Industries at the Intersection: Customer Experience

Sriram Viswanathan: Hello everyone. This is the TechSurge Deep Tech Podcast presented by Celesta Capital. Each episode, we spotlight issues and voices at the intersection of emerging technologies, company building, and venture investments. My name is Sriram Viswanathan, founding managing partner at Celesta Capital. This is the latest episode in our Industries at Intersection series, looking at customer service and call center industry.

One of the places where A.I. is expected to have a major impact. We're joined here at this episode by the founders of a pretty interesting company called Crescendo A. I. Crescendo provides businesses with outsource management of the customer service and contact center operations. Using AI to augment the human expertise and deliver a much higher level of service to its customers.

There is a lot of hype and questions around AI right now. And as you guys know, and our audience will appreciate the customer experience in general is a sector that is expected to be significantly impacted. And it's, uh, expected to actually be implementing large scale AI applications and use cases. And it's presumably one of the first large industries that might get disrupted by this technology.

And as our audience may knows, Celesta invests quite heavily. In deep tech, especially in the areas around industry transformation. So that is the context of, uh, crescendo. We are an investor in crescendo along with a general catalyst. There's a lot of stuff to dig in, into the session to help us through this topic.

We're joined here by crescendo AI's two. co founders, Andy Lee and Anand Chandrasekaran. Andy's the founder and executive chairman of Alarica, one of the business process outsourcing companies in the world. Alarica has more than 100, 000 employees globally and manages more than 2 billion customer interactions each year for all of its customers.

Anand, is a very seasoned product leader and investor, uh, a long history of, uh, investing and building large platforms. He most recently served as a partner at General Catalyst, one of the storied, uh, large scale venture capital firm in the Valley, uh, where he actually incubated this along with his other partners at, uh, General Catalyst to launch Crescendo.

Uh, he's been recognized as one of the top seed investors in India for the past three years. So, uh, congratulations Anand. And earlier in his career, Anand was actually a product exec at, uh, Meta, Five9, Bharti, Yahoo, and a bunch of other technology companies. Andy and Anand, great pleasure to have you both on this podcast.

Thank you for joining me in this conversation. Thank you for having us. Thank you. Sure. All right. Well, let's just jump right into it. Let's start by looking at the contact center industry. I think our audience probably knows quite a bit about the whole BPO space. And everybody, of course, knows about what is really driving the growth of the BPO call center industry over the last, you know, A couple of decades and largely it's driven by, uh, cheap labor and cheap manpower, which presumably, uh, has resulted in sometimes inconsistent quality, uh, sometimes, you know, linguistic issues with some of the customers.

And this is an industry that is almost half a trillion dollars. In revenue. Um, and so there's a great sector with which could get impacted by generative AI. And there are lots of people going after this segment. Half the customer service executives expect AI to have the most impact, uh, in their businesses.

Within the next year, according to Gardner. So, to kick us off, Andy, you know, why don't you give us a grounding on the dynamics of the sector? Since you've been a, uh, an entrepreneur catering to this industry, albeit not with AI, but with the traditional models, so can you just sort of give us some, uh, basic foundation on the market pressures, and what are some of the preconditions that make this particular industry ripe for AI adoption?

Andy Lee: Great question. I think, uh, it's actually a fundamental issue that I observed 25 years ago when I started Allorca was that I felt that the business construct, meaning the relationship between CX, the customer experience provider and the business was not totally aligned. Because historically, and even up to today, the entire business is based on units of labor hours.

And when you think about what businesses, uh, want to buy, what they want to buy is an outcome for their customer, um, but they want to do it in a way that is most effective and efficient. And traditionally, businesses have been challenged to manage and balance the, the issue of, um, wanting to provide the best customer experience for their customers with the challenge of the cost of doing so. And so historically, what businesses have done is they've tried to manage that cost by implementing deflection strategies. Right. And deflection strategies could be things like self help or, um, an automated voice response system like the, uh, IVR, the integrated voice, uh, response system where, where a machine is trying to triage what you're actually calling about before they route you to a human, or they'll slow up, slow down how quickly they pick up the phone.

