Using Machine Learning for Property Assessment with John-Isaac “JC” Clark from Arturo

If you’ve ever taken out an insurance policy on your home, you’ve likely had to answer a long list of questions, many of which are difficult to answer. Do you know how many feet your house is from the nearest fire hydrant? Do you know what type of shingles are on your roof? But answering just one of these questions incorrectly could end up costing you when filing a claim.

John-Isaac “JC” Clark offers a solution with his company, Arturo, which uses aerial imagery and machine learning to analyze your home.

With their technology, insurance companies are better able to understand the assets they’re protecting and construct the best policies.

More information:

Arturo is a deep learning spin-out from American Family Insurance relentlessly committed to delivering highly accurate physical property characteristic data and predictive analysis for residential and commercial properties for use in the Property & Casualty (P&C) Insurance, Reinsurance, Lending, and Securities markets. Leveraging the latest satellite, aerial, and ground-level imagery, as well as unique proprietary data sources, Arturo’s deep learning models provide differentiated property data unparalleled by any other provider – often in as little as 5 seconds.

John-Isaac “jC” Clark 0:00
You know, we did all of Australia’s residential property and 48 hours last July, just like crazy, right? You know, imagine saying you’ve got every image of every property Australia and you have, you know, less than a second to give me 35 attributes. And in doing that, like over 95% accuracy, as a human would be just like, absolutely impossible.

Alexander Ferguson 0:26
Welcome to UpTech Report. This is our applied tech series UpTech Report is sponsored by TeraLeap. Learn how to leverage the power of video at Today, I’m joined by my guests, JC Clark, who’s based in Dallas area, and he’s the CEO at Welcome, JC could have you on. Good to be here, Alexander. Thanks for having me today. Now, Arturo, premier website is an AI powered platform that derives property insights and predictive analytics from aerial and satellite imagery. So you guys are focused on on enterprise clients have been on the insurance space, specifically taking satellite imagery and being able to pull fascinating data and analytics, but also looking going into the overall real estate, real estate industry as well that I get that correct of where you guys are focused on.

John-Isaac “jC” Clark 1:13
Yeah, that’s absolutely correct. You know, we’ve really been able to build a very strong base, both in the US and abroad with insurers. But as we’ve done that, and we’ve derived these property insights and information from images versus having people go and attain them, which is lengthy, time consuming, etc, expensive. We’re also seeing an interesting demand from others that, you know, deal with those residential properties. Namely, the lenders, saying, hey, how could we use this information in new and novel ways to improve our business just like insurance are doing as well.

Alexander Ferguson 1:44
In a nutshell, if you had to just paint a picture of really what this this pain this concern that the biggest issue that your, your, your clients are facing right now, what is that?

John-Isaac “jC” Clark 1:57
So the it’s probably useful to share a little bit of background, you know, when you think about an insurance policy for a home, which is where we got started as a spin out from American family insurance, when a homeowner wants a quote, when your policy is coming up for renewal for your home, after you’ve got a policy, there’s information that insurers really need to have to accurately quote your policy. So you don’t go gosh, this is ridiculously expensive. Or go, whoo, I really liked this policy, it’s very cheap. And when you go to file a claim or have an issue turns out, maybe something about your property that was necessary, didn’t get read properly reported outside cover. And the way that’s traditionally happened is very painful. One is as you as a consumer, they’ll say, Alexander, you’re this home you’d like to insure? Here’s a list of 80 questions about that home? Um, you know, what’s the exact square footage? How many roof panels do you have? I’m not sure if you know, a roof panel is but most consumers don’t? Um, is it architectural shingle roof? Or is it asphalt shingle? Or is it three fat? I’m not sure if you know the difference between the three, I certainly didn’t before. But these are the types of things that insurers need to know about physical property. And it is very arduous and time consuming for consumer to do a accurately obtain or know those bits of information. And none of us want to wake up on a Saturday morning to get insurance policy and fill out a 60 to 80 questionnaire question questionnaire. So that’s kind of the problem we’re solving. The other ways that this data has been obtained is let’s say an MLS property record from when the home sold. One interesting thing about data like that is it may have been right then, but it might not be right now, because homeowners, like I’m actually doing at the moment, are remodeling their properties, adding pools, fences, you know, doing different things, so that data might be outdated. And then the last method of getting this information is really time consuming and expensive, which is a property inspection or an insurer. Just like in a lending transaction, we’ll have someone come inspect the property insurers at times do that when they’re unaware of whether or not the property is actually what’s there is supposed to be there. And that is take several days, two weeks to obtain. It can be anywhere from 35 to $350, depending on the size of the property and style. And guess what, even humans when they come visit a property and our experts domain mistakes, and so you miss things you might not accurately measure write something down wrong. So that was really the fundamental problem Arturo is solving for insurance is taking a recent image that we don’t collect but aerial imagery providers collect like those itself in Google and being an Apple Maps, right and other mapping technologies we pressings consumers, satellite imagery, companies, like max are where I was formerly head of commercial product before I joined Arturo CEO that take these imagery, this these images of properties. We then go and win ensure as a consumer, like Alexander saying, I want to quote or my properties coming up for renewal. What our customers will do is they’ll reach out to our API, and we will then fetch the latest images of Alexander’s property. And we will run those through a series of very proprietary machine learning models that we developed in concert with American families data, here’s the data. And then we’ll provide highly accurate information about Alexander’s property to our customers in seconds. And in a structured data format to like think of almost like an Excel spreadsheet, a CSV, a structured data response, which means it’s in the same format, they would have got it had you typed it into the form where they pulled it from an MLS record, they digitize that inspection report so that they can go Okay, you’ve got 22 roof panels, and your roof is so cheap, there’s no trees hanging over, we do see a pool. Good job summer. Now I can properly quote your policy. And again, we’re doing that in seconds, for around $1.50 to $2.50 for property, versus again, that 35 to 350. That takes weeks and hours to obtain and expensive. Two seconds.

Alexander Ferguson 5:46
Wow. It’s both speed and efficiency on on multiple levels. Now, all this ability, though, it sounds like is you’re pulling from that other company have all the insights and data and being able to train the machine learning. Well, how was that difficult? