the nature of business would take different time frame. [16:33] It's interesting you made the self-driving car analogy earlier because I was thinking about your business earlier this morning. And if you think about Tesla, part of the beauty of them getting to full autonomy is that they have so much data coming in from their cars, even when they're on L2, right? [16:47] For you guys, because you are the call center, you're the agent assist, you actually get full data of the conversation, whether it's, you know, voice, whether it's whether it's conversational based digitally. And that can become a training base for customers to automate more and more of their conversations over to the agent over time. Yes, 100 percent. And in fact, when we first my first the journey, when I first started seven, eight years ago, it's really automation only. [17:17] Fast forward, we run into all kinds of real deployments, and then we really actually broadened my own horizon. Then I believe that in order to really do the best possible automation, it's counterintuitively. You need to know what actually happened in the contact center, what are humans actually doing. So not only just the conversations, but also what they're seeing on the screen. That's super important to actually build. [17:43] the best automation possible. One of them is the sex appeal. It's the sizzle. It's what everybody wants to talk about. [17:51] which you have to have, otherwise you're a tired old company. The other is the realities of our business to run and what they need. And so if you are one of these new age companies, you're quickly going to hit a wall because you don't have the data and you don't have the systems that you really need.

[18:07] to run a contact center. But if you have the former, don't have the latter, then you're labeled as an online company. So here in our case, we understood this a while back and we make sure we invested [18:19] Not only we double down on the operational system for agent assist, but we also develop the sex appeal product because that's what a lot of customers want to talk about day one. Yeah, and another aspect of it is really just tied to the point I made earlier is that a lot of those costs shouldn't really happen. [18:38] People call in, there's no way to make them happy. [18:40] It's because they're not happy to begin with, right? And, you know, if your product works, if your process works, they shouldn't really happen. So if, look, if this room, we feel really, really cold, maybe the answer is not a heater. Maybe there's a broken window or there is a patio door wide open. The solution is turn on the light and see the root cause and then fix that first before you turn on the heater. [19:03] Yeah. [19:04] Love that. Customer support is one of those, you know, canonical examples of where people think large language models will be most transformative. And, you know, it's almost a consensus category for venture startups at this point. How do you compete? What is it like to compete when everyone has access to the same LLMs and, you know, latching on to the same big picture vision? [19:24] Yeah, so... [19:26] Again, in order to really deliver value in the contact center transformation, it's not just the models. [19:31] It's just not a model. Model is a bunch of weights and data, and itself is not going to provide a value, right? And now the question is how much you need to build on top of it to deliver that value. If that layer is very, very thin, then I would argue probably you don't have much opportunity to cue value, right? And then also, if that layer will be gone, when the model gets better, there's no way you have a durable business. But that is not the case for contact centers.

[20:01] where majority of the agencies are still on-premise and where a lot of there are so many look on average agents in the fortune 500 we look at some surveys they interact with eight to ten different systems remember these companies also apply other companies over years over decades those backhand system may not even talk to each other you know depends on where you book the flights or depends on where you book the hotel they may need to log into different systems [20:27] So that's the reality we're talking about. [20:30] So that's why, you know, we believe is our strategy is meeting customer where they are and and and then drive value on day one. [20:39] Vertical integration from the stake to the sizzle. [20:42] That's how you win. What do you think is overhyped and what's underhyped in the kind of contact center AI space right now? [20:49] For Overhype, I think it's the... [20:52] mindset of scarcity. [20:54] is... [20:56] The job displacement, I think, in the short term is probably a little overhyped. And what's underhyped... [21:02] is the mindset of abundance. Think about new experience that AI can enable. [21:10] For example, can you talk to a website? Can you directly talk to the app? And can you turn a synchronous interaction into a [21:17] asynchronous interaction? Can you talk to the airline app and say that I want you to do this XYZ and then call me back when you get it done, right? And then can you have that super... [21:29] multi-language AI agents to have those conversations. Or there are so many interactions that today you just cannot happen just simply because you do not have the staff. And then the other thing actually I feel is really underhyped is people really seem obsessed with one side

[21:48] off the conversation. [21:49] which is the workforce. [21:51] And then people ask, you know, how many of the workforce were replaced by AI? [21:56] But no one ever asked the question is how many inbound calls will be replaced by AI. So my belief is that there will be [22:04] over the next few years, you will probably see a race to getting the AI assistant. [22:09] on the consumer [22:11] aggregators. [22:12] And I, [22:13] And then a lot of things that consumer probably will delegate to the AI assistant, including making the phone calls. [22:19] So I think that's maybe an interesting thing to pay attention to. [22:22] That's really cool. Okay, so you could talk to the United Airlines app and have it, you know, asynchronously go figure something out for you and call you back. Is that something that you're working on? [22:30] Um, we're not coming on that. Okay, very cool. [22:34] Okay, I want to transition to talk a little bit about company building. [22:37] Doug? [22:38] You've been around the block for a while. You've seen the movie a few times. It means I'm old. That's what you just said. [22:45] I was trying to say it nicely. How is building a company right now? You're seeing this live with Ping. How is building a company in AI different from your last few decades of building legendary companies? [22:58] it's not very different. What I mean by that is you need a [23:05] Terrific founder. And we'll talk about the Cresta situation a little later, hopefully. [23:11] You need to... [23:13] Plug in world-class engineers at the very start. Unless you start with A-pluses, you'll never move up. You'll only be moving down. You have to plug in salespeople that are not administrators, that are fresh. Maybe they were regional sales manager early on because, one, you can't get the world-class people. And, two, if you get them, they're too big for the company. You have to figure out what the ramp is that you're willing to fund. You have to figure out what the role of marketing is.

[23:43] You have to solve this thing that I call the merchandising cycle that's been getting some play online, which is from product marketing to BDRs to revenue. Wherever that's broken, it looks like a bad sales guy, a bad VP of sales, but you have to get that right. And so I think the business fundamentals are very similar. [24:05] I do think one of the characteristics of the companies that are doing the best in AI right now is they just move with extreme speed. And maybe that's always been the case, but I think it's even more intense right now. How do you think about instilling the need for speed in the companies you work with and even at Sequoia? So I thought of answering that as part of my answer. And the reason I left it out, all the boards I'm on move with extreme speed. And that's because I paint a picture for the founders of a river. [24:31] a river with rocks. And the founders and the CEO's job [24:36] is to remove those rocks. So when you give me next year's plan, I don't care that's 150% net new AR growth. I want to know why the plan is the plan. And I want to challenge you why it's not 3x that. And maybe the answer is funding. [24:51] but we can get funding in this market. Maybe the answer is management experience. Well, that's often a good answer. Some people will say market. Well, no way that's market. We're a little company that is. And so in my mind, it's forcing [25:06] is forcing the understanding... [25:09] that these companies are capable of doing things online, [25:13] which they don't believe they are capable of doing yet, and to remove those rocks. And I push and I push and I push and I said, why can't we go faster and do it in a linear fashion? Because God forbid something isn't going to happen. If you hire 250 salespeople in Q1 and then you realize in Q3 something's wrong with the product, then you're stuck with a burn. So I'm a believer and I hear, no, we've got to train them all the time.

[25:43] so we can make mid-course corruptions up and down. And let's not be stuck by these numbers. We have 10 fingers, 100% growth. That's all bullshit. How fast can we possibly grow? That's always been the mantra and all the boards that I've served on. AI is not different. [26:01] What does Crestha need to do next? What does Crestha need to do over the next five plus years in order to become a great company, a legendary company? So, well, first of all, it has to continue to develop product. It has to continue to put one foot in front of the other. It has to always see whenever some people reach a Peter principle of their role, it has to be relatively aggressive in making sure it hires people that are capable of taking it from that point on and forward. [26:31] people that start feeling a bit like suits and administrators. Point one, that's the most important thing. But the other thing that Cresta has to do, it is to up its game in marketing. There's a lot of companies, I use the word the sizzle. There's a lot of company with a lot of [26:55] a whole bunch of steak. We're a modern company. We're best in class in one category. We're going to be best in class in the other category. We have beautiful growing run rate in both the agent assist and in the AI part of the product, in the automated part of the product. I just think we need to attach a marketing overlay so we become a household name out of the market.

[27:16] Wonderful. Well, glad you're on the podcast then. Maybe stepping back. [27:22] Doug, you've seen some market cycles. Are we in an AI bubble? [27:25] The word bubble implies you invest money in and you lose money because either due to lack of supply of companies or abundance of capital. There's certainly an abundance of capital. But I've noticed over the last two cycles, the internet cycle when Nets kept going public in 95, two great companies being built in the late 90s in Google and Amazon, a few others names like InktoMe. [27:55] Even the words I heard. [27:56] the internet is a fraud, it's not going to do anything and then [27:59] Three years later, the world went crazy. That latency was a lot less in mobile. I remember when we first looked at these apps and Jim Getzen, our former partner, said, how do you make money from a $19 app? How do you build a multi-billion dollar company? Never thinking of Airbnb, never thinking of DoorDash. A year or two later, we saw Airbnb and DoorDash. Again, that from initial birth to real market shrunk from the internet. I think this has shrunk even further. I think AI is here. [28:29] you have to invest. I think you're the front end of a cycle, which doesn't mean you have to invest in everything. But one of the mistakes that we made at Sequoia is whenever we see a bit of revenue momentum, we have some geniuses around the partners meeting that say, oh, it can stop. It can be substituted.

[28:49] Keep it very easy. You see a small company with very momentum in a front end of the market. I'm not talking about the SaaS market in 2021 where you're down to niche verticals. At the front end of the market, you start seeing the modicum of revenue momentum you lean in and you hold your nose on price. [29:06] Love that. [29:07] As you think about where value accrues in the market, there's compute. [29:11] There's other infrastructure. There's the foundation models. There's the application layer. [29:16] Where do you think value accrues? Up. Up? It always accrues up. [29:20] Just look at the gross margins as you move up markets. Look at the gross margins in the chip companies. Look at the gross margins of the system companies. Look at the gross margin of this. Well, but that's, and NVIDIA, of which we were the first investor, is a great company. Jensen was able to see the future many years ahead, and he... [29:41] Pulled one of the great, probably the greatest coup in Silicon Valley, what he did. It's just spectacular. But if we're looking over time, I think value is going to accrue to quote the application layer, what that ever looks like. It's going to accrue up near the customer, near the money, near the business user. [30:01] I agree. How do you think the AI wave is different than internet or mobile? [30:06] I thought... [30:09] of everything else [30:11] being [30:13] tools to make us more productive, [30:16] meaning we all became networked and we all became networked and mobile. I view the I-Wave as the Industrial Revolution 2.0. I think this is much...

[30:29] much larger. I remember thinking, boy, we have just seen the biggest market caps five years ago. Why is it? Because it was connectivity that created this revenue growth. Never imagined that there was this thing that was going to be much bigger than connectivity and [30:47] and the mobility. It was a complete redoing of humanity, of how humanity exists, works, lives, enjoys. And I think AI is both going to be a wonderful thing for us, and maybe even a kiss of death for us over the next 10, 20 years. [31:06] Yeah, totally agree with what Doug said. And I think one thing AI is very unique is that there are so many surprises. [31:13] There are surprises of underlying capabilities that you never seen before in Internet or mobile age. If you take... [31:23] If you take the world view in 2015 and take a time machine to give that to someone in 2007 when Steve Jobs first introduced iPhone, [31:34] I think someone can resonate with that. [31:37] And then same for Internet. I think people can kind of foresee what's coming. But for AI, I feel there's so many surprises as the underlying model gets better. There are things that even the authors for Transformer paper will not have imagined some of the capability that just came after the large language models. And that can change and surprise us. So I do think that a lot of the improvements is nonlinear.

[32:07] that's something that make it even more exciting. [32:10] You know, I'm going to remind you of something. In March of 2022, which now sounds like an eternity, [32:19] It was my last annual meeting where we meet with all the investors. [32:24] And it was a goodbye kind of thing, you know, where I present the performance and everything. And I had a slide. [32:30] that talked about all the waves, back from the chip wave to the systems wave to the land-slash-wan wave to internet... [32:39] to mobile. And the next box... [32:42] A short three and a half years ago was a question mark. [32:47] We did not know as a partnership, and we are as advanced as anybody. We are the bleeding-edge investor, right, in seed. We did not see the wave coming. And this wave has been a tsunami, and I don't think there's any end in sight. [33:03] Thank you. Thank you for sharing those insights. Do you want to talk about Cresta's technical stack or should we bug Ping on that? Well, in fact, I'm going to have to go in a few minutes because I'm in the process of recoding some of the… Are you Vibes coding the Cresta app? Yes, yes, yes. I'm Vibes coding everything. Ping, tell us about the tech stack. [33:22] Yeah, so we have a pretty broad surface or product, and I can maybe talk about the voice AI agent. [33:30] We, um... [33:32] We streaming end to end audio bi-directional and we orchestrate multiple different models. There are speech to text model and then noise cancellation model to improve the audio. There are models that detect the terms and the speech activities and to handle interruptions. And then, of course, there's a foundation model.

