Digital collections done right: hard truths & untapped opportunities

[MUSIC PLAYING] Welcome to the latest episode of Better Dead. I'm your host, Josh Forman, CEO and founder of InDeadin. And every month, we bring together speakers from across the global credit and collections industry to share insights and new perspectives shaping the future of debt collection. Welcome to another episode of Better Dead. I'm Josh Forman, founder and CEO of InDeadin. I'm joined today by Kristen Leffler. She went up to officer in InDeadin and molten CEO of tech advisors. Kristen, start. Welcome. And I'll give you both a journey to introduce yourselves. Just after some of the housekeeping for today. For those joining in live, a couple of quick things. If you have any questions, you want to post to the group. If you just go to the Q&A section, your right hands are there. Feel free to jump in questions, and we've got someone helping to moderate them as we move forward. If you're having any issues from an audio video perspective, you can switch into light mode. And if you do have to drop off at any point, that's not an issue. Everything will be recorded and sent around the event. And then just as far as a bit of an agenda and what we'll be agreeing on today. So intro and welcome, as I mentioned, from the speakers. Then really jumping into the data challenges exists between the creditors and reflect back and forth, and token management points that are hired in and so on. Or those that may have died into common mistakes and how to avoid them, the future opportunities that we see still outstanding in the digital collection space, and then again, set a leverage in that Q&A piece and jumping across from there. But I'll quickly stop sharing how Kristen, do you want to introduce yourself first? Then we'll jump in. Sure. Chris and Leffler. I'm the chief product officer here at Indenin. I've been in this role since there's a few months. It's November. Prior to that, I spent 12 and 1/2 years at a large US-based debt buyer building at digital strategy there, as well as inventory segmentation, and really kind of the value of data in the collections industry. So I think I'm really excited today to talk about some of that, both on the digital transformation side and that the importance of data and interoperability. So I think Scott's a great partner to talk about that. What's today? [INAUDIBLE] Great. Scott Hamilton just founded, co-founded, arm tech advisors. But prior to that, a couple of decades with the larger banks, Bank of America, Capital One, Chase, then spent a couple of years consulting, a couple of years at Prodigal. And now, arm tech advisors is, I guess, a bit of an advisory firm, but understanding where are you, where are you trying to go, and then how we can help you get there a little faster and safer. So glad to be here. Thank you, Scott. We'll probably back to you straight away. Scott on this phone. So just jumping in, as I said, I mentioned on the agenda side, sort of that the data exchange and sort of the data challenges between that sort of piece on the creditor and collect a handoff. I'm keen to get your view from your perspective on the most common challenges in that area. And just to take us from that. Yeah, I think in terms of the handoff, one, the obvious data handoff component is the breadth of the data and understanding what it actually even is. Sometimes as simple as a balance may not actually be the balance that you think it is. And then you multiply that by hundreds of other data attributes. So inconsistent or inaccurate data mapping or lineage, the other bucket I'd throw out is either inconsistent or really conservative sort of compliance limitations around consent or flexibility around what you can do or what you can say, how much you can customize it. Those are really two pretty difficult handoff challenges or partnering challenges. And from some of your, you know, your past experience is that do you find that those difficulties vary by the type of creditor, the type of client? For sure, those that are, you have a long relationship with, you've had time to work through, you know, all those questions around mapping and what a data field may result in or on the other hand, to build trust that you do have the right controls from a digital standpoint. So they do start to loosen it up. But it isn't so much a large versus small as it is a trust versus we're not there yet. Those are probably the two headlines. - I think that's, I think that's great perspective. The other thing that I think has huge potential in the creditor collector, like handoff, is truly like a go by letter, a handoff letter. That has always been in my past superimpactable because you think about what the customer is experiencing in this moment. Hey, the customer doesn't know who XYZ collection agency are. Sometimes they don't even know the charge off creditor as they only know what the merchant name is. So the ability for a credible trust and source. - Often, have you seen that not happen or the data mapping not actually be accurate? Is, from your perspective, how big of a problem are some of those handoff realities, I guess? - On the handoff side, I think it's pretty significant especially if you're working downstream paper. So handoff directly from like a creditor to a dip buyer, that has gotten much better