Right. All, all to kind of manage to a budget. And I think the result, the outcome of that is consumers not being able to interact with businesses, how, when, um, how, and when they want, and businesses not being able to hear a hundred percent of customer issues and opportunities. Um, at the same time, you know, because the pricing is based on labor hours and clients have, uh, you know, cost pressures. Ultimately, this service has gotten commoditized. The, the, the customer experience labor on component has gotten commoditized over, over time, and that's compressed providers economics as well. And so when you think about that ecosystem, there's not a lot of excess dollars or even incentives.

on either side to really think about how to improve the customer experience while reducing overall costs. And so I, I observed that 25 years ago and something that's, um, I've wanted to change. Um, but to change an industry, it takes, it takes a lot. I mean, it just, it's not easy to change how people buy, how they view a service.

And, and there hasn't been a, what I would call spark technology. That's, um, enabled that that transition and that transformation of how business is done. Um, so like my, my, my excitement of, of, of building this new augmented AI business is that I think we're going to be able to finally solve that. Um, industry issue, which is, um, historically, the industry has been trying to find a balance between cost and, and, and really serving customers.

And Gen AI to me is the enabling technology that's going to help us break through that.

Sriram Viswanathan: Yeah. So, so we'll get into that in a little bit more detail, but before we do that, just, just putting a fine point on Alerica itself. Alerica is a traditional call center. BPO kind of company. Uh, can you just give us some, some financial metrics and size of what Alorica, uh, is, uh, is catering to today?

Andy Lee: Sure. Uh, Alorica today serves almost every vertical. So anything from retail e commerce to financial services, healthcare, uh, consumer products, there's not a single. Segment of the industry that Alorica doesn't touch today. Um, it serves typically the largest and most reputable companies in each vertical.

It's around a 2 billion revenue business today. Our operates around the world, primarily focused on serving North American end users. But delivering that customer experience service from, uh, around the world. So countries like the Philippines or India or Mexico, Guatemala, Honduras, Jamaica, Columbia. So we, we have an incredible delivery, uh, infrastructure, but it's all with people, right?

And what we've been able to do is help our clients manage their customer interactions, um, whether it's customer service. sales, collections, uh, whatever, whatever the interaction may be, Alorica can handle those interactions and provides the labor to, to do so.

Sriram Viswanathan: So in a way, you're probably looking at AI as, uh, both a threat and opportunity for Alorica, because presumably, AI could completely disrupt your business if you don't address it.

Somebody else could just create what you're talking about with hundreds of thousands of people, uh, instead of doing that using a super smart AI agent, correct?

Andy Lee: 100%. I think if you were to talk to, uh, industry leaders across the board, I think there's a lot of fear around what AI and generated AI specifically will do to the industry on a, from a personal perspective.

Uh, I think it's a great opportunity because, uh, I actually think that because of generative AI, we're going to see companies engage in significantly more, um, interactions with their customers. Right? Um, I think the aperture at the top of the funnel in terms of quantum of conversations and types of conversations are going to ex expand.

In my opinion, I think it's going to expand by, by 10X. And because of that, we also see that AI can handle somewhere between 60 to 90 percent of conversations. transactions depending on type of transactions. So if you start to think about that kind of dynamic, if the overall pool increases by 10 times, right?

And AI is handling, let's say seven, seven X of that, there's still three times more human interactions that are going to, that are going to be, that are going to need to be handled. And our belief is that those interactions are going to be significantly more complex than they are today, which means that there's going to be an upscaling of labor as well as an expansion of labor

Sriram Viswanathan: and we'll, we'll get into this in terms of quality and how reliable the AI agent is, uh, with respect to the responses and customer satisfaction and whether that's equal to, you know, a human in the loop kind of, uh, Uh, an experience. So, so, so to, to sort of expand on that on an, you know, I want to turn to you at General Catalyst.

You've been sort of, you know, you joined there more as an investor and you're in your personal capacity. You've been sort of, uh, funding and seed investing in a number of companies. What was the reason why you saw this and how, tell us the backstory of how you and Andy got hooked up to create Crescendo.

Anand Chandrasekaran: Yeah, thanks Sriram. I think that, uh, largely it was this realization that Contact centers are a natural area where automation can really make a huge difference. So my last job before, uh, joining General Catalyst was at Five9 where I was leading product and saw sort of the previous generation of software in the space.

The thing that strikes you is Customers would buy a lot of software from us and the more software they bought and the more complicated the integrations became Uh the more their cx actually sucked, right? Which is ironic because they're making all these investments to actually improve their cx, right?

