Adam Parks (00:01.043)
Hello everybody, Adam Parks here with another episode of Applying AI. Here with my co-host, Mike Walsh. And today joining us is Karan, coming from EXL to talk to us more about, well, keeping your AI voice and written and communications from an Agentic AI standpoint on the right tracks, making sure that your compliant systems from a technology perspective have the backbones they need to keep you safe, be productive, and start actively being able to deploy these types of artificial intelligence tools.
So Karan, thank you so much for joining us today. We appreciate you coming on and sharing your insights. You know, I'm getting to meet you for the first time today. So could you tell everyone a little about yourself and how you got to the seat that you're in today?
Karan Sood (00:47.764)
Absolutely, Adam. First of all, nice to meet you. Glad to be here. I lead the AI solutions and product business at EXL. EXL is a data and AI-led company that is listed on the New York Stock Exchange, and we are 27 years young. And the reason for that is that the space is changing so fast that we keep reinventing our organization every few years. So we're gonna stay young.
As part of this, and as part of my role, I have a series of AI solutions that span different workflows. One of those workflows is collections, where we actually started focusing on this about 2018-2019 before generative AI was cool. Right? So this was based on the earlier machine learning models. And given we were already, I would say, a few steps ahead.
So when the AI came, rolling in 2022-23, we were the first one to adopt that and take it to the market and production as it is and deliver value. So that's one of the portfolios. Previous to this role, I was CEO of Godrej’s North America business, where I ran the country operations as well as a transformation for the global business, and that's how I made a switch and jumped deeper into AI in ‘23.
Adam Parks (02:08.763)
And very focused on the product side of the business now, which I think is very interesting. So now that we've got a little bit of understanding of your background and of the organization itself, talk to me a little about the solution, the main solution that you're providing to the collection space. And then I want to ask and talk a little bit about how you're keeping that on the rails. Like, how are we keeping this type of technology on the train tracks?
Karan Sood (02:33.526)
Excellent. So our solution is called PayMentor. And the core thesis behind this is very simple. The core thesis is that for any collections, you need to solve for two problems. You need to focus on customer engagement, which means reaching the customer at the right time of day and at the right frequency. Once you reach the customer, you need to persuade them to pay, right? Those are the two axes with which you fundamentally the entire collection's operations run.
So with our solution, we reimagine both of these axes, and we make them better. So we improve the customer engagement through our AI models, which are dynamic in nature and which at every interaction predict what's the next best channel at the time of the day, the right frequency to contact this customer. And unlike collections operations, which run this on a static basis, saying, Okay, contact everybody on day four, what we do is we iterate, we do it on a more much more dynamic basis by overlaying the risk and the behavior segment of the customer.
And the second thing we do really well is we leverage AI for persuasion. So this is where you know the human ingenuity comes in, right? You have to understand the context, and you have to be able to converse with the customer. You have to, in some cases, you have to empathize, in some cases, you to explain. In some cases, you have to explain to them what the consequences of non-payment are, right? Within the guardrails.
So with AI, with specifically generative AI, where the speech has become much more human-like, and our solution, you know, sometimes people cannot distinguish between whether it's human calling or AI calling. So with that technology, we've been able to significantly improve the persuasive power of collections. And combining the two, which is a better strategy, better outreach with AI-led persuasion, fundamentally delivers a higher collection. So that's what we've been doing.
We've been doing it for 20+ clients globally across industries, and it's delivering significant results.
Adam Parks (04:32.604)
Well, so many people in our audience equate artificial intelligence to their, you know, instance of ChatGPT, right? Like, that's, I think, a lot of how people are viewing it. And so they're pretty used to some crazy hallucinations. And I believe one of my lawyers said that ChatGPT is like a drunken frat boy. It's gonna be very wrong, but it's gonna be extremely confident while it's as wrong as it is, right? So I think that there are some challenges there, and with those hallucinations comes fear.
But when it comes to deploying these types of things at scale in the wild, it's not just about deploying one model that's doing voice, right? It's a series of tools that are being tied together. The term I've heard used before is "judge LLMs". One LLM is judging another LLM, which is judging another LLM, and different models are basically feeding different pieces of the equation to the whole math problem.
Talk to me a little about how you approach that as an organization from a product design perspective.