over the years. But handoff from a creditor to an agency, I think is still very, very poor. Especially as that account moves around through the lifecycle from seconds to trash, you're almost never getting a handoff at that stage. And Kristen, just elaborating on that a little bit more from like a product development perspective, or just from a product perspective, how do you see it addressing those friction points? And then sort of that maybe it'd be good for both most of all three of us to get a little bit deep dive into that sort of data hygiene piece. - Yeah, I think like if I could put like magic wand and say like how it would be the best way for it to work, would be for say I'm a customer, I logged into my bank and in my bank app, it says, "Oh, actually, this account is now being worked "by EditSlide busy, click here, "and it directs you right to x, y, v, a website, "for it all, whatever the case might be, right? "That's a super speedless bread crème transition "from a trust in source." And that's product work, right? That's integrating across the banks, that's the APIs, the data stack. Now, I don't think for most banks and I'll think they have that app in sight. I've heard some rumors that some may be having at the site like that or perhaps some of the fintechs to the BNPLs might be more amenable to a solution like that because it is so customer-facing and customer-friendly, but that is not the reality for most of the collection. Most collections are still flat-fat cheese first back in source. Maybe since the SD's back in fork with fields that say, "Bow," and you're like, "Ooh, what's that?" Where you can use potentially AI or really good ETL tools to try to streamline and standardize the data feeds that are coming from your various clients and your various creditors to try to provide the best customer experience you can provide in your ecosystem. - Yep, and question for you to your Scott, but we think about how important that data piece is in terms of actually ensuring the good outcomes on the collection. So do we have accurate contact data? Are we collecting on the right balance? Potentially the ability to sort of connect that account lineage? Where do you think the resistance comes from? Is it purely technology and capacity constraints within the organizations? Or do you think there's also an educational piece around just how much of the collection performance actually does come from the data quality, not just purely the collection strategies and activities that are happening within the agencies? - I would say it isn't, they have their data map. They know what "Bow" means, right? And it just so happens on the recipe inside, you have 30 different data feeds that are all different. And they all say, "Bow" or "Bow" or "ArrangeBow" or you're like, "SeeBow," and you really have to sit down and map out your version and their version and make sure they reconcile, then copy it over to 30 other clients somehow. So the ETLs mechanically are fairly, well, kind of straightforward, but misunderstanding. And oftentimes the giver doesn't really even know what the lineage is either. And that a lot of those data feed go into many modern validation and customer conversations, result in dispute, it can get really messy, for sure. - Yeah, I think I would echo Scott's point about, if there is some technology constraint, there are also people constraints. And I think it's less so about the actual technology itself and more about resource allocation. The folks who are dealing with a collections arm of large creditors tend not to get the resources that they ask for. They're forced to do things in the most manual ways, which then creates these problems down the line for folks like us. I don't think you could say that. Or any sort of lack of abilities. It's just lack of resources. - Yeah, absolutely. And double clicking on something, you said they've got around sort of the business context piece. And I think Kristen, you and I have had numerous conversations about this as we interpret files coming across and making sure engineers and others understand what that means and where it goes. But like given the level of complexity, 'cause you also mentioned Scott the end there, completely the things like model validation, notices disputes and other things. Like, how do you approach tackling that business context understanding? I think it's easy within the group that I'm speaking to here, you guys both know it really well and that's what you do all day. But when you think about the actual team members that are typically interfacing with that, you've got data engineers, you've got business analysis that they may or may not have that collections background. So you just elaborate a little bit on there on sort of sharing that business understanding context to make sure that those things map correctly so that the data is used while that comes across those details. - What I have seen used before is back to Kristen's bow, for example, is that there are five or six different possible meanings behind that field header and to list out, these are the five choices and be very conscious that you're choosing option three because you had that validated on the client side. And then because it's definition three, this is how we're gonna use it as opposed and just making it rather than guessing, like having a