And what we saw was that Uh in addition to everything else Just the rollout process and the pre integration and everything not being available as a service that delivers an outcome was sort of at the heart of why they were never able to get great CX, right? So when I was starting to explore how Gen AI can impact our industry, we started to look at what is the biggest problem that we should solve, which is a 25 year old problem that Andy talked about, which is the unit of value is labor.

Uh, the unit of value is not automation and, uh, the software that helps, uh, you know, uh, sort of deliver an outcome instead of just delivering labor. It's just a bad surrogate for, uh, an outcome, right? And so we have this moment where the level of automation that we can provide is so high, but also, you know, the level of outcome that can be delivered, both, uh, accountability as well as the monitoring, as well as the alignment.

Between the enterprise and the service provider is finally possible for the first time in 2025 years. So when Andy and I first met, you know, we were connected by one of the founders of Observe. ai where I was a early investor and Andy was working with them super closely, uh, you know, uh, to deploy their technology.

My, I reacted to three things. One was Andy was just like an incredible first principles thinker. Which when, you know, the companies become really big, uh, the founders become more of CEOs and they lose a lot of the entrepreneurial first principles thinking. But despite having built a lot of car to a top five BPO, Andy was just like starting from first principles and being curious about what Jenny I could do for our industry.

Secondly, Andy had really been one of the few people in our industry, uh, from really any background who had scaled a company through the stages, right? Early stage. Uh, sort of hundreds of millions, now billions of dollars and had taken the customers through all of those journeys and managed to retain, you know, most, not all of them.

And so we knew that we could build a very enduring business together with that actually business scaling experience. And then the last thing I would say, which has been super helpful to us is the. inorganic and kind of non dilutive ability to grow through whether it's buy and transform or just thinking about the capital architecture, right?

And there are lots of founders who, when they build these, uh, businesses, don't think about the capital architecture that's needed to build and scale the business. And Andy had done that very intuitively through his entire a lot of our journey. So when we met, it was just like a great fit. Uh, for us to, you know, start from a technology perspective, start from a, uh, you know, BPO business perspective.

And then we really built a multidisciplinary team around the two of us.

Sriram Viswanathan: So, so this is great. I, I want to, I want to double click on a couple of things that you said here, which is, uh, you know, the BPO industry historically is, uh, is notorious for lack of innovation. As you said, it's, you know, the unit of value is the labor or unit of constraint is labor.

You can't scale the business because labor is constrained. And, and if the, if, you know, every day the BPO's, uh, company's value walks out the door in the evening and you have no guarantee they're going to show up the next day. So with that as the core business that, Andy, you've built, uh, You know, it's conceivable that AI could completely replace your core value that you've built in Alerica.

I mean, do you worry about that? How does that impact? Because Crescendo and Alerica have no, no connection today. These are two completely separate companies. One could, you know, it's kind of like the, you know, one could eat the, eat the other. So how, how do you think about that?

Andy Lee: So I think one of the things that I, I talked about a bit earlier is that I think generally fear is one of the most, um, powerful emotions that drive people's, um, decisioning and actions.

And I think when something like this happens, a lot of times people get paralyzed with the fear of what if, what, what, what's going to happen and how is this going to change? And if you're gone on point, if you're running a big business, you're, you're, you're generally in the, in a protectionist mode.

You're in the mode of like, how do I protect this business and create sustainability going forward? And I think a lot of folks in our industry today are, are, are, are faced with that, that question. I think when you're faced with that question, a better way to do it is to lean in and really think about what is the world, what's the world going to look like in five years because of this technology.

How are things going to be different? And what, what things will consumers be willing to embrace and how will they change their behaviors? And I think the smart businesses will observe those, those consumer changes in behavior and think about how to leverage technology like Gen AI. To meet customers where they want to be met and help them solve and resolve and buy, um, at in, in the way that they want to and when they want to, right?

So to me, this is an opportunity again, like I mentioned, this is an opportunity to completely rethink. Um, how businesses connect with customers and that's, that's, that's beyond exciting because nothing really transform transformative has happened. Um, in, in, in like the last 25, 30 years. So this isn't like, to me, this is an opportunity.

Um, is it scary? Yeah, it can be scary. But, um, I think the opportunities far outweigh the, the, the, the negatives. Yeah. All right. So,

Sriram Viswanathan: so let's, let's get into some of the technical details. Anand, you in your prior life, you helped scale Metas, Facebook's messenger platform to over a billion users. What exactly is the technical architecture of Crescendo because I understand the, the, the old traditional business of, uh, you know, BPOs like Alorica.

