$350M Fintech Founder Steps Down as CEO to Rebuild With AI | Josh Foreman, InDebted

I have to think about a world where the business exists without even me in it. This is probably one of the craziest stories in the Sasapok Olympics era. I don't want to get to a position in three years time ago. I didn't work, but I only gave it like 50% of my time. Ten years into building one of the biggest step collection software company in the world, then he stepped down a CEO to rebuild it from its core in an AI-native way. I remember the first collection like it was yesterday, and making the first $12 of revenue. And today you fast forward and we make over $1 million a week of revenue. That new person that you meet for 15 minutes, they've spent 15 minutes, and you've spent 50,000 hours that sometimes you need to trust a little bit more. The company is already doing $80 million in revenue, while either over $350 million and profitable, yet he's taking a step back to rebuild from its core. This is how Sas founders are staying competitive in the era of AI. My biggest regret was probably my name's Josh Foreman, the founder of Indebted, and this is Founders in Motion. You said publicly that the competitor you're most scared of. It's a lean AI-native version of yourself, starting today with nothing to protect. So when did that realization land? I would say it became very evident from end of last year, early into the start of this year, where I think a lot of other founders, we noticed that the quality of the AI tool it got so much better, that for me was a big push that I have to make a decision on how much time I'm going to spend back on the tools. So I started making a concerted effort to clear my calendar a little bit more and have more time doing that, and then started to be able to see just how much you could get done and how much you could essentially build with less resources, and that sort of then started to constantly have me thinking about what if this was the first year that I found at the company it wasn't 10 years ago, and how different would that first three, six, 12, 24 months of getting the company off the ground book like? And that was where I think it sort of hit me and just said, "Okay, well, I know personally I would have been able to go a lot further with a lot less." When you think about the approach to make indebted more AI native, so you took quite a drastic approach, stepping down a CEO, focusing now on product development, why the decision and what do you think are core fundamentals that lead you to make that decision? Yeah, I think there's three things for me to stand as. The first one is the company itself, and then what does it mean to be the CEO of that particular company? So I'd say, like the first thing straight away is that I don't necessarily know that there has to be the approach for everybody, but I think being the CEO of indebted requires a lot of the other non-technical components of being a CEO, just because of the nature of the type of business we are and the sort of reporting structure, etc. So that's one part. I think the second is the stage of where the company is at. You know, partly maybe it was under sentimental, but it's 10 years since starting the company, and invariably I start thinking about two things, which is what is the right sort of, you know, path to make sure that the business has an ability to continue, you know, on with or without Josh. One thing I didn't realize when you sort of started this journey is that you always think of like there being a clear start and a clear end. Now I'm so confident on the ability for the business to compound over the long time, long term, that I have to think about a world where the business exists without even me in it. Whether that's me because I want to do something else or whether it's me just because literally I'm not here anymore, I want to be able to see this idea that the company could go on for, you know, decades or centuries. And so therefore you start to think a little bit differently. And so that has definitely caused me just to think about how do I decouple, you know, the need for it for it being myself. And then the third part is I just think the challenge is incredibly hard. I think the embedding the AI component into the product roadmap, I think we've done an incredibly good job of that. But I am concerned on how difficult it will be to rebuild things that have been around for 10 years. Everyone I was talking about tech debt and organizational debt. And I just think it's going to be a really, really hard challenge and I need to go to give absolutely everything I have in order to know if I can do this. And there's only one, I don't want to get to a position in three years time ago. I did a part time. I want to know whether it works or not that I gave it everything. As a whole, there's this whole gigantic rhetoric of SaaS apocalypse, SaaS is dead. I think there's elements where you can look at one particular product and we could say, okay, well, maybe you don't need that product in an AI world. And I think absolutely there's examples of that. But I think putting every single SaaS application into a bag and going this whole thing's crap, I think is just insane because obviously there are the best companies that have ever existed over the last decades. That being said, one of the things that is where I put my time and focus on is what these