pre-set definition choice and just making it explicit by field type. I've seen that used. The givers often don't know what is meant, the receivers can misunderstand it and you just need to, some sort of mapping mechanism to validate on both sides. What it was meant to be given, what's then received to be actioned, a key. - Yeah, I think I'll highlight what you just said there, Scott was on both ends. So I think a place where agencies can fall down a little bit is not wanting to interact with the creditor quite as much and saying like, "Oh, the creditor gave me this file." They must know what was in there. I'm just gonna take it and map it the way I think it should be mapped rather than having that dialogue. Does it take more time up front to get it mapped correctly the first time around? Yes, does it make the downstream process into your point, Scott, reduced disputes, reduced frustration? Like, yes, all those. And so it ultimately improves collections, right? So the investment up front is very important and the dialogue. - And the last part, before we just sort of jump on to some of the next pieces, but we have, the thing we've talked a lot about sort of typically focusing on large financial services, creditors and we obviously make a lot of references to U.S. related stuff. And I think probably two pieces here, maybe Scott, you can take the sort of vertical side and maybe Kristen, I know it's only been a few months, but take the regionality side to it. But like, how do you see the similarities and or differences Scott on the sort of vertical piece in terms of, you know, a large bank versus a large telco versus, you know, a debt buyer and others and another change and one of the nuances, something, Kristen, to you on sort of how that maybe deals with as you think about different markets around the world. Honestly, I've been surprised that I thought larger firms be it banks or big card companies versus small fintechs would have the rules paved out, clarified a data dictionary, the lineage, all documented. But every time I think that, you know, the bigger firms have it, you should have it or do have it all mapped out and validated. I'm sort of proven wrong. So no offense to any, but and then on the good news, some smaller firms have fantastic processes and double and triple check in partnership with their clients at the smaller end of the spectrum. So I don't necessarily, those that are experienced in a given partner and have already worked through this, I think is the rule that they've figured it out and maybe learn together a little bit, maybe less so bigger is better and I don't think more mature just because they have scale. - I think to go to your point about the regional innovation, Josh, like it was this, it was a very eye opening experience for me to see just how different some of the different markets are, not only in the data that's captured, but also the expectation on how you use that data, right? Like being in the US market were so conditioned to model validation, here's how the concept works and here's what seven and seven means. And while those technically unnecessarily data elements, you use your data to drive that decision making and that can be very different regionally. Hey, what are the requirements for contactability? What are the requirements for affordability? What information do I need from the consumer? Where do I need to put it in my system and then transmit it back to the creditor to tell them we've done this affordability assessment? That is not something that I had any experience in and it's kind of a whole new world. Oh, he said I didn't know, I didn't know. So I definitely didn't appreciate the US as the hardest market, but I definitely didn't appreciate the differences between markets and what data you need to capture. So in a company like ours, where we have to serve us all those markets, well now actually we need fields for these different markets. And can you use the same field? Like can the bowel be the same? The bowel that we have at undeaded, can it be the same bowel from all these different clients, from all these different markets? So do we actually need to separate them 'cause the way that we use them is different, the way that we decision on the front. I think that is very interesting. Yeah, absolutely. Awesome, I appreciate that. And so jumping on to the second piece, we were to begin to get into, it's got obviously your new organization in terms of our tech advisors and you've been in the industry for a long time. You've seen a lot of things. So we've tried this the horror story, it doesn't necessarily have to be a horror story, but if you can just sort of give your insights into sort of common pitfalls and things that you've observed in these sort of large digital transformation projects and sure that'll often all bunch of other questions. Yeah, that's a wide open. I think the biggest, I don't know if it's a horror story, but the biggest headwind that most firms face is either some fear or hesitancy or a compliance partner that really just boxes them in, either out of choice or out of requirements or maybe it's the client and they really end up with a pretty bland, watered down weapon to engage customers, email, SMS, chat. It's just, it's got so many guardrails on it that it's really doesn't have much chance of being wildly successful and they get it off the ground and it's not wildly successful. It's solid and so the results aren't fantastic and so additional investments or additional creativity around compliance or something isn't a place that people are running to because the original test wasn't super successful. So that puts sort of one strike right out of the gate if you aren't sort of leaning forward a little bit. That can work against you. That's one bucket. The other bucket that I'll share is most of these within creditors or agencies even are kind of side of desk. If you don't have scale to dedicate the resources, they're generally run pretty simply on the side of somebody's desk and again, you end up designing or executing or more importantly not learning and testing and getting better with the velocity that you really need to. So either you start slow and results sort of our sideways or you start and you turn it on but other things come up and you really can't iterate and learn nearly as fast as a lot of your peers are and again, your results tend to lag and not get double the investment that they probably should if it was sort of wildly successful. Those would be my two. - I think baked into people's organizations, assumptions on what the general could do, say digital specifically could do. I think they talk to industry, other industry peers who say like, oh, we were able to double our investment and they assume that that is just a blanket, not appreciating, hey, actually the reason you were able to double your investments because you built up this infrastructure, your investment was very big and you were able to double it versus, okay, I can double my investment, I'm gonna make a small investment and double it and that doesn't actually scale that way, right? - I forget where you had brought up that the, you almost have to start backwards. Your portal has to be fantastic. Then you can get to messaging 'cause if it goes the other way, the conversion rate won't be there. It really, it's hard to do it in piecemeal. To your point, you really have to understand the entire customer journey and make sure there are too many roadblocks in that experience. And that's hard to do when it's 10% of your job and you're running a call center at the same time. - Right, right. I think that can back to kind of one of my soapbox things that I always talk about, right? Which is you have to be able to measure what you're doing. So, if you have this customer journey, how do you measure every step of that customer journey so that you know if you make a tweak, it was impactful as you make a tweak. It was impactful. I think that comes back to your test and learn. Thanks a lot of folks are testing on the side of the desk, disconnected single channel processes, but they're not learning because you're not capturing the level and the granularity of information that you need to be able to learn from those insights. - I think I'm comment I was just gonna make and sort of I guess re-like here as well from Jim's put in the chat too. But I think that there's a, the first part I was gonna bring up was with respect to the fact that I think there is a view sometimes that it's easy, right? It's just easy to send emails. It's easy to send text messages and I think, you know, Chris and you and I could go on this journey for hours and hours and hours. And I face this particularly on the investor front when I speak to potential investors in the business 'cause they're like, I get it. So you send emails to people and they pay accounts like, you know, where's the evasion in that? And I think there's just a lack of appreciation, but you see that all the way up even sometimes in the predator and in evening the competitive space, right? Where it's like, hey, I can shoot a point stop side of desk. I can deal with this myself. And, you know, I remember early, early days that in debt of business working with closely with the Australian debt buyer and we were helping sort of get things out and they're like, oh, we decided just to do it ourselves and put it all into mail chimp for the three days later, the domain had basically been listed. And I sort of like, hey, what do we do? And I say, well, this is the thing. There's so much nuance to how you unpack it. So I think that there's this, you know, there's a component of it. And I think Jim's touch on it well, which is you go to the cost mitigation piece. Well, if I can send emails versus letters, that's way cheaper, that's good. As opposed to, no, I can actually lift up election and recovery rates, but to do that, there's a lot of slices of layers of data and technology that need to be done. And I imagine Scott, that's obviously why, you know, you see a big opportunity on the advisory side as well, because you're like, where do you start with this? And how do you sort of build this up from this ground up? I think back to, it actually is kind of easy to start. And I was surprised I was at something, I won't name a year and a half or so ago. And most of the people in the room had just started emailing SMS. And it was a party. Like they went from nothing to something, and it was like the greatest gift in the world. And I don't know what percentages they were at, but it was about as basic as you could imagine. And what was interesting, I thought was interesting, is there was little