Jake Aaron Villarreal: Welcome to our podcast, From the Ground Up, where we interview startup founders exploring their journeys, their success, challenges, and lessons learned. We hope you be inspired in what it takes to build a thriving startup. I'm your host, Jake Aaron Villarreal, and here with us today we have Cameron MacArthur, CEO of AI Insurance. Cameron, welcome to the show.
Cameron MacArthur: Hey Jake, thanks for having me.
Jake Aaron Villarreal: Great. So a little bit about Cameron. He's a... he graduated from Northwestern and has been awarded for his research in human-computer interaction. With a background as a cognitive software engineer at IBM, Cameron is focused on bringing a data-focused approach to the insurance space so that companies can unlock the potential of AI. He's building AI Insurance, which is a SaaS platform for running small insurance programs. Uh a lot going on there, Cameron, and excited to dive into it. Um before we do, um where are you joining us from today?
Cameron MacArthur: San Francisco, right in the middle.
Jake Aaron Villarreal: Great. It's always uh kind of a mixed bag when you think about the weather there. Are you originally from San Francisco?
Cameron MacArthur: No I'm not. Moved, moved out here to start the company. Um born outside of Chicago and grew up in New Jersey.
Jake Aaron Villarreal: Great, East Coast. I like it. I spent some time in New York, so I understand that whole culture and that vibe. And both upsides and downsides to each coast, but happy to have you on the West Coast where we're located as well. Jake, I grew up in Silicon Valley in San Jose. So spent a lot of time in between Apple, IBM, and Hewlett-Packard (this was way back in the day). And then of course Google and Yahoo and all the other companies came up. And you know, everything is startup world now in, in the Valley. So that's where, that's where I started, and I'm now in uh Southern California, so south of LA and north of San Diego. Different weather. Um so anyway, thanks for uh hopping on here. Um you know, before we dive into your company, everyone has experiences that shape them as an entrepreneur. Whether it's modeling their parents or experiences from their youth. What was it that kind of led you down a path of eventually becoming an entrepreneur as a youngster?
Cameron MacArthur: Good question. So I, I'll zag on this. Uh Scott Galloway has this great um I guess it's a line where he says that a lot of people think that to become an entrepreneur you need this special thing, you need some X factor that nobody else has. And his take, and I fully agree with it, is that it's quite the opposite. That you become an entrepreneur because you, you lack this X factor that other people have. Um and in my case, I, I think the, the lacking was around being able to work at a big company. Um like I, I think it requires a lot of political savvy and patience to work at a really large corporation, right? Where I, I had friends that were at big companies and you watch them kind of, you know, take some hits to the jaw, right? Their manager takes credit for their work. And you know, it takes three and a half years to do you know, an idea that they had on day one that they could have implemented in a month and a half, right, just given the resources. But you can't do that because it overlaps with this other team at the company. And there's polit- that team's managed by somebody else, right? And all that kind of like political big company slow stuff.
Um I think if you're good at managing that, and it is a skill set, I, I think you can often be far more successful in life than, than if you started a company, right? If you, if you're able to play the game and you can go be a partner at Goldman Sachs or something like, you know, your, your financial outcomes are, are going to be fantastic, and your quality of life will certainly be better than starting a company. Um I think I, I started a company because like, for the goals that I wanted to achieve, I just didn't have... didn't have that skill set for the big company.
Jake Aaron Villarreal: Um yeah, I understand what you're talking about cuz I worked at Oracle and it was at the time 20,000 employees, I think there are 100,000 today. And it was always uh navigating the, the process of how to get things done, who you have to work for, collaborating with your team members. And you know, "oh by the way I actually have to do a job and, and you know, hopefully get paid and progress." And yeah, there's a different core skill set, you know. I always ask you know "what is it that, what's the main byproduct of being an entrepreneur?" And for a lot of us, myself included, it's about the decision of making your own choices, having the freedom to decide what you want to do, how you want to do it, when you want to do it, and who you want to do it with. For you, what's kind of... as you've now gone through this a little bit and you're an entrepreneur and gr- what are some of the things that you're getting out of being an entrepreneur that you didn't think about before you started leading a company?
Cameron MacArthur: Well, one of the things that I did think about was the freedom, right? Of the thing I tell people is I certainly work more hours than I did when I was at IBM, but I have much more freedom and flexibility and ownership right. I, I think when I was in a big company at a team, I', I'd feel guilty if I you know, took a day or two off because I'm like "I'm accountable to my whole team" and you know... um and I think when you're a founder it's sort of like no one questions how much work you do, right? You don't need to prove to anybody that like you're ride or die for the team. Um and so I, I just felt more, more flexible I guess.