organizations, these AI native organizations fundamentally look like. And why is it that they're able to achieve things at such a fast period of time? And what's the nature of that? And one of the things that just sticks with me a lot is that the companies themselves will look so different in terms of how they're constructed. Less people, yes, but less people also and different types of people doing different types of roles and constructed in certain sort of ways. And I think one of the things that the market as a whole is looking at and saying is, well, these companies have far fewer people in order to get to these incredible revenue milestones and are doing them faster than we've ever seen before. Tech debt is quite a big hurdle to overcome. When you think about that on like a practical sense, do you think tech debt is actually something that is that difficult or is are people blowing it out of proportion? I think tech debt's not too bad. Depending on your, obviously, I haven't seen everyone's sort of co-based and stacks, but I think there's org debt, organizational debt, and that's the one that I freak out about. Systems and processes that exist outside and around the product that you have a look at from a technology lens. You can take any part of the org, as a matter of whether it's accounts payable or recruitment and chances are it's a fairly, it's definitely probably not AI native. It's potentially tech enabled at best using some sort of software platform or SaaS platform, but you haven't had the opportunity to throw engineers at that and say, what's the way that we're going to do this at scale and make that work really differently? And again, I go back to my hypothesis is that these AI native businesses have started, I'm like, okay, why do you scale from one person to 50? They haven't probably gone and reached for the same HR system that indebted and most of the other companies use, and they haven't approached it in the same way. And we can now look and see that they've been able to attract incredible talent and do it very quickly whilst remaining incredibly lean and fast. And so I think the tech debt piece is a real thing, but I think it can be overcome. We haven't yet seen real cases of people overcoming or getting through the sort of the org debt, and that's the part that is front and mind for me. Before we go any further, for someone who is listening that has not heard of indebted, what do you guys do? Yeah, so think of us as an AI-enabled debt collection business, which is definitely not the sexiest of industries, by far, but is a fundamental part of the ecosystem of how credit is extended. It started the business 10 years ago with a belief that technology could transform that, and in do so in three ways, the first is it has always been, and still is largely a human and analogue process. People on the phones, checks and people to pay their bills, and we thought that's got to be out of it digitized. The second is it's a very unsophisticated process, and we thought at the time machine learning represented the biggest opportunity, and now we think AI is the next lever to that. The third is it's a very ugly, cede industry, to be honest. Part of that is honestly because there's never been this idea of customer experience. You don't think of it as the same way you or I would if we went to a really good hotel, in certain age, in four seasons. You don't go, "Ah, I mean, at the same type of experience at the four seasons, I am with a debt collection company." But why not? Why can't you be respectful and treat people with empathy and provide a really good user experience? And so those were the three things that we thought about. And somehow we've failed forward into a business that has done a really, really good job of that. We have a long way to go. But we help millions of people every year go through that process in a far better way than they have before. Talk to me a little bit about that. So obviously debt collection carries this very specific reputation. So being very aggressive, dehumanizing, and in some cases slightly violent. What were some of the changes with software that you were to do to create a more positive experience for both parties? So one of the things that I think is helpful for the audience to understand is that someone is technically in collections from the moment they're one day over Jufe. There's a lot of reasons why that could happen. So potentially there was a lot of things that were happening. That was-- money in an account or a card was declined or I just traveling or a busy building a company and I forgot to pay the bills which ironically happens more than you would expect. So that in itself means that this idea that the human that is in that process is some type of you know archetype of person who doesn't want to pay their bills or is incapable I think is one of the first things. And so the question is how do you identify people onto that sort of quadrant and we look at it in sort of these two accesses which is what is that person's capability to pay like I actually have the money I want to pay it or have the money rather and then the second access is their willingness to pay. So I do want to pay. So in the top right hand corner you have Josh has money to pay his $30 bill and wants to pay it. That user experience is then very different. So if the way that I have to pay it is print a barcode and go to Australia post and scan it and give cash. Wow that is a hard process. So like don't don't expect that to be reflective of my capability or willingness but you haven't necessarily made it easy versus an SMS with a leak and then I double tap with Apple Pay while I'm walking down the stairs to grab a water. That's a totally different experience and so we can change that. And then you've got to identify people on the other spectrum which is maybe they have a willingness but they don't have a capability. Maybe they haven't been paid on time or there's a mismanagement of funds and things like that. And so at its core what technology allows us to do is to understand well where do we think people are on that quadrant and as we engage with them even if they don't realize that they're giving us signals as to where they may be on that quadrant and we can use that to change how we engage with them. And so one of the examples I'll give you is let's say in Australia for a traditional bank if you haven't paid after 90 days they commenced a lawsuit. Well what happens if you were in Europe during that 90 day period and the only way they were contacting you was via a phone and you happened to have a different SIM and they didn't email you and they sent you physical mail to your address in Sydney. The hypothesis would be we need to sue this person because they're absolutely not willing to pay but they could theoretically be sitting in a villa in Italy having a great time enjoying a holiday and be like oh I just didn't know. And so if you flip that and think about emails, text messages, phone calls, etc. happening during that process. If you know the phone is not ringing the SMS is not being clicked the emails are not being opened. It allows you to draw a very different conclusion as to maybe what that person's aware of and therefore how you should engage with them. I think I might have landed in one of your debt collection pull at one point. I have to actually oh really. Which is pretty funny. Yeah. Well then you know it's working so that's good. Exactly. And I'm so curious since you mentioned the idea of like calling. So I think traditional debt collection is very focused around like calling and making sure you're reaching people via their cell. Do you see voice AI playing a big role into the evolution of the company? One of the biggest. Yeah we're spending a ton of time in this area now. I think it's the final frontier for technology to own 95% of the collection stack. I'm very bullish on it. One of the things that was so unprepared early on in the journey of the business was this idea that yeah you have to pest a people we have to call them to get them to pay and what turns out is that people really don't want to have a conversation about this. But there are a percentage of people that need to have a conversation for whatever reason. And I always use this analogy but ideally when I book an airplane flight I do not want to speak to somebody right. I want to go on to the quantus app or the Emirates app but make a book in and whatever. If you have to call it's usually from a very bad reason right. Plains being cancelled last minute. Someone's lost my baggage. I can't do it through the app. I always ask myself would I prefer to be able to call have an AI agent answer me within one one dial. Solve the entire problem and have me offer phone for five in five minutes. All the experience that you and I are familiar with when you have to call an airline which is like a 90 minute wait period. Chaos is it no one's going to want to do it right. And I think the same is going to be true for debt collection. And there's too much belief that the humans were overwhelmingly not want to have that conversation with AI agents and our early data suggests the complete opposite. And so we're we're investing in a big way. That's definitely super exciting. And I will say I always I also fall into the camp of I would love to talk to AI. So at least you have one customer that's very happy with it. I wanted to understand like when you think about a new technology frontier like voice AI what's the choice between buying versus building internally. We've done both. I think early on the the humble answer is we have to have no idea. We just went down an approach which is let's do both like let's go to market and find what's the best. And then through that process let's learn let's also build and test and understand it. What I've learned over the last particularly 90 days to 180 days because we've been very deep in this process recently is that voice AI is a multi layered sort of system of complexity. So things like text to speech. We are not going to be the best in the world at that. There are phenomenal companies out of the valley doing incredible stuff raising awesome rounds. Let's use those products. What is the right way to engage a customer for collections and be the best at negotiating those outcomes. I think we have a proprietary data set that allows us to be good at that. And so I'd say it's build in our case but with a very big asterisk above it which is building using industry leading layers of the stack to help us do that. I think voice AI at the cursory lens you could think it's super easy but I just had a conversation with someone also building voice AI in the valley and the test and developments and like all the interchanging and all the small little tweaks that they do is actually