[33:53] and to handle the conversation. And the other side is the TTS text generation model. And then in parallel, we also run multiple smaller models to do guardrail checking and to make sure that nothing is going crazy. And as well as those models will do company-specific kind of checks, for example, never give out text advice or never give out financial promises, things like that. And then that's the runtime of voice AI agent. [34:23] There are components like running large-scale simulation to really stress test the AI agent to cover all the edge cases. There's test case management components. And similarly, if you think about our voice AI assistant, so it's also streaming audio, but... [34:42] Again, so there's a lot of similarity between the infrastructure, but it's not bidirectional, right? It's one direction. And in listening to the call and then understanding what's actually happening in the call with two humans, right? And then orchestrating 10 plus more models, actually. And, you know, in fact, similar to Vertex AutoML, we have a platform that can allow customers to build their custom models to detect interesting events in the conversation. And then marry that with workflows. [35:12] And people use that to detect fraud, even used to detect fraud, call center fraud, to train agents to how to handle objections. There are so many use cases that now with that tool we call Opera, they can express and trigger workflows. And underneath is teacher-student distillation to distill into very small models that we can run in real time and to understand two human conversations. What's the latency when I talk to one of your agents?

[35:41] So it's around below 800 milliseconds. [35:45] Wow, so it feels like talking to a human. Yes. Huh. So you're running all these models in your real time then? Yes. [35:51] Are you running open source models or are you running 11 Labs in the equivalent? So across the platform, there are 20 different models. Some are open source and fine tunes. There are small models that, for example, we only do chat or email for human agents. And we auto-complete their sentences and type ahead. Those are very, very small models. And for TTS, yes, we use 11 Labs. They're a great partner. We also use other vendors. And we constantly compare the performance. Really cool. [36:21] And then the actual meat of the conversation, though, the dialogue or the conversational flow, how do you... [36:28] How do you control that in a way that's not so rigid that it's like the IVR systems of yesterday, but not so freeform that, you know, customers can go crazy and get their refunds on airline tickets and, you know, have the bots say crazy things and embarrass the customers? Like, how do you how do you control the flow and get the best of both worlds? Yeah. So it's really just how you train humans. You give them the specification about what's the goal and these are the tools. [36:58] So there's a lot of discussion about what's workflow, what's agentic. Workflow is anything you can write it down in code. That's step-by-step. That's workflow. And car wash. Car wash is actually workflow. If you think about Oboba tea, milk tea, those are physical workflows, but they cannot do other things. For human conversation, it's very messy. It's nonlinear, right? So that's how the agentic workflow come in. That's where LLM is really good at. And...

[37:25] And then on top of those, you want to determine mism, right? And that's how we introduce the testing, the simulation, and then the guardrails to make sure that whenever you have a change in any part of the system, the behavior is still expected. [37:40] Do you tune your customers' models to, because you also have this agent assist product, so you're in the flow of all these customer conversations. Do you tune the agent to that training data or is it completely net new forward deployed engineers on site mapping out conversations? [37:57] Yeah, so we have a tool that can map from... [38:01] you know, what's actually in the human conversation to extract the blueprint. [38:05] of the conversation. [38:07] So I think the beauty of that, again, is to discover a lot of unknown unknowns. So there are a lot of topics and there's a lot of things that reason people call in you may not even know that may actually contain the call volume, a very large call volume. And then once you have that, you can now look deeper and you can use ILM to do all these analysis and extract what are 57 different ways that people express the same intent and what are the different ways that the call flow will go. And then we can summarize and extract that. [38:37] then in fact the tooling gets better, the forward deployed engineers will just be a lot more efficient. And then there are also other ways we use the human side of conversations. For example, [38:47] we extract the model for the visitors [38:49] So that's how you build your simulation. And the simulation is a huge part of improving the AI agent. And we believe that having access to exactly how your real customer humans come in and describe ways and in different ways, sometimes very messy. You can extract a model and then to better simulation on your AI agent as well.