discussion around what's next or what's possible to then justify to your point, Josh, the adequate investment or to make that a reality. Like people were still sort of celebrating doing something that worked. And the portal didn't get locked up, but they got a handful of payments. So a false sense of achievement, but not knowing what's possible in order to figure out if are my 2% or 92% of the way there. - I think where it gets difficult, I think you're right, Scott, being able to do it, you could send, you could have being organization sending emails out of Outlook, right? You can do that, but Gale is the hard, how do you do it at scale? And that's actually a completely different than doing it on the side of somebody's desk. And in order to achieve, I think what Jim was talking about over indexing on the cost piece of it potentially, being able to do it at scale is really how you get those cost benefits. - When you can also see line of sight to this not being a channel on migration activity, and you can begin to see path to a customer experience optimization and a better relationship and more trust and greater and faster payments, the whole different space opens up. But tier one is, most people just focus on the cost side. And that's not super easy to do, by the way, that the first tranche is, but the majority isn't. But when you can really start to see line of sight a year or so away with some tech that's newly out and some that's coming, it does feel like we're in the second inning at the moment. - Absolutely. - Chris, then pushing across to you on this one, sort of how organizations should have crouched sort of that technology implementation, or call a digital implementation to ensure success and then you can bet your abuse and things. - Yeah, I think we touched on it a little bit, right? I think folks tend to give up too soon, from like the digital transformation. And I think, you know, Staten has some of these conversations at the Arntag Conference. People were saying like, oh, it's didn't work. Like, we don't get any of our payments to the portal, so we don't really want to invest anymore in that. But not taking the time to understand why don't your payments come from the portal, right? Is that you, multiple problems there, is you have too much friction in the login process? You solve that differently than actually, my email student gets delivered. They logged in, but they had a really high break rate or a really high first pay default rate. All three of those could lead to not enough payments coming from my portal, but they have three different root causes. So you really have to understand what is the metric and then how do you remedy that? And without, you know, throwing in the towel too early, without dedicating an analyst to understanding why is it not working? I think it folks give up too early. I'm, you know, how to front see and I've seen the success the digital collections can bring to an organization. Like, I am a huge evangelist. I fully believe in it. So I'm like, disheartened when I hear people say, like, ah, it doesn't work. Like, it doesn't work on my paper. Yes, it does. It does if you just do it, right? And I say this time and time again, like, an analyst will always pay for themselves. They'll always pay for themselves. Give them a problem. Why not a problem? And like, why is this that way? And a good analyst will always pay for themselves. - I need, let's just go, you go. Go for it. I was just gonna say you touched a little bit on it just earlier, Chris, but like dealing with it then at scale is an entirely different challenge. Can you pick on that thread a little bit? I think it's clear. We know this well. Bigger things get the harder problems get and the data gets bigger and things get messier and it gets more and more and more complicated layers and layers. But how do you approach that? I think if you take the view from what everyone's been saying that you can get that sort of excitement, you can, as Scott said, you can get something up and running. You can see, hey, customers do interact with this. But then you decide you want to invest and there is an ROI and there's a business case. What does that path of scale look like when you start to deal with millions of consumers and sort of dealing with this at real scale? I think it depends on how you win about the starting point. If you start in, I'm gonna use my example, if he started like, hey, I'm sending a thousand emails out of Outlook or a thousand emails out of Gmail today, scaling requires a complete tear down and rebuilt. Unfortunately, if we started, hey, I'm gonna integrate with an ESD like email chimp or send grade and I'm gonna have some infrastructure set up so I can monitor my engagement rates and I know what's happening and I can make tweaks. Then you can start to scale a little bit more easily. You need better data seeds. You need closer to real time data seeds. You need to move away from batch and towards webhooks. That's how you start to achieve that scale because you start having to monitor where at the beginning, you can say, oh, I'm monitoring this on a monthly basis. I'm gonna tell you how many payment came from my Coral List month, the bigger you'd get, the more frequent the interval needs to be that you're looking at your results because small problems at scale magnify