very meticulous to get to to get to an answer that sounds like a human. So you guys have quite a unique pricing model for like a software stack. You charged not on seats but charged on results from very early on. Why did you guys make the decision and what has that done in terms of helping the business grow? I've so many mixed feelings about this one because it's been a huge challenge of growing the business. So the industry as a whole operates on this contingency performance model generally speaking which means we take a percentage of the dollars we successfully recover and if we don't make we're not successful we don't get paid. It makes a hundred percent sense why the industry operates that way because it basically just comes down to accounting standards but once accounts get to a certain age over due in majority of the markets around the world global accounting standards require you to write those accounts off and so no good thinking CFO is going to go I should spend a dollar chasing an account that I just lost a hundred dollars on right so I gave someone a hundred dollar loan I had to write it off now I'm lost a hundred dollars this stupidest thing would be to spend a dollar to go and get it because now you just lost a hundred and one dollars so you're increasing your potential loss pool so that makes a lot of sense venture investors did not they did not think so in the early days of growing the business and I understand why predictability recurring revenue etc so we spent a lot of time trying to make a different way and it was that was mistake big mistake and you try to push too far against the grain you try and get too creative you don't just accept the reality for what it is which is this is still a enormous multi hundred billion dollar you know 10 of revenue and one of the things I love about the whole AI world now is all the sudden C fees are no longer cool and everyone loves predictable sort of like so similar predict that outcome based revenue and so we align with that in a really in a really good way it's a very nice thing for indebted when we can say to a client we helped you recover a hundred million dollars last year and we kept 10 and we gave it in 90 and most of that was all written off debt there's now 90 million dollars of new revenue into your business and we got 10 million and we took all the risk it's just such an easy way to align incentives and outcomes isn't it great when the tides turn back to your favor so good so good and I mean you are one of the biggest collection software company in the world so maybe there is something to the pricing model itself so you scaled indebted from a solar carter in 2016 to a global company with over 200 people across seven countries somewhere in there you transition from builder to CEO so when you went through this transition of being a CEO and having all of these additional responsibilities on top of just building the software what was that like did you enjoy the transition did you I did enjoy the transition I think it's always I think one of the things you have to lean into when you start a company and it starts to scale is you as an individual will need to change but I know one thing is true I love the early stage of building I love the early days like you know as you recruit different people you find people are good for different parts I think I can clearly scale because we've been through the journey but I just have like a love for those early days like a nostalgia when I think back back into those times and not to be be, you know, against the reality as well as very hard. I didn't have a lot of money and no one had any good cash. And it was, there were tough days and very, very difficult, difficult times. But I, I like to be in the engine room and see something come to life, if you will. I remember the first collection, like it was yesterday, and making the first like $12 of revenue and being like, that was the coolest. And it, it honestly, it meant more, I remember the first dollar. And I also remember the first million dollars collected. We have a video of myself and a few other team that was like, four of us at the time and I think we popped a bottle of mowey and like celebrated because we collected a million dollars. And today you fast forward and we make over a million dollars a week of revenue. The early Builder's Day sounds amazing. Was there ever any conflict or tension between like your builder side and your executive management side of things? The has been was and is in that I, I lean into problems in a big way. And I get a lot of self reward and satisfaction from solving complex problems. So I don't do well as an executive working with other executives who are scared of, I use this technology a lot, but like just staring into the abuse and chewing glass, I think is, you know, I am definitely not the best engineer. I'll be very clear on that. My biggest skill though I think is that I'm technical enough, but I also have a good commercial acumen. And so I think from a product perspective, what that's allowed me to do is I'm not the engineer that paths myself on the back because I think the code's really cool. I'm the one that passes myself on the back because someone just paid us some money for something that we built. And that's the part that I can bring and all have always brought to the team and it's the ballers between again, that executive side of I know