[39:09] And then what methods do you use to make LLMs really bespoke for a customer environment? Like, is it? [39:14] RAG? Is it prompt engineering? Is it fine-tuning? Is it all of the above? Reinforcement learning? What are you most optimistic on in terms of techniques? [39:22] Yes, so we use almost everything. So definitely prompting and then RAG and for those simpler agents. But we're still exploring, you know, by looking at the human behavior and then the outcomes, how do you use RL to improve this end-to-end performance? [39:42] and [39:42] But... [39:43] For AI agent by itself, I think the foundation model itself is already pretty good. You just need to get the best out of it. [39:50] at least for a digital channel for chat. [39:52] but for [39:53] Other use cases, there's a lot of opportunity to fine tune the models and to make them, you know, for tasks like summarization, for tasks like auto completion sentences and that kind of stuff. I feel like there's a lot of room to extract from the fine tuning open source models. [40:11] What goes into building a successful fleshy demo versus production-ready AI systems? [40:17] Yeah, so that's a really interesting question because I think one thing, [40:21] unique about AI is that there's a huge gap between the demo and production. And on one end of spectrum, you have rocket launches. The rocket launches, the demo is the production, and the production is the demo. You cannot fake it, right? But for AI, it's a little different. And, you know, I can just give you an example, right? So auto-summary. Auto-summary feels like a commodity capability that, you know, anyone can use ChatGPT to create auto-summary. But in order to deploy in

[40:50] 20,000 people across multiple continents and call centers and a huge list of challenges. First, how do you get the real-time audio? [41:00] In a demo, you can demo very easy on Twilio in the cloud. But remember, 50% of the conversation happened on-premise, right? And then... [41:09] And then sometimes, you know, how to access that will cause you a lot of money as well. And how do you go around that? And then in the real call, 20,000 agent calls, there are transfers. There are a lot of transfers. And then there are third parties. [41:23] callers that come in as healthcare specialists. All that needs to be transcribed and summarized. And sometimes the conversation goes so long. How do you handle like three hour, four hour calls that go beyond the contact window? Right. And then things like, you know, is there background noise? And then things like, you know, for different core reasons, there can be different templates. You really, really want to extract these type of information. You cannot miss that. How do you make sure you do that almost 100% of the time? And by the way, how do you handle PIs? [41:53] And then you cannot have the personality find information, you know, on REST. And then, by the way, how do you handle, you know, data residency if you're talking to – [42:04] multi-continental, multinational bank or our healthcare provider. So all these have become additional requirements that make something that would feel very commoditized, like, you know, out of summary, become very, very much harder to do in actually a contact center. And that's why you need a product-minded chief executive officer for one of these companies. Absolutely. And this is also why all the pain and all the value is in the last mile. This is why

[42:34] Thank you. [42:35] Yeah, I tend to agree with that. Talk to us about the future. What happens if everything goes right? What does that mean for Cresta? And what does that mean for the world? I think that AI will just like any technology before it. [42:47] like electricity, it will disappear. [42:49] It will disappear into workflows. And I think, you know, 20, 30 years later, no one will realize that they may actually talking to AI or is a human assisted by AI. There's one thing I'm really excited about is that today, if you think about a business, [43:04] right and they feel like multiple personality to the customer so in the sales phase [43:10] or the marketing phase, they really, really want to talk to you. [43:13] They call you very, very aggressively. [43:16] And once you sign up and become a customer, [43:19] you're dealing with entirely different personality, right? And you're dealing with service departments and they tend to use the terms like [43:27] tier defense, deflection, [43:31] And to just handle, you know, to refer to the exact person that they were quoting just a few days ago. And then even if you have a long conversation on the customer support line and share a lot of feedback, two weeks later, another department will come in. What's your feedback? How about you fill out this survey, you know, to our business? I feel like these are really disconnected, right? And I do feel like AI agents can make this entire experience a continuous, long-going conversation throughout the entire customer journey. [44:01] tool to do that. And that will really bring the level of personalization, the level of customer experience that wasn't possible before. Yeah. The point that really stuck with me that you said earlier was about kind of the scarcity versus the abundance mindset and how much can business to customer communications really evolve and app experiences really evolve if you take the abundance mindset to bring LLMs into this field. Thank you, Ping. Thank you, Doug, for joining us today. I love this conversation. Thank you. Thank you for having us.

[44:31] Thank you. [44:39] you