And then of course, Crescendo is, you know, infusing a whole bunch of technology, but is it building large language models per vertical and then incorporating that into the core workflow? We can just talk about the technical aspects of it.

Anand Chandrasekaran: Yeah. So I think what, what we attempted to do, it's going to become a bit more mainstream in terms of how companies get built in AI in our view, Sriram, which is, I think we need to have multidisciplinary teams that cover off all the key elements that we need to be excellent at.

In our case, it is a combination of AI, BPO and contact center. And why is that the case? So we have Slava who, you know, was CTO of Genesis. We have Matt who is, uh, you know, one of the SVPs and leading sort of architects of Zendesk as they scaled. And then Todd Famous from Genesis, Andy, myself, and, and many others.

The reason I think we need that multidisciplinary team is because we're going to have to build an architecture that actually delivers the entire outcome, right? So whether it is something we can fully automate, partially automate. Uh, or we shouldn't automate at all. We're responsible to deliver the outcome, and we get paid for the outcome that we deliver.

Right? So while we don't build the underlying LLMs, we've built a lot of layers around it to ensure that that outcome gets delivered. Right? To give you an example of that, why has this not happened over the last Right. We've thought really long and hard about the reasons why this has not become mainstream.

And one of the biggest reasons is accuracy. Another big reason is brand exposure. And with AI, you have hallucinations. So that's where we focus a lot of our tech and our architecture is not to, you reinvent the wheel and you know, what's already been done, but to really ensure we take care of the reasons why this has never gotten into the mainstream.

And those are reasons like the ones that we just talked about. And so one thing that we do is we guarantee that the brand will not be exposed. Um, and we focus the questions and answers on customer care, for example. So that's a key piece of technology that we've built.

Sriram Viswanathan: If I may interrupt you, is that based on a set of language models for, let's say for healthcare or financial services or it support, what, what are you, how are you building the knowledge base?

Uh, for that question. And how are you continuously improving the knowledge base?

Anand Chandrasekaran: Yeah. So we, uh, by owning every interaction, we're able to learn from literally what happens in every interaction in terms of, are we automating it right? We track the sentiment of that interaction and say, was the customer happy when we automated this transaction?

Is it better than if a human had done it and so on. And by owning the entire interaction, sometimes we may actually start by doing it manually. But within a few weeks, we know how to automate it and, you know, then we can sort of automate it and deliver it at a, at a margin profile. Uh, what we are loyal to is sort of The top line and the growth profile of a BPO because we're going after the labor budget, which is for every 1 spent on software in contact centers, about 10 to 12 are spent on labor.

So we're going after the labor budget, but in order to go after the labor budget, we should be confident of delivering the outcome one way or the other.

Andy Lee: Right. Yeah.

Anand Chandrasekaran: So we mostly automate, as Andy mentioned, 50 to 70% through ai. But we do have the humans in the loop for the other 30%, which makes us incredibly confident of delivering the outcome.

Sriram Viswanathan: So, so are you, are you built on, uh, some of the established models, uh, LAMA or GPT or what, what is the core architecture built on?

Anand Chandrasekaran: Yeah, so we, uh, uh, leverage all of the top LLMs. Uh, in fact, uh, you know, embarrassingly, I wrote in my original memo that we'd have to build our own LLM. Not, not being aware of how fast the LLMs would evolve.

And, uh, we've really been able to kind of ride the shoulders of giants, so to say, and take advantage of all the capabilities that have happened, including on voice. So originally I thought that a lot of the LLM innovation was on the data side, but on the voice. Uh, we'd have to recreate our own, but it's been incredible how much of it we've been able to, uh, kind of piece together, which has allowed us to focus on innovation layers on top of it, which is what a lot of application software companies do is that they don't recreate the underlying, um, technologies.

And so we've been able to, you know, bring that, uh, same, uh, Uh, kind of innovation to the, to the AI space as well.

Sriram Viswanathan: So, so Andy, given the fact that you've actually worked with large enterprises from your current role as the founder of Alorica, how did you guys decide to build Crescendo AI in its current instantiation?

And do you think that you would actually go to market by licensing the core? architecture and engine to large enterprises for them to build their own contact centers? Or would they have to come to Crescendo? Are you, are you kind of the aggregator of the contact center experience for any number of enterprises?

Or would you do a white box kind of an offering to enterprises? It's great questions.

Andy Lee: I mean, this, so we've spent a lot of time on that question, right? Um, because when you think about, Why CX has not made quantum leaps in terms of improved improvement. And you look at the quantum of different technologies that have been tried, that people have tried to introduce into the marketplace.

The one thing that the commonality that we see across the board, whether large enterprise or small enterprise, is this concept of complexity. And if a business wants to, um, enact a new strategy in CX and bring in technology, they ultimately have to behave like a conductor, so to speak, um, and have the ability to select the right technologies, integrate them, and then find the right labor.

Providers and pools and then orchestrate all this to ultimately deliver a seamless experience to their customers. Historically, I mean, well, it sounds simple. This is where we've seen companies not able to execute crisply. And so when we thought about how we were going to go to market, our number one intent was to make buying CX AI services extremely easy, right?

And what does that really mean? Mhm. Right, so if you start this concept that Crescendo's underlying goal is to help companies leverage and enable AI to serve their customers in a more effective and efficient manner. To sell them a tool and then ask them to go do it, that would require that they would have to have talent, expertise and bandwidth to, to not only, um, select, implement, deliver, and then, and then ongoing manage, right. Um, because all these models need to be continuously managed to stay, stay, stay relevant and accurate. And so we thought like, how do we create a product slash service offering that. eliminates the complexity and just delivers an outcome to businesses and their customers.

And so to that question, what today, if you want to work with crescendo, the, the, the, the, the only way that we're, we're selling today is that you have to buy the entire experience. Right. Which is crescendo shows up. You say, please make it happen. And, and crescendo builds all of your, um, AI, uh, for, for both voice and digital channels, um, creates that front, that first level of experience, but also augments that technology with humans behind so that if the technology is Isn't able to resolve.

There is a human behind it that can resolve at the same time. These humans who are resolving at the, at the back, at the back end, they're also our knowledge engineers. And they're, they're also responsible to help continually feed our models and continue to improve them over time. And so like, as a, as a business, As a buyer of this service, you don't need all the technical, uh, capabilities in house.

You, you don't need to, um, have CX practitioners to figure out how to design the most, um, effective and, and, and consumer pleasing experience. You just basically say, crescendo, make it happen. And, and we show up and, and, and effectuate it. I think that, that, that model, I think will work for a certain segment of customers.

To begin with, and our goal is to continue to prove this out, you know, I think many larger enterprises may have, um, different large departments that have different perspectives on how they want to solve for that. But our, our, our approach today is to simplify and deliver an outcome. And we want to work with customers, uh, who, who, uh, want to experience the benefits of AI quickly and without complexity.

Sriram Viswanathan: And what sectors or industries are you targeting first? And what are the ones that you don't think is appropriate for Crescendo or AI to actually be able to deliver the best experience for?

Andy Lee: Well, today, I think based on where we are, I think where we see the most applicability of, of Gen AI is on informational opportunities.

So doing pre sales, doing customer service, answering, uh, questions that, uh, consumers may have about the products or services they bought. I, I don't, we don't see a lot of delineation between different verticals, but we do see delineation on complexity. So obviously as, as, as a newer business that's just started, our focus is going to be on those businesses that we feel like we can deliver that experience faster with less friction.

Anand Chandrasekaran: uh, if I can add one, um, no, we, we use this metaphor for the full stack offering that we've been building, which is, uh, when, you know, smartphones and GPS became mainstream on smart on, on, uh, iPhone and Android, that was sort of what enabled Uber. To go build a taxi hailing software, right?

And they sort of had a choice to sell that software to other taxi cab operators or build a taxi cab service themselves. And obviously they would be competing with the folks who had little to no technology with a big technology advantage. So I think the models exist. And of course we know, you know, which choice Uber made and resulted in a generational business.

And so I think that that's sort of the, what, what has convinced us to say, We have to take this entire stack to market. That's how, you know, if you look at the cloud penetration in the contact center industry, it's 12 to 15%, right? So if we don't bring a full stack model to market, we'll have sort of this anemic adoption that's happened over the last 10 years for the next 10 years as well.

And the way we sort of spurred adoption from. You know, 12 to 15%, like 30, 40, 50%, because if this is not fully digitized and not fully on the cloud, there'll be no chance for AI to be adopted. Right. So this has to, this is the best delivery model.