my investors want to see X, you know, a million error or whatever it is. And how do I translate all the way back? And then I'm in front of a customer and going, well, if I just build them this, I can get said number. And then I can achieve said goal. And so that's sort of where those two have worked well. If the refounding journey works, what does success look like two years from now? All companies are a factory. And our job is to produce a widget. And we should have a brand new factory and produce widgets at speeds that seem like completely impossible to the business that we are today. So, and the reason I use that analogy is I don't want to get too caught up on revenue growth rate numbers and revenue per employee, all of which are important and have tapioids for all of them. And I want us to hit them all. But I want us to be able to go like whatever it is that we produce in the company, let's say it's a line of code or product, a higher performance review. I want us to think about it as well, how long did it take us to do? Let's just use performance review. That process in the company before. And today, I'll be honest, it feels a little bit like a conveyor belt bringing luggage out of the airport. I want to see it being like one of those high precision sort of things being stamped out of it. So that's how I want the company to be because I think if we are doing things like that, it means we have mastered the automation component. We could harness all of the brain power of everybody in the company and say, guys, let's focus 100% on this problem and go and solve that. And that's the next unlock for the company. I'm excited to see this factory come to life. You can spin out widgets super quickly. So in the past 10 years, what do you think is your biggest regret in your founding journey, something you wish that you knew, something you wish you could avoid it? My biggest regret was probably the times where you haven't entrusted the instinct as much. And even though I had done businesses, a couple of businesses before that were successful, indeed, it was a different level of scale and success and in terms of where it is, that you invariably doubt yourself quite a bit during the journey. And you do it because you meet people that you overwhelmingly respect. And then they say something and then in a go, you sort of sit up at night and like, have I got the idea wrong? So an example is like you mentioned on the pricing, the amount of times that I met VCs where I was like, oh my god, this person, they run this firm and they're invested in these companies and they go, you're pricing model sucks, I would never invest in that company. And then you go home and you're like, you know what I should do? I should do their right. I'm just going to convince an entire 170 billion dollar machine just to just completely change overnight because Josh says so because this person says so. But they haven't had, you know, they really got to meet me for 20 minutes and they didn't get to understand the industry. And you go down some stupid rabbit holes and you do some stuff that ultimately you get to where you are today. And it's like, it's the same for a comprising model that it was when we started. And I attempted to honestly rebuild that 15 times because of feedback from people. I look back and I regret it because I wasted not only my time and emotional energy and things like that, but it also causes you to make the wrong hire. And then you've got to get rid of that person or you make the wrong decision. You waste a million dollars or five million dollars. And it's just, I think for founders, there's one thing that you definitely have over everybody else, which is you spend more time on this thing. It's your baby, it's your child. And so if I had to cumulatively add up the number of hours I've spent on indebted, there's one thing for sure, no one will ever spend more time. And so that doesn't mean you're perfect on it, but it means there's certain things about like how the machine operates and your intuition on how the company should move forward. That I think if you looked at it in some sort of metric scale, that new person that you meet for 15 minutes, they've spent 15 minutes and you've spent 50,000 hours. That's 50,000 hours in front of clients and customers, investors and pitching and things like that that sometimes you need to trust a little bit more. And for me, it's like some of the worst moments, some of the unmask decisions and things I've made because of those conflicts where it's not really a conflict on data, it's a conflict on this person that said something and you go, I just think they're the best person. I think trusting your instincts is really difficult, especially starting out, because you do suffer from a lot of self-doubt, especially when you haven't seen the fruits of your labor starts to manifest or the fruits of your decisions start to manifest. Okay, Josh, I feel like I've learned a lot about the founding journey and I'm so excited to see the next version of "Indedits." Thank you. I really appreciate it. It's great chatting to you today. That's a wrap. If you liked this episode, please hit the like and subscribe button. It helps us bring on more awesome guests, love our production and bring on new series you'll want to watch. And if you want to hear more early-stage builder stories, check out our other episodes. Okay, see you next time.