Um what is a surprise? That's a good question.
Jake Aaron Villarreal: And maybe there is no surprise, maybe you're just, as you're learning you're going and you're growing. Um but you know you always oftentimes start with what the idea uh is about, what you're about to do, and you already have some thoughts about what it's going to be like. And you get there and it's nothing like what you expected. So maybe thinking of it that way, what's, what's, what's it been like so far that's been eye-opening for you?
Cameron MacArthur: There's a lot of surprises that are eye-opening. Not so many of them fall into the category of "what's really good." Um like I think one of the big surprises is around the kind of connective tissue work. Like that, that sort of "fill in the cracks" work and how much of it there is. Um you know on, when you start a company, when you have an idea, it's just the idea, right? For me it was, "oh I wrote a machine learning algorithm in Python that predicts the cost of a claim." And you're like, "well what's the company at that point?" It's you know, building a product and working on this algorithm and then doing sales, right? And you're like, "okay cool, those are the three things." And then that balloons a little bit. You hire some people and you're like "okay, now we need to do payroll." You're like, "well that's okay, you get you know Rippling or Gusto for HR, and QuickBooks, and you know you, we use Brex for banking, and it's like, great all that, all that work's done." You still like... but every one of those things adds like a tiny you know, an almost grain of sand of work.
And then as the company grows you know you, especially as a, as a technical founder, I think I always just saw the work as the technical side of things. It's like "the building, that's, that's the work." Um and then you kind of hit a point where you're like, "oh I don't, I don't do any of the sales and I don't do any of the building of the actual thing and you know now we have like a Head of Ops who's doing all of the, the payroll and the HR and the running stuff and you're like, so what am I doing? But also I'm so busy all the time! Like I promise all my hours are full." Um and it's all that kind of yeah, I don't know, fill in the cracks kind of mortar work of, of running a business and that I think sneaks up on you. Um you know, I, I think I took too long to even realize that there was this big mountain of non-technical work to hire our, our Head of Ops. Um I had a very funny conversation with her when I hired her because initially the position was part-time, and I was like... because I just didn't think that there was that much you know work, there wasn't that much meat on the bone. And yeah she's very experienced and I was like "yeah this, this role's part-time. Are you okay with that?" And she was basically like "yes, but you're wrong. Like I'm gonna take this role because it is extremely full-time and you just have no idea." And she was right by a mile right, on day one it was a you know 10-hour a day role. Um so I think that that's snuck up on me. All this, all the other stuff if that makes sense.
Jake Aaron Villarreal: Yeah, lots of stuff going on that we all have to work through and manage and learn as we grow. Um let's talk about you a little bit. So you, you worked at IBM, you were there, you got the big company experience. You understood what that's like. Now you're doing something that's innovative and you're building your own dream here. What was the inspiration behind what you created and what you're building? And for the listeners, what space is it in? So I'm going to kind of throw a lot at you. What is it that you're building? What problem are you solving? And what was the inspiration behind wanting to actually even create something in this space? How did you even think about a problem here?
Cameron MacArthur: Great question. So alright, I'll give a, I have a good analogy for the "what are we solving". So if you think about an insurance claim, like you know you get rear-ended in your car right, the whole thing is kind of open and shut, right? We know how much roughly a Subaru bumper costs. You know your, your insurance company is like "oh we got to replace the bumper," they talk to an auto body shop, it gets replaced right. Maybe it takes two weeks or a month or something. There are other kinds of insurance claims that are high complexity. An example of one of those is medical malpractice. If you sue your doctor, they have medical malpractice insurance to protect them from that lawsuit over an alleged mistake. That insurance claim and that lawsuit will take 5 to 7 years on average.
Jake Aaron Villarreal: Wow.
Cameron MacArthur: And at the end of those seven years, about 15% of the time they'll pay you 100 grand and they'll pay their attorneys 350k. And they're like, "oh my gosh, if I knew I was gonna give Jake 100 grand, I would have given him 200 grand six and a half years ago. And Jake would have gotten twice as much money that you know, the person who was injured. The doctor who made an innocent mistake wouldn't have you know a med mal case hanging over her head for six and a half years. And the insurance company would have saved $150,000," right? So kind of everybody wins if they can figure out how to close these claims more quickly and more efficiently. But the problem is nobody knows ahead of time what the claim is going to cost seven years down the road.
Um and so what got me into this space is um when I was living in Boston, my girlfriend at the time actually worked at a medical malpractice insurer. And I got exposure through just hearing her talk about her job and just being sort of aghast at how insurance companies ran and how much money they were investing and spending on specifically this problem of predicting the cost of a claim. And so that's, that's what kicked off the whole company, was finding out that humans are actually really bad at predicting the cost of a claim. That you know, in a lawsuit their median distance off is $107,000, about 50%. So you're kind of just you know, shooting from the hip on these. And I thought, "you know I think machine learning could do this better."
So there's a public use data set specifically for medical malpractice called the NPDB. So good job federal government. Um you have to report to that if you, if you have one of these adverse events. And the first model was based off of that data set. And that's kind of what kicked off the whole company right. What we told you know, Y Combinator and then other investors was, "Hey, you know humans are not that good at this, this data is not even that good. You know, if we can get better accuracy than humans using this public use data, imagine what we could do with you know some real high quality data at scale where you know we're owning the whole flow." Um so that's the whole company.
And then we ran into this "cart before the horse" problem that happens in insurance. Which is that a lot of these insurance companies weren't even on the internet, much less had data. So we said "oh you know," (even back in 2018), "hey here's my money, give me the AI" (because AI was like the sexy thing back then even). And then folks would say "here's our claims data, it's on a f- in a filing cabinet in Orange County and we fly down there once a quarter and spread the claims out over our card table and that's how we review the claims." We were like "Whoa, okay, you need you know internet-first uh software to run your insurance company before we can start looking at the data coming out of that software to do things like predict the cost of claims or help you underwrite better."
Jake Aaron Villarreal: So the platform itself is a technology platform, a SaaS platform, Software as a Service. And the user of your technology are ins- insurance companies?
Cameron MacArthur: Correct. On the nose. We make a SaaS platform for running a small insurance company end-to-end. So they're doing their underwriting, their claims management, you know managing applications, issuing policies, making payments all out of AI Insurance.
Jake Aaron Villarreal: So if I own an insurance company in Orange County, which is where I live, and I have claims that are being managed that I need to dive into and do all the things that have to go with you know, along with the insurance business, but I'm not very technical but I definitely want technology. What's the user experience like? Who are you really selling to in that organization?
Cameron MacArthur: So we're selling to, to you in that case. A lot of these small insurance companies kind of break your mental model of insurance in that they might have one employee or four employees or five employees tops. And a lot of times they have just a board, and they outsource everything to vendors. And so we kind of offer two things. In the cases where things are outsourced, we're the connective tissue, they get all these vendors working together in one location right, everyone kind of plays nice in the AI Insurance sandbox. Or if you're you know doing everything yourself... yeah you're a four or five person team. So the, I think third hire we made um is our Head of Design, Allan. And from the very beginning we've invested a disproportionate amount of time and resources in just interviewing all of the different personas around insurance. You know we spent literally years just talking to claims adjusters and saying "hey you know, tell me about your job, what are you really good at, what are you really fast at, you know, what is really frustrating for you?" Because we had the opportunity for these small companies to build from first principles right, they weren't on software, they were on an Excel spreadsheet. So there's this sort of golden opportunity to say "hey this is going to be jarring and different, but you know we can really revolutionize this whole experience."
Um and that's because of the space really right. If you think about in our guy example of you getting rear-ended, your mental model of how an insurance company works is pretty spot-on right. You call Geico, some dude with a headset's like "Hey Jake, what's your member number?" And you imagine headset guy typing it into a computer and pulling up your policy history and how long you've been with the insurance company. And you probably applied into that same system on your phone right, you filled out the digital application and you can log in and see your policy and all that. And that's it, that system exists on the nose. Um there are a couple companies like, there's a company called Guidewire that's sort of the you know the, the one famous for that, and they sell to all the big insurance companies right. So they have 175 customers, they go after the top thousand insurance companies, everyone that you've seen an ad for right, your Progressives and your State Farms and your USAAs. And that sort of left this opening in the space of all these thousands of really small insurance companies that are too small to be served by this custom-built software and really have to get you know lightweight SaaS kind of plug-and-play self-serve stuff. And that's great [to] bring to the table.
Jake Aaron Villarreal: That's really cool. So you know the InsurTech space has been really hot, there's a lot of investment going into that area. What is it... talk a little bit about your niche and I think you hit on some of those points already, but what, what do you do differently than some of the other vendors in the InsurTech space that um you can share with the listeners?
Cameron MacArthur: Yeah of course. So we focus on specifically just the small insurance companies. And there's actually a kind of small insurance company that's a non-admitted carrier called a captive or risk retention group. Um and these are essentially self-insured, they're, they're member-owned insurance companies. Um and they, they suffer from a couple of problems. One is because they're very small, they can't even dream of affording existing software options right. Where if you're writing 5 million in premium a year, which is an insurance company's revenue, you're not spending a million dollars a month on software. You're not spending half a million dollars a year on software right. Um and so currently they're sort of priced out of the market.
And then the second component, and even if they could afford it, they have this big fragmentation problem. Where often the insurance company is not all under one roof. They might use Cameron to manage their claims, and he works at a different company, a claims management company. And Jake does their underwriting, and he works at a different company. And then they have a third person who does their investment management, and a different person who does their actual sales of the insurance and their brokerage. And a different person who does their regulatory filings and back office management. And so m- most existing software is charged by login. So even if you got... even if you could afford it, you get this situation where Jake says, "Well I'm not paying 1,500 bucks a month for Cameron to log into my agency management system," and you know "just because we share a client."
Um and so those are the two big things we solve. Because we're SaaS and cloud-based, we're very scalable right. We have folks that are just starting up a brand new insurance company right. If you're an InsurTech startup and you're saying "I'm going to start up an insurance program or an insurance company," then AI Insurance is a great fit for you because we'll say "great, like we, we already built the whole tech stack so you can just plug right in and, and run your insurance company." Um and then the second piece is often these small guys outsource everything, and we sort of use that you know fragmentation bug and turn it into a feature and that becomes kind of you know, we get a network effect out of that that helps with our growth.
Jake Aaron Villarreal: That's great. You know when we had our first call you talked about "building the dream data set." You know, walk through what that really means for, for, for people that really don't know what your product is and how you would even go about doing that, because that's something that I think a lot of companies are trying to figure out in the AI space in particular.
Cameron MacArthur: Yeah, I, I hope you've got big data folks that, that listen. So first problem we tried to solve was predicting the cost of a claim, right? And so initially we were like "great, give us your claims data." Then we found out that it was all on filing cabinets. And we said, "okay, we're going to need to create their claims data right, at scale." So we started with just claims management and we said, "great you know, you enter the data about your claim and you know put in the files and things like that." And then it became really clear that about 30% of the data on a claim is about the insured and the other 30% is about the policy right. And so then we were like "okay we're missing you know 66% of the data that we need." So, so that sort of started to balloon and we said "okay, let's do policy management and let's do insured management so that we can track that, the other data that, that plugs in here."
And then it sort of you know started to just spread to the rest of the insurance process. Where we said "oh well, you can't really manage the policy if you're not issuing the documents for the policy because you know if someone else is issuing the documents then that's where the data is." And yeah then we did that. And then it was "well you know if you're not doing the underwriting then that's where the data is." And then once we did that it was "well the application is really where the initial data comes in, and that feeds the underwriting, which feeds the document generation, which feeds the premium billing right." And so we ended up having to build out the entire value chain. Um and now you know it's all built on this machine learning layer of you know, we knew what our end goal was. So now everything from digital application through to rating, underwriting, quoting, binding, issuance, premium collection and claims management is all collected on AI Insurance and then we can train on it and do all that good jazz.
Jake Aaron Villarreal: Wow. Really valuable. I can understand what companies probably would be interested in. You know, you never know, there's a lot of acquisitions happening and you know "data is the new oil" which they always continue to say. But that's really cool that you thought about that. When you went into Y Combinator, was your thesis what it is today? Or did you start with one idea and as you went through left and, and made some pivots or kind of morphed into something different?
Cameron MacArthur: We applied to YC, or I guess I applied to YC (using the royal 'we' there), um with the idea of predicting the cost of a claim in medical malpractice. And then the pivot that happened is you know, that's still our end goal um and you know we have that running um. It's going to be going live to our first customers hopefully in 2024, we've been running it behind the scenes as you need to do for anything where you have a machine doing what a human is, is currently doing. Um but that was our initial application. And then the only slight pivot that happened was this "cart before the horse" problem of realizing "whoa, the, a lot of these companies aren't near the margin of machine learning they need. We need to get them on the internet first. Let's, and then we can talk big data and then we can talk machine learning."