Sriram Viswanathan: Well, you know, what you're really saying is that the traditional BPO model, as you said, the labor intensive, low margin, you know, highly scale inefficient sort of model, uh, moves to a software model.

With AI, you know, very low CapEx, very low OpEx, and you're, you're really driving the value out of all the AI implementation. Uh, and, and so the margin structures and steady stream, you're, you're, I assume you probably are in the 50 plus level, uh, margins, uh, profile.

Anand Chandrasekaran: Yeah. And, and constantly increasing, right?

So, uh, you know, again, the, these models are actually quite familiar to us, right? So if you are, uh, Call it an Amazon prime subscriber, right? You know that there is some sort of an interplay between robots and software and warehouses and people and trucks, but you honestly don't care how the medley of all of that comes together, right?

All you care is that the network is more, uh, uh, has more throughput. And that your outcome gets delivered better and better, right?

Anand Chandrasekaran: If they're able to do it at a, at a better margin profile, that's a little bit transparent to the, to the consumer, who's like still paying the subscription fee. And so these models have existed in, um, in the consumer space, for example, but it required someone to thoughtfully invest in the entire solution, the full stack solution to bring, uh, two people before having a single subscriber, right? So we've built six products on the backend that we bring together that are all pre integrated so that the customers can go live in a matter of like two to four weeks.

Sriram Viswanathan: I got it. Okay. All right. Well, let's just talk about how you guys are scaling, because you've had a bit of news in, in the, in the market already, or barely nine, nine months old.

And you have, uh, uh, you know, you've done a financing at a pretty, uh, amazing valuation, you know, no complaints. We're participants in that and we're very happy about the fact, but tell us how, how you scaled so quickly and your capital structure, and how did you actually scale the overall business in the last nine months?

Andy Lee: Maybe you can talk a little about the capital structure and I'll talk a little bit about the, uh, some of the M& A work we're doing and why. Yeah. So maybe I'll just start off with the M& A side. Um, look, one of the, one of the things that is always challenging for a startup is to get a market penetration and, and get new customers.

And, um, one of the things that, uh, we thought about early on is like, how do we accelerate, uh, adoption and confidence. Um, and have, uh, proof points in the marketplace, uh, that of what, what our, what our product service can, can deliver. And so our strategy was like, let's go find a company that is serving the customers that we're interested in serving and use that as a sandbox, so to speak, to prove out our, our, our model and our, um, our services.

So that's, that's kind of the, the, the rationale on the fly, right? Um, because I think the market. will move fast and, uh, and the technology is available. And so therefore, like, how do we get this into as many people's hands as quickly as possible? Right. That's, that was really the impetus on, on the why of M&A.

Sriram Viswanathan: And you're, you're specifically referring to your acquisition of, uh, a partner here? That's

Andy Lee: correct. That's correct. So, you know, obviously we've, we have our own internal sales teams and go to market teams that are focused on new customer acquisitions, but we felt that this strategy would help accelerate, um, by buying partner hero, we felt like we could accelerate the adoption of AI and, uh, create, uh, Okay, for the

Sriram Viswanathan: benefit of our listeners, Crescendo, a small startup, barely gotten out the gate.

Why is PartnerHero, a much larger firm with larger profile employees and all of that, it's kind of like the snake eating the goat. Do you guys worry about indigestion or, you know, how do you think about that?

Andy Lee: So, so the, uh, I think, you know, we essentially, if you think about partner heroes, probably like an 80, 85 million business, a year business, 3000 plus employees that it does on the face of it seemed daunting to, um, to do something of that size and scale, but, you know, over my, over my history, I've bought many Um, companies in this space, um, you know, Lorca today manages a hundred thousand employees.

So for us, it doesn't feel as daunting it, uh, you know, it's, it's, it's much more manageable having, having gone through it many times, um, understanding the intricacies of trying to, uh, meld cultures together, um, you know, understanding that. Like that whole, that whole motion of, uh, aligning, aligning organizations behind one mission, understanding how important it is to get the entire organization to understand the vision of the company.

Those are things that our team at Crescendo has been really focused on in the first 30 days, because we know that if we get that right, we'll move a lot faster, um, down the

Anand Chandrasekaran: And, and I think Shuram, on the, on the capital side, Uh, we had underwritten each milestone, uh, with a structure, uh, you know, in the, earlier in the conversation, we talk about, uh, thinking a priori about the capital architecture.