Karan Sood (05:36.526)
Absolutely. So most people are used to AI in voice, right? And that's what they refer to it. But we take a step back and say, you know, we also use AI in chat. So think of the common braid and think of either chat or voice as a channel of communication. So when we think about, and when we actually deploy AI, what is a given and what is a non-negotiable for us is that it has to not hallucinate.
Karan Sood (06:05.292)
And for that, we put a series of guardrails that catch the errors before they are actually communicated. So here are a few techniques for how we are doing it. In some cases, if you think about the technology, it can be fundamentally broken down into three components. Let's talk about voice. You have a component that converts what the customer talks into text. That's the first component of the technology. Typically, we use the best technology available.
This gives about 95% accurate word translation. Right. And about a year or year and a half back, this was about 85%. So this technology is gonna keep improving. So the input to the model for the next step is improving. So as data becomes better and cleaner, model output improves. So that's why. The second step in this journey is where the model comes into play. This is where the LLM, the brain of the solution, is. But we don't just pass this text into the LLM directly.
We actually have a series of prompt guardrails built into it that are paired with the customer's response before it is fed into the model. So the model automatically receives input through a series of built-in guardrails. Once the model does its magic, predicting the right answer in that context. And because these models have been trained by us on the collection-specific workflows, these models also generally produce better results than the standard off-the-shelf model.
So that's another thing we improve. So one, we bring in the guardrails at the time of the prompt. Two, we fine-tune the model so that the model is generally more accustomed to giving collections relevant and compliant answers. And then once the answer is generated, that's where we have another LLM as a judge to figure out whether the answer that is coming out before it gets communicated is meeting the right set of guardrails and the checks and balances we have put in place.
That then gets converted into what we call text-to-speech, where we use some of the leading players like ElevenLabs, which fundamentally converts that text into almost like a human-sounding voice. So, while the core heart of the component is LLM, and if you didn't have all these guardrails and prompt libraries built around it, you could have a hallucination that could completely run amok.
And as you said, it'll be very confidently wrong. But we put in place these prompt trees, guardrails within the LLM, and fine-tune the LLM to control that hallucination.
Adam Parks (08:30.098)
So you do it by breaking it down into three components and then guardrailing each one of those components as it goes through so that there's nothing lost in translation. Now, Mike, you've been out in the marketplace talking about this product and this process, right? Because compliance has been the number one thing. What kind of questions are you hearing directly from the marketplace as you try to explain these technical tool sets to a simpler audience?
Mike Walsh (09:09.492)
Yeah, well, an audience that has been kind of taught fear, right? So they're worried, right? Yeah. And that's kind of the frustrating thing is they do think of the drunken chat, frat boy, horrible experience. That's what they have. And I think I mean the only way to do it is to, like, hey, this is different.
The technology has even, you know, Karan and I started at EXL about two months apart. We have a totally new virtual agent. It is incredible, and it's way better than what we had before, which was really good and very productive across the globe, right? So I think when you have to dig in, you have to demo it, you have to show these different components and break it down where you know.
I think compliance is a gatekeeper in the beginning, but it's your best friend at the end of these onboardings because, hey, collectors have bad days, right? Like, I've seen really good collectors lose it. It is a hard job. I trained; that was my first day in this business, it was as a collector. It was miserable. I wanted I'm like, hey man, I'm gonna leave if I don't, you know, you don't get me off this floor. It's like, no, no, we're gonna switch you to a new type of client. So you know what that is.
But you realize it is such a tough job. And these people, meaning our customers, our friends, our neighbors, they don't want to talk to a collector, right? Like so, I think this tool gives them such a better option: private, clean, easy, at their own pace, at their own time. They have so much more personalization through AI. It seems crazy, but it's true.
Well, once you go through those steps and you see that this solution or any solution, you know, is there to help and make it better, more efficient, I think you get through it, but you have to go through some great questions. You know, how much PAI do you need? You know, all these different steps that, you know, we're used to. And I encourage people to ask those questions. Get it all out.
You like…there are no stupid questions. If you three years ago, I was brand new to this technology, and now, you know, I live it and breathe it every day. But I still do like, you know, do a lot of talks where we started stage one. And I think it's great that people are all, even I worry about this podcast once in a while, is are we too far ahead? Are we too far behind?
Like it changes so fast, as Karan said before, that people are in all different stages. So you get a ton of different questions. And I think people are learning fast. There's a lot of information out there.
Karan Sood (12:25.568)
I'll actually say something. You should select the partner not based on the technology. You should first select the partner based on their collections experience. See, because technology is gonna keep improving. Right? So you need partners who are flexible, who bring the best of breed to you. But technology is only as good as how you use it.
And technology in the hands of people who have never done collections or who don't understand this deeply is what results in hallucinations and all those nightmares, complaints, et cetera.
If you give it to the people who have done it, Mike has been here for 30 years, some of our team members have been here, but besides the point, if you give it to the teams who've been doing it for a while, who understand the pain points, who understand the process, and then bring the best of technology, that's where the magic happens.
Adam Parks (13:11.482)
Technology doesn't change the learning curve for the business itself. And so that experience of being able to build it out, and Mike and I have talked about this at a few different shows. If I told you how many voice AI companies reached out to me in the last 18 months looking to enter the US marketplace from all over the world, mostly I mean Africa, Europe, whatever. And every time I ask them two questions: " Who's your lawyer, who's your compliance?”
Because unless you don't even have an answer to those two questions, then there's nothing that I can do to help bring you to the US marketplace because you're a threat, not an asset to the industry, if you don't understand those compliance guardrails. The technology will continue to evolve. The tech will change rapidly.
Karan Sood (13:52.27)
Correct.
Mike Walsh (13:53.273)
And I think too, Adam, like, this is a tool for collections people, right? Like, part of what Karan said is like it's gotta be adapted to your procedures and policies, right? Like, it's yes, you know, that you wanna let the tech go, you don't wanna control it, right? Like you, but it's gotta say, Hey, this is our policy.
How do you do it for this client versus this client versus in this industry we serve? Utility is different than medical, right? Like, so it's gotta be adaptive to, and you know agencies, they'll have government, they'll have credit card, they'll have medical, right? All the same agency. It's gotta be able to do all those things. Otherwise, you're just buying a headache, right? So make sure it's a yes.
Adam Parks (14:36.86)
Well, and all those consumers are different, right? The balance ranges are different, the intent is different, the approach is different, and for each one of those, you have to have something that's flexible enough to be able to accomplish these things. But it sounds from the other conversations we've been having, it sounds like a big chunk of the change. And you've mentioned that you're on a totally different agent than you were when you first started three years ago. And now you guys are onto something new.
One of the shifts that I've seen happening, and I think you guys are kind of leading the charge with it, is that push to intent-driven. The old agents were reacting and responding to things, but the new agents are trying to understand the intent behind the statement, which is where I think there was the biggest differentiator for the human versus the bot communications. But the more that we can understand the intent of what the consumer is trying to say, the more value we can ultimately provide with these tools.
And that intent, I would expect, changes culturally, whether that be regional or city-based. I mean, that's got to be a giant spider web in and of itself as to how to predict that intent. And I'll use a silly example, but coke, soda, pop, right? Like they're gonna use different language for it in all different areas just of the United States itself or even within an individual area.
What does that start to look like as you're deploying tools like this on a global scale? How deep does that web go? And how different has the business become as you've started to better understand the intent of the speaker?
Karan Sood (16:21.474)
The beauty of these language models, specifically the ones we are coming out with, they have been trained on tunnel data. I can't even- you can't even imagine the amount of data they've been trained on. So one thing that these models do really well is that they understand the nuances of language. That's fundamentally what has been improving on each of these models. Obviously, the reasoning is improving. That is basically like the way they write the algorithm, which is improving to make the model reason more.
So I think that's one thing. As models improve, the context understanding of the conversation or the context understanding of what the customer is saying is becoming better. But that's just half the story because there are still gonna be nuances that the model will not understand, right?
And that's where our collection experience comes in because we take the intent classification and then you bring in our expertise of that classification and then do the reward or the answer, the answer to the customer.
So I think that's where the partner you choose has to be experienced enough to understand the nuances of intent. And I think, and actually I'll tell you one step further, if anybody misclassifies the intent, the entire AI journey goes off the track. So I would actually argue that's the single most important point of differentiation in how accurately you can classify the intent of the customer.
Mike Walsh (17:41.977)
Because it doesn't matter how good your message is if it's the wrong message, right? Like we've all been on the call, and you're like, I'm not asking for that. You know, like so, especially in debt collections where you have someone who's not thrilled at the process, right? Like it's a customer who's got a problem. I think it's hypercritical that the understanding of the AI is done.
That's where collection experience is huge and quality of tech, those two things have to marry to make it really productive for you.
Adam Parks (18:24.1)
I started thinking about when we're looking at this intent, you know, that's where so much confusion can exist and live, is in a misunderstanding of intent. And we can't afford to have anything going off the rails, being misunderstood, or being unable to bring it back. Now, when we talk about stacking these models on top of models, is intent its own model, or is that handled as pieces of the other models? But as you've gone through that evolution, what does that look like from a tech perspective?
Karan Sood (19:01.346)
Best intent classification models are machine learning models, right? Because you don't want a hallucination at the time of intent. So when the speech gets converted to text, that's where you have a specific intent classification model, which are pretty good, pretty accurate these days, which fundamentally starts segregating the journeys.
But here is the other bit. Best of the models will also make mistakes, right? The problem happens is when AI is not able to track back and go to the right intent. That's where the next problem comes in. So the first run comes in with the wrong intent classification, right? Which, you know, customers we have seen are still okay if the AI walks back and talks the right language, immediately admitting the mistake.
So I think that's where the second thing comes in, where your AI bots need to be flexible enough to go back to the starting point of the journey and then reroute it to the right intent and then trigger the journey again. So I think those are the two bits that have to go hand in hand. But I think the intent classification models are becoming very, very accurate.
Mike Walsh (20:04.854)
I think that's the drastic improvement, too, right? That flexibility currently and the scalability. Like if it used to be, you know, maybe two tries and send it to a live agent. Now it can be that second try, if it eliminated the first mistake or the first utterance where it didn't get it right.
It might have them rephrase and then boom, I'm sorry, and then take that on the right track. I mean, if you think about how many times a human being even does that wrong, a collector, right? Like they think you wanted this or said that, that's always gonna happen. And it could be back to your pop versus soda versus sub versus hero versus, you know, grinder. You know, it could be a collector in Massachusetts versus, you know, someone in rural North Carolina.
And it's just a different version than they come to an understanding and then they go back down the track. I think that to me is the biggest difference in the last three years is. It used to be boom boom, very robotic in its understanding, and now it's way more flexible.
Karan Sood (21:17.654)
And that also actually, you know, if you think about how customers get frustrated? They get frustrated when the bot doesn't understand what they say. That's the core source of frustration. They say, you know, connect me to an agent. And that is actually improving because bots are getting smarter. They can reclassify the intent; they can trigger the right answers. I think that's how it is. Evolving is gonna keep getting better.
Adam Parks (21:40.218)
And retrace their steps back to the beginning. So if there's a misunderstanding from an intent perspective, the bot can go back to the beginning of that call; it can rewind, relearn, and try to determine with additional context. Whereas even as a human, I mean, you might be able to remember the call that you're on, but you're not reciting it again a second time, right? Like, it's just a different animal altogether with that level of context to that conversation and understanding that intent and where that consumer is trying to go.
Now, when we think about these models continuing to improve, and as you've mentioned that you've gone from one agent to another, how much of the process is moving from, let's call it, one agent to another versus supercharging one of the existing agents?
So like, clearly there was a versioning breakpoint between those two models where they were like, Yeah, this one's as good as it's gonna be. Now we're gonna change whatever it is core infrastructure or whatever the version change catalyst was. But what does that look like? How much does the model or that agent expand before it's necessary to move to the next one, and have those timelines been shortened at all?
Karan Sood (22:59.628)
We've actually designed it slightly differently. We have designed it as a multi-agent orchestration for our voice bots. What that means is we don't have a single agent that is doing multiple things. Because what we have learned is that one agent with multiple goals will make mistakes more often than multiple sub-agents, which are dedicated to a goal.
Because at the end of the day, the agents are a goal-seeking piece of code, right? If you muddy it by adding too many goals to that agent, it is going to make a mistake, and it is going to hallucinate. So that's one way we solve it. So we almost have an orchestrator agent at the top, and we have specific agents for each specific goal, which then intelligently get routed to, and there is obviously a connection between the agents to hand over and keep the journey rolling properly.
So that's how we design it. And as models improve, we fundamentally change one component of these agents rather than having to rebuild these agents. So these are built in a modular manner that, okay, if out of the let's say the tools improved or the memory has to be improved or the model has to be improved, we can take that piece out, and we can bring in the new one. So that's how we've architected this solution so that we don't have to, you know, boil the ocean every time a better model or a a or a better tool orchestration comes out.
Adam Parks (24:25.583)
Makes sense, modular management, right? Be able to place the carburetor, not the whole car. That, I think, makes a lot of sense from that perspective. But what does that learning process look like? How are you constantly evaluating those models to determine what the next best thing is? Because the new thing, the shiny thing, is not always the best solution. So how do you look at that learning process?
Karan Sood (24:50.939)
See, you have to always balance three things when it comes to bots. You have to balance the latency. You cannot have a conversation that doesn't feel real-time, right? So it'll spoil the experience. You obviously cannot get the accuracy wrong; you cannot hallucinate. At the same time, you can also not incur so much cost that you know it just becomes unprofitable to do this, right? So it's always the balancing of this trifecta of accuracy, latency, and cost that we have to manage.
And there's also, you know, while the world is probably, I would say, accelerating at a much faster pace than ever before in terms of the new things coming out. But we also have a philosophy that if it's not broken, don't fix it. Right? We don't need to use the best-of-the-breed model today because that's just overkill, which will just give you more cost. And which is probably not the best thing. So what we do is, we obviously have our RD team, which constantly monitors
What is the best model coming out? What is the best suited? What are the best use cases for these models? So we obviously have our RD team, given we are part of a larger setup, which keeps an eye out for the net new. But our production teams always focus on the outcomes. And if the outcomes are on track, they're delivering the right outcomes. We only change something if it misses the outcome we are looking for.
So it's all towards delivering the best outcome rather than worrying about whether I need to upgrade each component periodically.
Mike Walsh (26:16.46)
I mean, it comes down to collecting performance, right? Like, that's also like, yes, the tech is doing its job, but what is the job? The job is to collect money in a non-compliant environment, right? Like in a customer-friendly environment. And if you're doing that and it's improving, that's great. If it's missing a segment or something like that, it's when you're looking, you start tweaking.
If the tweaks don't work, then you say, Okay, what's wrong with this thing? And then you kick it up and that R&D team's like, we can solve that, right? Like so, it's not just the check that comes into play, but also the strategy of collections. You know, you can have a great model, but it's not designed for collections. What is it gonna do? Right?
So, we do update our strategy quite often, though, just to build on Mike's point, because while execution, if it's not broken, don't fix it. Strategy, however, has to evolve because I have to collect more today than I collected yesterday. So, which means our strategy and our intelligence modeling team are always on their toes to make sure that the next version of the strategy is getting released periodically to improve the performance.
Adam Parks (27:22.725)
Well, and then you've also mentioned that each one of the tools that you're deploying is customized for that particular deployment, for that organization, for that product type. What does the process of learning or teaching the model actually look like, and what's required to optimize it?
Karan Sood (27:43.896)
When we started entering into new markets and new industries, it used to take us, I would say, a few months to learn. But given now we have done across, I would say, like four or five industries, about six or seven geographies, we have started to see patterns emerge, which gets us started much faster. I think we've been able to compress our timeline to about a month. Actually, once we go live within 10 days, we can broadly tell you with 90% confidence which customer has a preference for what channel in about ten days of going live.
Adam Parks (28:14.353)
Wow. Okay. And so that's, and Mike, we've talked about what data is actually needed. It's not necessarily the PII here either. So talk to me a little about what does an agent or an agency need to provide to you in order to have a successful learning model.
Mike Walsh (28:33.726)
Agencies are, like, they're kind of tricky, right? Because they usually go across multiple industries and multiple products. They have multiple clients with different settlement payment plan requirements. So, really, you know, getting that set up is part of the process. But usually they've already done that with their payment processor, nine times out of ten, right? Like, they have someone collecting digital payments for them and there are rules for each client. We just copied those, right? Like that's a quick cheat. You get that up.
So you need your guardrails; you need your collection process. There's slight verbiage; people are comfortable with different verbiage, like a little bit of compliance, you know. And then it's really you set up these parameters and then it's connecting, connecting to get that data back and forth, and there you go. It's much easier than people think.
Adam Parks (29:33.637)
And but like, what are they sending over for that understanding? It's just account-level information. Is there any behavioral analytics being collected? You know, what do you ask for from that agency to be able to actually execute on that? Right? Like, I know that you can copy the rules from whatever other digital channel that they have, but if you're gonna be doing this reinforced learning, it's just that simple. It's seven fields, and you guys are off to the races. Like, what's it really looking like?
Karan Sood (30:10.69)
What we generally need is just a file about the person, the contact information, and the loan history, which are typically the three fields we need on a daily basis. We will learn about your policies as a one-time exercise, unless they change, in which case we can obviously update them.
So those are the two major updates, and there is a reliance on the compliance team to just sign off on the templates we already have built so that you know your organization is comfortable with what we are communicating.
That's pretty much it. Actually, that's our only reliance. All the behavioral signals we measure on our own because that's our solution, I would say, differentiating factor because we understand how the customer is reacting to our outreach. And that's what the signal is that goes into our models to decide the next best action.
Mike Walsh (30:57.836)
Yeah, like consider a model suite that's pre-built for collections, right? It's not like it's gotta learn everything from ABC agency to get started. It says this is agency second placement, you know, a twenty-four-month-old or, let's say, a twelve-month-old after charge-off. That's enough to prep the model.
And then it's gonna do a lot of rapid experimentation and see what's working, what's engagement, what time, all those different good stuff to get people to, you know, you're trying to start a right party contact through digital means, really, or a virtual agent. And that's really what you're trying to do. Get the person engaged. The advantage is that it can be 247 with the tech, right? It can be at the customer's convenience.
Adam Parks (31:58.31)
So you already know what patterns and correlations to look for within the data set. And now it's just about putting a new data set in there and identifying those same patterns and correlations because the behavioral patterns across portfolios are going to be, at the very least, similar. And so that makes a lot of sense. How does that start to work as we start deploying with these organizations?
Does the learning st–we do the one-time learning, we've got the behavioral analytics. Is there additional learning that's happening around that? Or is that kind of the focal point of the compliance concentration?
Mike Walsh (32:43.8)
Well, it's like you're recalibrating to each portfolio, right? And/or sub-portfolio if it's an agency. And there is a lot of learning going on, right? So there's reinforcement learning going on. So your behavior or your lack of behavior is gonna teach the machine about that specific portfolio or that specific agency's paper. I always say it's just calibrating, like, it works, it's proven, it's been working for years. Now, how do we fine-tune it to make it work specifically and learn?
So you're using, basically, the models that are there that know collections, then you're doing reinforcement learning. So they adjust on the fly to adapt to that specific portfolio is the best way. I say it simply. Karan can probably give a more technical explanation.
Karan Sood (33:58.631)
So think of us as partners who will bring on day one a strategy for your portfolio, and we'll say, know what, from our experience looking at your portfolio today, without any further information, we believe this is how the segmentation of day one should look like. And in the next two weeks, we will do A-B testing to start rearranging people into different buckets based on our models.
And that's how, so then, we will bring a day-one strategy, which is our learning across countries and markets and is suited to your portfolio. And then, based on your data, we will continuously improve it. So then there'll be a version two of strategy, a version three of strategy, a version four of strategy, all of which will be fine-tuned to your customers' intent and how they interact with us. So that's how we iteratively improve it. And that's how we
That's what we call it, the reinforcement model, because in the back, our models are regularly doing A-B testing. So for example, if our strategy says to send this code of customers three SMSs a week, there will also be a challenger within our own strategy, which will say, okay, to 5% send two times, to 5% send four times, and see how they react better. So that's the constant A/B testing we have built in as a reinforcement loop, continuously improving the strategy and decisions and leading to the next version of our strategy.
Adam Parks (35:27.887)
And so looking into your crystal ball, what happens over the next five years?
Karan Sood (35:33.436)
Man, the way the technology has changed.
Adam Parks (35:34.609)
I went with five years on purpose because I, you know, in five years, flying cars are possible in five years, right? Like, the whole world opens up. But I'm curious because I'm just for someone who spends so much time engaged in these tool sets, where do you see that technology evolving over, let's call it two years, five years?