Jake Aaron Villarreal: Getting them on the internet, wow. That's a real transition there. Um interesting. So as a company when you went in... or as an individual when you went through YC and you shared you know "I'm going to build this and I'm going to build that," and you started to see the sort of the, the, the, the amount of work that it would take at some point. You got some advice that they said, "I think at some point you got to hire somebody." And if you've never hired anybody, you've never managed anybody, you've never led anybody, what was that experience like for you? Because there's a ton of people on the fence that are going to be building companies in 2024 and they're going to go exactly through the same experience you had. What were some of the learnings [that] came out of that for you?
Cameron MacArthur: Oh my gosh, yeah, so many more, more, more learnings than can be counted. Um but yeah, you're on the nose. Came in as a, as a solo founder, was over my eyes you know, didn't know how to do machine learning well, didn't know how to do sales, didn't know how to run a company, and also didn't know how to hire, and didn't know how to manage. Um so sort of patched the hole by just bringing on a friend. So convinced a friend to quit their job and come work with me for you know the three months of YC. Um so you know, he, he didn't have interest in insurance you know, but, but was a hero and did, did the engineering for you know those three months while the company kind of got off the ground. And then uh you know, he left and then at that point we'd kind of gotten out of the you know, again using the royal 'we' there, AI Insurance was sort of out of the woods. And you know, we were able to hire our first person, which was a, a gentleman named Matt, who wanted to be you know the, the first engineer. Um and coming out of YC, I was also able to raise you know our first amount of money after demo day. Um and that makes things a little bit easier, less from a capital perspective, and more again because you know people want to be the first engineer at, at a startup right. They see the, the first engineer at Box, or the first engineer at you know Yelp or something. And you know, if you're an engineer a lot of folks fit in that persona, they want to be that person um and you know get a life-changing amount of equity.
So I think what you get out of YC is that kind of proof of two things: "okay this is a, a good idea and now we have enough capital to go see if there's product-market fit." Um so yeah, I think the big learnings are probably "don't sweat the small stuff," I talked about you know all that like mortar that fills in the gaps. Um yeah, and I'd say the biggest takeaway from YC is at the end of the day all you care about are signing new customers and making money. And if you do those things, nothing else matters. Like your landing page doesn't matter and like your financial modeling doesn't matter, and your... none of the other little stuff matters especially in the beginning, if you can get customers and get revenue on a weekly basis.
Jake Aaron Villarreal: Makes total sense, yeah. We hear that a lot when you get out of Y Combinator and you're really on your own in the process, but also once you're out, I mean you still have the connections and potentially even some mentors you've met. Um what's, what's it, what, what was your strategy to get those first customers? Because from what we understand for a lot of acceler- a lot of accelerators they say "don't even build the product, go out and get buy-in that you have an idea that companies want to pay for, and then you build it." But you still have to start with a list of targets and then you got to make the calls and you got to go meet and you got to hope that they like what you share. And you you know, by the way you're trying to build a company, you're probably getting a little bit burned out but you're also having to you know keep things rolling. What's um, what was your process in just having those connections to start?
Cameron MacArthur: On the nose, do all that before applying. So I, I applied right around, I think I applied a day after the deadline. And had already done a lot of lift of going up and meeting with you know medical malpractice insurers in Maine. And like one of the things that kind of kicked it for me was I cold reached out on LinkedIn to the CIO of a major insurer right, a... one of the largest in the world. And said, "Hey, I, I like did this thing, you know wrote this algorithm to predict the, the cost of a medical malpractice claim. Is this like good or useful? Like I don't know anything and I really just would like to understand and learn. And you know, I'm coming in ice cold to insurance." And they invited me to their headquarters. And I met with their Head of All Claims for North America and their Head of Medical Malpractice. And it was supposed to be a half hour meeting and went for an hour and a half.
Jake Aaron Villarreal: Wow.
Cameron MacArthur: That really drove home to me like, "Okay, there's definitely pain here," right? If, if these people who are you know way orders of magnitude more important than me are meeting with some schmuck who you know has an idea. And they were really helpful. They were like, "Look, you know, we're, we're way too big for where you are now right. Go try this on a bunch of small insurance companies. And then if it works, we'll probably you know, buy those insurance companies and buy you, right? Like we're, we're not a you know, first customer kind of thing." Um but it kind of demonstrated okay, there is real pain here if these people are willing to invest their time to even meet with me for that long. There is real pain. Um and that sort of catalyzed for me that I was onto something.