So we had talked about, uh, like we'll underwrite the initial risk to build the product, assemble the team. Then we wanted to deploy in like a dozen customer sites and get them all up and running and prove the margin profile expansion work that we were doing. But once we had done those, uh, two or three milestones.

Uh, we wanted to underwrite the, the buy and transform as well with PartnerHero.

Sriram Viswanathan: Yeah, uh, really what you're saying is that not only AI is a disruptive technology in the PPO industry, you also had a very disruptive approach to the capital structure and the financing of this. I'm curious, you know, the, the PartnerHero founding team, how did you convince them?

Because they had 3000 employees. And if you are widely successful in your strategy, You probably need 300 employees in PartnerHero. So, how were you able to convince the guys to integrate with you, uh, knowing full well that they're going to be skinnied down pretty dramatically because of the technology that they're going to start embracing with you?

Andy Lee: I think there's a couple of, um, Things in the, in there one, um, I think partner hero, um, obviously like every other BPO company, um, was concerned about, um, where, like how the industry was going to change and how it would affect them. Right. So I think they had a strong belief. Um, uh, Shervin, the, uh, founder of partner hero had a, was very aligned with our vision of, um, where, where the industry.

Um, was going to go. And so he really had two choices, right? Um, he could lean in and try to do it himself, um, lean in and find a partner or just let things happen to them. Right. And, um, like, I think when you think about the things that are necessary to transform an industry, it is very, very hard to do it alone.

Right. When I started at Lorca, I started with 10, 000, uh, never raised any outside capital. So really was doing it Alone. And when I, when I sat down and said, okay, why did I fail to transform the industry? What were the things that I lacked? And I, and I came to the conclusion that to change an industry, you need the power of many.

The power of many is, is the concept that of bringing people from many different, um, walks of life who all have ability to influence, um, how people think, um, what they believe and, and ultimately how they act. And what we're really talking about is changing people's mindset on what customer experience is.

How it should be bought and how it should be measured and how it should be evaluated. Right. And so when you think about that, you need a much more flexible capital strategy because it's going to take time. Right. Um, but you need lots of people in the ecosystem, in your ecosystem to be able to influence every aspect, uh, whether it's on the finance side, on the consumer side, on the industry analyst side.

I mean, everybody needs to kind of Be educated and, and come along for the ride, so to speak. And, and I don't think that, um, doing it alone is, is, is, is, is necessarily, that's a low, low percentage shot. Like I've, I can only think of a few people who've done that, like Elon Musk, you know, Jeff Bezos, Mark Benioff, right?

Those are people who have created new industry sectors, right? But those, that, that, that, the, the, the quantum of people who can do it on their own is really high. Very, very, very, very far small. Can I, can I

Sriram Viswanathan: tell you what is amazing about what you guys are doing? And, and this is going to lead me to my next question, which I think you may choose not to answer in the case of the example that you gave about Elon starting, you know, EV or Tesla, whatever, he did not have a old, you know, uh, ice, uh, car company that he was disrupting.

And here in your case, specifically, Andy. You have Alorica, 100, 000 employee, traditional BPO, low tech, low margin. You know, you didn't take any money, you know, kudos to you for having scaled it to a 2 billion revenue business. And now you're starting another company with Anand and building something that can fundamentally disrupt your existing business.

So, so the obvious question I'm going to ask you is Alorica the next acquisition that you guys will make? You know, I think, uh,

Andy Lee: I think there's, there's lots to be learned from both sides, right? I think what Alorica does today is unique and special in that it serves its customers the way that they want to be served today.

What Crescendo is working on is to show customers another way, right? And do I think that one day those, those roads may intersect? I think it's very, it's very possible, right? Um, but I think the, the, the key is to enable each organization to continue to be successful, right? Do I think that, um, that humans, that the entire, uh, CX, uh, industry will transform in six months?

No. It's not possible. Like, I think when humans tend to extrapolate outcomes much faster than they actually happen in, in real time. And so, uh, companies like Alorica will need to continue to exist because their customers need them to continue to do what they are doing. Um, Uh, for, for, for the foreseeable future.