Karan Sood (35:44.595)
Yeah. I think the models will continue to get better and smarter, right? Just the comprehension, the intent classification, the ability to answer back to the customer, that's just going to improve, which means the number of calls that can be fully contained by AI is going to increase. That's one. Second thing which will happen is all the associated pieces around the models, the speech-to-text, the text-to-speech, all of that is also going to improve.
So fundamentally, the number of calls you can directly contain with AI is going to increase as the technology improves. But I think the second axis, which gets ignored, is the human behavior. People also, in three to five years' time, are going to get super comfortable talking to AI. You know, I'm just throwing it out. There could be a cohort of customers who will build their own personas for their personal agents, who will actually talk to the other agents.
That is also not out of the realm of possibility, right? But I do know that the average population is going to be much more comfortable talking to AI two years from now than they are today. And that curve will only improve and probably flatline in about five to seven years when enough people in the population are comfortable talking to AI. So I think those two trends will converge, and that's what will drive this adoption massively.
Adam Parks (37:10.768)
Well, we saw consumers' behavior change over the past couple of years as they've become more comfortable with subscriptions. And that modification took about ten years, from going into a blockbuster to just paying Netflix on a monthly basis, right? And accept even your car's got subscriptions now, which is ridiculous. But I think you're right. I think that consumer changes.
Now the bot-to-bot conversation is a whole other ball game because that's already starting. Consumer bots are already emerging. The debt settlement companies are creating their bots to reach out and do those negotiations, and what's that gonna start to look like. I can only imagine that there will be more services in the coming twenty-four months that will be selling that service to consumers directly. Like “Don't talk to collectors, we'll handle it for you,” or “our bot will handle it for you,” and then the bots are talking to the bots– in the Terminator, right?
Like this is where it all goes off the rails. But Yeah. It's like an alternate reality. but I think it's interesting as we start looking at that consumer behavior, because if we're collecting, right, everything that we do is driven by how the consumer is going to react, how they are going to behave? What's the next action that they're going to be willing to take? And as we look at all of those pieces coming together and we give it a little bit more time.
Adam Parks (38:30.788)
I mean, the acceleration of use case of what we'll call the retail LLMs, the ChatGPTs, the Claude, Perplexity, whatever your favorite flavor of ice cream is, is irrelevant. But as people get more comfortable with that, the bot conversations start to improve, and everything goes down the line. I remember the first real conversation I had with, like, a ChatGPT just trying to learn was when my daughter was first born. I had no hands because I was feeding the baby. So now I'm just talking to the models and trying to learn.
But from that came some really interesting article ideas, like the eight personas of consumers and debt collection, things of that nature, which kind of freed up that time to think. You know, I'm curious to get your thoughts on this one. And I'm sure you've seen the Anthropic Report back from March. But do you, as we think about the consumers, do you think that we're going to see that AI will have a direct impact on the collectability of consumer accounts over the next five years or so?
Mike Walsh (39:34.135)
I think it will, right? Like you know, debt has been around for a long, long time. And really, you know, to me it's almost like it's gonna come down to the job market, right? That's gonna be its always number one, right? If AI takes too many jobs, then it won't be as collectible, right? Like i you know
This will be looked at more, as I have to use AI because it's cheaper. You also have the stress that, to save jobs, the government's looking to keep US jobs, and US call center jobs in the US, right? Like so, I think there's gonna. I think the market's changing for sure. I think consumers are- I think it's overblown that the whole AI's gonna take everybody's job.
I think it's gonna create so much growth, you know, and you know, I know a lot more. I had a conversation last night at the baseball game. My son's decided because of AI, he's gonna become a plumber. And I'm like, well, he's gonna make a ton of money in Charlotte. Like, all we have, you know, people just keep building here, right?
Like, so he'll be busy. So that's great. Like, so maybe people look at different jobs differently, you know, and different jobs like are gonna be more in demand than they were, or I think get more prestige than they used to have, right? Like starting your own business. But to me, there's gonna be correct…
Adam Parks (41:08.6)
Each job becomes more important. Right? As we accentuate our capabilities by augmenting with artificial intelligence, each individual job has to be more important. And in that anthropic report from March, they talked about how there's been no observable reduction in the labor force. The only thing that has happened is that new jobs at the entry level that used to be created are not necessarily being created.