And then I started you know just again cold reaching out to tons and tons of folks. And cold reached out to a, a great guy named Troy at Risk Services. Um they're a captive manager, which is a you know vendor in this ecosystem of these small insurance companies. And said what I was doing and he said "oh I have a, a customer who might be interested in that." And that was you know how we met our first customer. Um and she was you know running one of these smaller insurance companies and we got lucky enough that she has a very innovative mindset and you know was interested in trying these things out. And like you said, you know sign a small little you know, back of napkin looking contract. I look back at that contract now and I'm like "oh my gosh, like [it] was basically just you know the first slide of a keynote presentation being like 'hey we'll try this and you'll be our customer.'" Um and yeah, she was you know very forward-thinking and innovative and is still a customer to this day. And you know that's all you need is a few of those that are, are willing to take a chance on you and see if this will work. Um yeah.
Jake Aaron Villarreal: That's great, great. You know, I think the key takeaway there is that basically just having the courage to send the, the InMails or the emails and to do the outreach where the outcome you know, you know, quite frankly could be a hundred "no"s before you get to a "yes". But you still have to get to those "yes"es and um, no one's there to kind of coach you and help you. As a solo entrepreneur it's kind of on you to make it happen. So...
Cameron MacArthur: You know, there's a risk profile there. Like I joke to the team that our first round of funding was a savings account, right. Because I was flying all over the country before YC. I went from my YC interview to Chicago to meet with this first customer. Um and that was all just you know, taking trains up to Maine... anybody has any interest, you say like "oh I'm gonna be in your city next week." Um and yeah, there you know, you have to believe in yourself. And, and I think that's an important part of raising money too. I think investors like to see like, "okay, this person's betting on themselves. They, they, they must think they're onto something if they're you know willing to put this much of their own money into it."
Jake Aaron Villarreal: Yeah that's great. I agree with that. What's, what do you think has been the biggest challenge so far as a solopreneur? Because there's a lot of companies we work with that or talk to that, you have two, three co-founders and there's a little bit more of sharing in the responsibilities, the risks, and you know every other part of the operation. But as a solopreneur, what's, what's been one of the bigger challenges that you've had to face or, or [are] working through?
Cameron MacArthur: I think obviously you know, the loneliness and the lack of empathy is always hard right. I think every founder talks about that, that you're going through a very unique experience that not a lot of people can, can truly appreciate just how difficult it is. Um and every founder's experience is unique right, because you're dealing with your own baby, your own company. Um and so not having co-founders I think makes that a lot more challenging. Um I think the other piece is that typically in a founder role you kind of want to be a jack of all trades, master of one kind of thing. And if you have you know let's say three co-founders, you can you know spread responsibility and say, "look, right, this is the thing that you're going to be exceptional at." And we all have to do everything, but there's a division of labor there. Um and I, I think that's just tougher when you're a, when you're a solo founder, right. I find that you know, I don't feel exceptional at any of the, the one disciplines right. I'm... that obviously gets easier once you can hire folks, right. Like I said, we hired our you know, our first engineer Matt and yeah, I remember walking with him and him saying "hey you know, what do you, what kind of testing stuff do you have set up?" And I was like "nothing right! What do you mean? Multiple environments, we don't have any!" Right? And this was, this was all like pre-revenue, pre-customers. But it was, it was you know like "oh I could not be born more behind you know." Um and so you, you got to I think be a little bit more comfortable with being kind of wrong and bad at, at a lot of those areas until you can hire someone who's, who's better than you at it. Um and yeah that, that can be tough. Um but yeah I'd say those are the two big, two big ones.
Jake Aaron Villarreal: Yeah that's, that's, we hear that a lot I think. Um you know you, you have to also strategize oftentimes on your own unless you talk to others that can help you solve problems. And as a leader your employees are looking for you to solve problems and to lead. So I think um, it's a little bit more challenging as a solopreneur, but at the same time there might be some real good rewards that come out of that too. As a company, you look at the market you're in, h- how big is that market for AI Insurance?
Cameron MacArthur: Yeah, good question. So just the United States, just property and casualty insurance, um just administrative expense, so just the waste basically right. Your, your running of the company and paying salaries and all that stuff. So just administrative expense, just US, just property and casualty is about $360 billion annually. Um and then if you look at just the small guys of that, so segment it one more time and just look at your you know tier four, tier five insurance companies, you're looking at like 247 billion.
Jake Aaron Villarreal: Wow.