very, very, very quickly, whether that's a revenue problem, a cost problem or a client's problem. So you have to start looking at things near real. - Awesome, jumping tack and no, 'cause we're running on the time pace, but future opportunities is one of the next big things and I can already see some questions coming in to sort of touch on some of these here. But Scott or Chris and whoever wants to sort of go first, what's the most exciting stuff you're seeing? There's obviously one particular word that's getting used over and over again. There's got two letters in it. But like, keen to see sort of, yeah, what you're seeing, what excites you, what are you passionate about and what are you going to see? - Let's go from there. - I'm wondering what the two letters are. AI. (laughs) The buttons. - Oh, okay. - Not real, but. My, I'll start, Chris, if that. I think what I'm very excited about is when you can not only do the analytics within a given channel, but start to connect it across multiple channels and they all learn from each other. At the same time, in real time, like, that's not flying cars. It's just called real time data aggregation and analytics. But you can pull it from your portal, email, SMS, voice, chat. When you can understand where the customer's engaging to what degree, when, what they do or say in those, and have the other channels learn, I'm really excited about that space. And you begin to then be able to have an integrated conversation with the borrower instead of a transactional conversation. I don't think we're quite there yet, but all the tech pieces are there, and I'm confident that's where the consumers want us to go. So I'm super excited about that. I think to some degree, RCS will get us a good step of the way there when it shows up. So that's what I'm really looking forward to. As each of these channels are optimizing within themselves, when those wires start to get connected, the data gets smarter and the consumer notices, and really engages at a much more intimate level. - I'll echo some of what you said and make a point that I've been trying to like evangelize in the industry here, of late is that our customers, whether we're in elections, right, are Amazon's customers. Or not fight, I'm not agency A, fighting agency B for that customer. Like a customer has a limited wallet share. We're talking about a customer, not an account, and a customer has a limited wallet share. We are literally competing for that $50 versus the Amazon Buy Now button. Like that's who we're competing against, that's where we need to get to. So I think everything that you said has got like the technology in some place, right? Amazon can do it, Amazon knows what's in my part. They follow me around the internet. They send me a follow up email. It's so convenient, it's so easy. I like to think about it when I don't even mean to think about it because they have managed to, like another one of my buzzwords, like attention saturation, like it's all around me, all the time, shows up in my Instagram feed, it shows up in my email, it shows up on my phone, it shows up in a push notification. Like the technology is there and has been there, e-commerce has been using it for years. I think our problem in this industry comes back to the being a quality data hygiene and willingness to invest in the tech, willingness to invest in that same technology stack to enable it. So what I see as kind of happening here in the, oh, the medium term is we'll see industry consolidation because the pace of change, whether that's through AI or not, the pace of change will fulfill some folks out of business, forced consolidation, which will then allow for, you know, a little bit thicker margin at bigger players. And then we will start to see this transformation to more of an e-commerce personalization, customer level conversation happening across kind of all the channels. So thing one, and two, speaking more directly to the AI piece, I think there's an interesting potential if you take the hypothesis that all the predators have to date, not devoted to resources on the collections and recovery side because it's like never the most important thing. We're going to dedicate those resources to originations. Advancements in AI, whether that's chatbot, whether that's sentiment analysis, whatever we want to talk about, that could actually lower the threshold for investments so that the predators start to invest more in their own book, which would decrease the third party. So I think that may be the more pessimistic side versus the optimistic transition to e-commerce. I don't know, yeah, exactly where to put my chips. Yeah, that was the Patoist interested to pick in on, let's start for you. Let me see. Chris, and one thing you mentioned, the whole closing loop back to the beginning of the conversation around predator collector handoffs. I, as an account goes predator agency agencies, you debt buyer agency, every person that grabs it sort of starts over, right? The relationship starts over. The data gets a little bit misjumbled back to our mapping topic and the consumer loses and is frustrated. And there's just, there's just tons of misaligned incentives and inefficiencies in my opinion. And I don't see how we solve that, but hypothetically, you just connect those dots, you share data as the consumers lifecycle or