But nobody's being replaced. It's too big a risk, and it doesn't add enough value. There are a few reports out there that talk about the augmentation of these tools, right? Using copilots and other things to enhance the capabilities of an agent so that they can better handle those exceptions. And as the models get better, the exceptions get smaller, and, you know, fewer humans are needed for it, but that gives organizations the capacity to go out there and sell, right. Now they've got additional capacity for their organization.
Mike Walsh (42:07.402)
And I wonder if those new job creations are just because it's moving so fast. They don't know where to land that job, right? Like, they don't know how, like the job's gonna change and they know it, right? So do we do we launch it now? What skill set do we need? I think this is a transition period that, yeah, some of these, like, even think of a collection manager at an agency, right? Or operations manager. They're gonna have to know some data analytics or find a tool they can understand that teaches, gives them what they want to know quickly, right? Like so.
As the roles change, I think we're adjusting to that now. And and I think it's I think they're gonna be there. It's just that we have to figure out exactly what they are and exactly what we're looking for as organizations because they're gonna change. And then people are gonna find great jobs that they love and are cool, and they learn stuff, and they're gonna be involved every day. And just think of like what you've been doing with AI, right? Like I'm gonna plug Adam here, like how many
You know, the index is coming out. How much, just because you had time to do things like you do with a baby and do all these things that you created different roles in your company because of that, right? Like, so I think it's changing. I think this is a transition period. And I think that was a really good article. You know, I know you shared it with me, but I think those entry-level jobs are just gonna change, and we just have to make sure we're training our workforce for them, or we're gonna have to do it ourselves, right?
Karan Sood (43:54.956)
I think that the workforce that is now entering the labor market is just much, much AI-savvy than we are even today, right? So as the work of the future evolves, more companies will need people who are comfortable with AI and can work with and on AI, right? Actually, I would say that in the world, about 90% of the work is going to be people who are comfortable working with AI.
It's only the 10% of people who actually work on AI, would say, maybe less. So I think that's the shift that will happen pretty rapidly: you will need, fundamentally, a skill set that you know how to work with AI. And I think as people learn through that, I think the new jobs will emerge. There will be, I do believe that there will be some industries, some areas specifically, which might be more impacted than others.
Software engineering for example, right? That software engineering has just in span of three years has gone from the hottest, know, hottest market to, you know, the most stagnated market. But even today, there are more software engineers than there were four years back. So the pace of growth obviously has tapered, right? It will create more issues later because, you know, if you've never coded and something goes wrong, I think you have to then diagnose it.
So I think there will be some learning curve over the next three to five years, but I think 10 years out, I think it's going to be like the internet like people know how to use it, like people know how to get to the internet. I think AI is going to become like the internet, where without it, you're not going to function.
Adam Parks (45:34.096)
I think that's a pretty fair statement. And gentlemen, I can't thank you enough because every time I have a conversation with somebody from the EXL team, I walk out of it thinking about my future, the future of the industry and really the direction that all of this is going, because we're there's no way to avoid it. And those organizations that are not deploying the tool sets are going to find themselves behind the eight ball. And, you know, even though the cost of computing power is is getting less expensive.
Being a first mover allows you to get out in front of challenges and go through the organizational learning processes that you can't shortcut. No matter how cheap the production power becomes or the computing power becomes, we're never going to get to a point where you can just magically learn as an organization how to deploy these tools and the nuances of making them successful within your organization.
Karan Sood (46:27.242)
Very well said, Adam. The learning curve can only be learned by actually doing it. So I completely agree.
Adam Parks (46:34.98)
Fair. So for those of you who are watching, if you have additional questions you'd like to ask Karan, Mike, or me, you can leave those in the comments on LinkedIn and YouTube and we'll be responding to those. Or if you have additional topics you'd like to see us discuss, leave those in the comments below as well. And hopefully I can get Karan back here at least one more time to help us continue to create great content for a great industry. But until next time, gentlemen, thank you so much for your insights today. I really do appreciate you.
Mike Walsh (46:56.845)
Thank you, sir.
Karan Sood (46:57.117)
Thank you. It's wonderful to be here. I'm looking forward to coming back.
Adam Parks (47:00.558)
I look forward to it as well. And for those of you that are watching, we appreciate your time and attention today. We'll see y'all again soon. Bye everyone.