Cameron MacArthur: Um which is interesting because you'll notice that that's a bigger piece of the pie. So large insurance companies actually spend less by a percentage on administrative expense, right. A big insurance company might spend 30% of their revenue on their operating expenses, and a small insurance company might spend closer to 60 or 70. Um which you know makes sense right. If you're, if you're at the scale of, of a Geico, yeah you're spending you know, I don't know, a million dollars a month or something on software, but that actually per employee ends up being a really, really low cost right. Um so yeah I think the, the crazy thing about insurance is the, the market is certainly there right. There's a massive amount of money being spent on you know waste and manual processes here that can be automated just by workflow software and you know further automated by AI.
Jake Aaron Villarreal: Well AI seems to be everywhere these days. As you look into 2024, what, what's on the road map? It sounds like you've kind of painted a picture of what I think is coming, but for, for you know just I guess from your perspective what's really exciting for you? What are you excited about?
Cameron MacArthur: Yeah there's, there's a couple things. So one is like the reading and writing skill level of these large language models is getting crazy good. So we've already rolled out a large language model for reading and auditing invoices of defense counsel against your guidelines. And I think leaning on that in the coming year is going to be pretty exciting. We're, we're looking at things like you know drag and drop an application that you got from somewhere and we'll just read it and parse it and you know pull, fill up your application with whatever that information was. Which is a big pain point in commercial insurance. I think we're going to start to see some of the reading and writing roles get handled better, right. You know a claim comes in, is there coverage? You know, the ability to read a policy and make that coverage decision um is something that you know right now if you said that to an insurance person they would say "no way, like never." Um and I think we're going to start seeing some of that stuff happen over the next, over the next year.
Um we're also particularly excited about um the underwriting piece. If you look at just how banks underwrite risk, like credit default modeling and stuff, it's actually pretty good. They figured out "okay you ask these questions, they draw your risk curve and they say okay give Jake a loan at this interest rate or don't give Jake a loan" right. And that doesn't really exist in insurance. I, I feel like there's, there's not yet the data to really well paint those curves. And so we're looking forward to, to start just optimizing, finding a local maxima of saying "okay you know you should adjust the credits and debits that you're offering you know it shouldn't be 5%, it should be 2.6% for this," and things like that.
Um and then I also think question and answer is, is going to take a huge step forward. There's in insurance massive amounts of files. And I think the ability to just say like "here's a question, apply it to this whole folder of a bunch of junk right, Excel spreadsheets and PDFs and all this stuff, like what's the answer? Read through all this and understand it all and synthesize it and parse it and then answer my question." I think that that domino is going to fall um in certainly the next 12 months. Like we've, we're already seeing GPT-4 is going to take PDFs right, you can say "here's a picture, answer me questions about the picture." And so I, I think we're not too far away from saying like "alright here's a giant just data dump of, of junk and like answer all my questions."
Jake Aaron Villarreal: Um yeah that's great. You know you're in a unique position, really a bird's eye view because you're working with a lot of uh agencies, a lot of insurance companies, mom and pops. What... looking and understanding how they operate today, do you think it's a, a good business to be in? Like if you wanted to start an insurance company and use your technology, how easy would it be to do that? To take a, a market and say "Look, I'm going to focus in you know this part of the US and this is going to be my niche. I don't know anything about insurances yet, but I know that if I have the right systems I could kind of learn who my target audience would be and start selling the technology like, or selling the service." How, how, what's your, what's your pulse on the success of those kind of smaller companies?
Cameron MacArthur: Absolutely. So we're actually seeing the rate that small programs and captives are being formed increase. So more companies are making the decision to self-insure or you know spin, spin up a captive or risk retention group or a small program than to go to the go-to-market. Because A) costs are ballooning and a lot of go-to-market insurers are just pulling out, right. You've seen what happens in Florida where more and more insurance companies are just like "alright, we're, we're just leaving," right. And that opens up a gap in the market for more and more small insurance companies to fill the space. And we're you know focused mostly on the commercial side more than the personal lines side, but one tends to follow the other, right? If you're you know, if a company's pulling out of Florida and not insuring any personal homes, it's not like they're enthused about insuring the buildings standing next to those homes right. Um so I, I definitely... that, that is currently happening and I see it continue to happen.
And you know that's sort of our whole MO right. We have, we have tiny little startup insurance companies that are getting started on AI Insurance for 89 bucks a month.
Jake Aaron Villarreal: Wow.