Sriram Viswanathan: Great, great answer. I didn't expect you to tell me that that's the next acquisition you're going to do next Tuesday. But, but Ana, I think to, to, to the point that Andy makes, you know, notwithstanding the pace of innovation happening in AI, is it fair to say You know, I can see why human in the loop is going to be critical for automotive industry, for self driving cars or robot taxis.

You know, in the case of customer experience, even if you didn't have a human in the loop, you can deliver a certain basic level of customer service and quality and all of that. Uh, do you see a need for human in the loop for the longterm? And if so, what kind of ratios, uh, you know, and scaling?

Anand Chandrasekaran: Yeah. Uh, I mean, I think if you're trying to go after, you know, I think we started to talk about a 500 to 750 billion TAM, um, as we start to serve that large of a market, you're going to have an assortment of capabilities that you deliver, right?

Like it'll be tuned for vertical, for You know, uh, for each scenario, you know, for geographies, there'll be, there'll be areas where there'll be a little bit of nationalism where data is stored and where humans are based and so on as AI becomes the default way to, uh, provide these services. So, you know, as we're building a generational company, we have to think about sort of providing that entire outcome.

to the customer. And, you know, when we think of like what kind of business we're building, we're, you know, there have been like a few multi hundred million software companies. Because we're going after labor, this can actually be a generational business. Like think Constellation Software, Danaher Energy.

Kind of businesses, which also use some of the same industry transformation theories that were inspired by, uh, were also inspired by some of the multidimensional multidisciplinary team builds like the Levant goes off the world that, you know, Haman, who joined our board, kind of co founded back in the day from general capitalist, you mean from general catalyst and what we're trying to do is bring that creation ethos where, um, you know, we're, we're not sort of predetermining the outcome.

Okay. But in, in the reality that we live in, uh, back to your question earlier, it's got to be that harmonious combination, right? Exactly what that combination is might change based on what use case and what vertical and what point in time, but that that's what builds a generational business, right? If you only have a technology that you're trying to solve every problem as if it's a technology only problem.

Understood.

Andy Lee: Yeah. I think just, just one, one, one add on to that point. Cause I think people tend to over estimate the impact to labor. Like you think about all these consumer based direct to consumer app businesses, like whether it's an, it could be an Uber or DoorDash. All these apps were built with the mindset that there wouldn't be no humans in the loop.

The funny thing is, behind each one of these apps that are supposed to have these digitally seamless intera um, transactions, There are hundreds of thousands of people behind, you know, like, bridging the holes that, that exist, and there's no, technology is never perfect, and, uh, humans will always need to kind of help to stitch together.

Sriram Viswanathan: Which I completely, completely believe in and, and buy into. Look, gentlemen, I have one last question before we wrap this thing up. It's been a great conversation. The question I have is, you know, just, Put your forecast hat on. And what do you think this looks like for the industry at large? You know, the industry, we said, it's a, you know, hundreds of billions of dollars in, uh, in overall TAM in the traditional BPO, what do you think crescendo will represent, let's say in two or three years, what is your vision on how big crescendo is likely to be?

Anand Chandrasekaran: You know, I will say we see the outcome based billing model as really a new way of building businesses. For the last 25 years, that's all that enterprises wanted. And for 25 years, uh, you know, everything has been an approximation of delivering an outcome. And we sort of compare that with, uh, the transition from on prem software to SaaS.

Um, at that point in time, there were like tactical technology differentiators that SAS had over on prem systems, but over a period of time, everybody that was doing SAS had sort of a baseline, um, set of capabilities, but the early movers to SAS had a big advantage and they were able to get into new categories and so on.

And we think sort of outcome based, uh, deliveries is going to be just a model in the future. And we feel like we've just Uh, developed an early mover advantage in this space that we intend to take advantage of.

Sriram Viswanathan: Fantastic. Uh, gentlemen, this has been a great conversation. I learned a lot. Sounds like you're building not just a transformative business, you're using a transformative technology.

To really, uh, disrupt a known industry for the last 20, 25 years. Thank you very much for joining me in this conversation. Good luck to you both. And hopefully when you guys are making your next big billion dollar acquisition, you'll come back to this podcast and tell me how you did. Thanks Sriram. Thank you very much.

Great chatting with you. Nice to chat with you.

Thank you for tuning in to the Tech Surge podcast from Celesta Capital. If you enjoy this episode, feel free to share it, subscribe, or leave us a comment. A review on your favorite podcast platform. We'll be back every two weeks with more insights and discussions of all things deep tech. Bye for now.