situation changes. The next person down the line benefits from what the last person learned and sort of everybody wins. I'm hopeful that we're able to crack that piece of the puzzle as well. I had had some more hope in that particular bucket. This is gonna sound a little bit counterintuitive. I had hoped that potentially the CFPB's 1033 rule around data portability and the customer is moving their own data. Push the industry in that direction, but the current, on the surface, the year round, I'll switch that out indefinitely. Because that is, right? If you wanna talk, that has always been the talking plan. Like, hey, customer's can, but with their feet in this particular industry, well, if we gave customers the ability to vote with their feet, now would it be a very painful transformation? Yes. Now collectors can start competing on customer experience, which I think is, you know what, like I said, what to do, but we still have to fit into the ecosystem. - Great. Jumping across to an Q&A that we've got one in and again, it sort of touches a little bit on the AI piece. But Cisco, Chris, everyone said jump in, but what experience is there with integrating sort of fintech SaaS virtual agent providers? So virtual agents, I think, is the thing. Got you wanna tie this one? Got any? - Hi, I spent a lot of time in this space. I'd love this. I think RMAI was my eye opener. Last year, there was none. This year was, there was 15, so I don't, and I don't know if anybody's there yet. But next year, they will be. And I can't even, I don't know how to size the impact it will have, but when your largest op-x line drops to tied for your lowest op-x line, it really is gonna change the dynamics a lot. I also think from a consumer perspective, the jury's a little out, but I could hear both arguments that they don't like it as much, and I can hear arguments that they might like it a lot more. Either way, it's more efficient. Arguably could be more effective. And this time next year, we could pencil in another session. I just, someone's gonna figure it out, and when they do, it's gonna be big. So I'm waiting on my tippy toes to figure out, how that unfolds. - Thank you, it's quite on the op-x line. It's an interesting one. That's what everyone thought digital was gonna be for them, which was like, oh, we're gonna get rid of all this costs, we're just gonna create all this agent-belts, customer-self-service. And then that really did materialize for most agencies, right? So just because the technology is there and can be integrated, it doesn't actually mean it affects your costs, you don't integrate it correctly. So I think that's where the doubles and the details, if you will, the technology might be there. You might sign the deal with the Fintech provider, but if you don't integrate it correctly, you're not mapping it correctly, you're not giving it the right information, you're not getting data back at the right frequency, they're gonna just cancel the contract. - Yeah, I think just starts point earlier around the, do you get mixed feedback? People like it, people don't like it. I think ultimately when they come down to it, I don't know how effective is it, right? If I can jump on and get instantaneous service and not need to wait 30 minutes on American Alliance hotline to try and get service, that's great, but if I call the button, it can't help me re-book my flight, or then I'm sitting there and I'm just wasting time. And that I think goes also back to what you mentioned, which is the quality of the data. And I think to your point Scott, I think everyone's running at this. The question is, so maybe has a data advantage and is able to leverage that and in what use case, because I think the problem is, I think everyone probably looks at it as we have a panacea and it's just gonna solve everything. And I think as we know, it's just never like that. It's gonna be more than potentially even lots of other different technologies come before. This is gonna be very niche, very verticalized down to each piece. I don't think it's even gonna just be as simple as, here's a collection they are. I think it's gonna be much more nuanced than that. You're gonna have different types of vulnerable customer stuff. There's gonna be different pieces fine tuned to dealing with more basic queries. There's gonna be stuff dealing with different languages. It's gonna get much more nuanced and how that all pies together, I think it's gonna be a really, really interesting approach. - Awesome. - Air. And pulls in other channels all at the same time in real time while flying your car to work. It's gonna be interesting to watch. And I think data cleanliness, data integration, AI's gonna hopefully help in the middle, but the pace of change will continue to, I think surprise us, but we'll see, it's exciting, a couple of years, I was for sure. - Yeah, absolutely. Well, if there's no other questions that I can see in here at the moment, I feel free to jump in the chat otherwise. Scott, Kristen, I appreciate you both taking the time. Jump on today, and for the audience, I found a really insightful. - Yeah. - Awesome. - Thank you for doing it. - Yeah, thank you for stopping. Ready? Have a great day. Thanks, everyone. (upbeat music) Thanks for joining us for another episode of Better Dead. Don't forget to subscribe. (upbeat music)