Cameron MacArthur: Because maybe they're in a position where they say "Okay, um I don't know, there's a, there's a gap in the space. People don't understand the professional liability associated with you know uh building guitars," right? Like "oh you make guitars by hand and you know it's really expensive for you to get you know like general liability insurance for your store because people are playing the guitars and you know go-to-market insurance think they're going to drop them or something." And maybe you say "well I know the you know the people that own like 500 of the 2,000 guitar stores in this whole state, and so I've got you know really good look at the market here. And I can start up an insurance program you know, go get fronted by a carrier and say 'okay, I'm going to offer insurance just to guitar stores' right? And we'll offer them the whole suite and we're you know just for guitar stores and we know it well." And you know you would start up an insurance program in that case. And you know that's, we, we see a lot more of that happening. And you know that's kind of right down the fairway of, we would say "hey yeah go talk to AI Insurance so that you don't also have to build a whole tech company to be able to run that insurance."
Jake Aaron Villarreal: That's great, that's really cool. Wow, I love it. Well let's, you know let's kind of wrap up here. Um really excited about what you're, what you're creating. And, and love to check back in with you in 6, 12 months and see how things are progressing as you execute your plan in 2024. I'm going to ask you three questions. This is just three simple questions with three simple answers. So first one is where do you go to think big or to brainstorm?
Cameron MacArthur: Um I do it in the middle of the night. I often wake up or maybe you know kind of in that pseudo wake-asleep state, and I have a little notepad by my bed and I'll often kind of have all my ideas then and write them down. Um I also do a lot of voice memos on my phone then. So uh my, my better half is you know, she's a heavy sleeper but is plagued by me getting up in the middle of the night taking my phone, going down to the guest bedroom, recording a bunch of voice memos that I don't remember any of them. And then you know, wake up in the morning and I'll type them out and dictate them. So I'd say that's where I do like the big, the problem solving. Um when you kind of like step away and aren't thinking about anything then some of the answers come through.
Jake Aaron Villarreal: That's great. Wow. What advice have you gotten from another founder that's been priceless to you that you would like to share with others?
Cameron MacArthur: Um when I was trying to decide whether or not to start a company, I asked um Cai uh, the founder of Ironclad. And his response was "Don't do it." I was like "Wait what? Cai, I'm asking uh like I'm a potential founder asking a founder if they should start a company and your answer is 'don't'?" And he said "Yeah. It's, it's so hard and nobody appreciates how hard it is. And you shouldn't do it unless you have no other choice." He was like, "If you think there's anything else that you could do where you would just be happy, go do that thing. Only start a company if you think there's no other option for you and you couldn't be happy any other way." Um and I think that really codified the decision for me because it helped me realize like "oh there, there is no other option. I can't, I can't do anything else but this and you know wake up and feel like I have you know happiness or, or purpose." So, dark message but, but a good one.
Jake Aaron Villarreal: Yeah, really cool. Um what do you do to stay positive in the roller coaster ride of a startup?
Cameron MacArthur: I think keeping positive people around you is important. And yeah, don't sweat the small stuff. As time goes on all the things that you used to freak out about you know one or two years ago, you realize how trivial they were. And then that starts to give you perspective of you're like, even the horrible things that happen you go "oh well you know, I'm sure in three years this, I'd all look back and think of this as a little minor problem." Um, perspective.
Jake Aaron Villarreal: I like that. Yeah, really cool. Um if somebody wanted to find you or find AI Insurance, where would they go?
Cameron MacArthur: Website's aiinsurance.io. Um and yeah, I'm on, I'm on LinkedIn. I can you know, you can cameron@aiinsurance.io. Maybe, maybe I'll regret that one, I don't know. Cold outreach.
Jake Aaron Villarreal: Yeah, I like it. Um very cool. Well you know, appreciate you uh Cameron coming on today and sharing your story. And for all the listeners, thanks for listening, it means the world to us that you spent your time with us. My name is Jake Aaron Villarreal and excited to sit here today with you now, and look forward to our next episode. Until then, take care.
Before we wrap up, I want to give a big shout out to all the entrepreneurs that have joined to make this podcast possible. And for all the listeners for listening, it means the world to me that you chose to spend your time with us today. I'm your host Jake Aaron Villarreal signing off for now, but can't wait to connect with you all soon on the next episode. Take care.
This show is sponsored by Match Relevant, a company that helps venture-backed startups find the best people in the market. And they do it in three simple steps. First, they sit down with founders to understand their story. Second, they tell their story into multiple candidate channels. And third, they schedule interviews within 48 hours. Find us at matchrelevant.com to learn more about how we do it.
.jpg)
.webp)
.webp)
.webp)
.webp)
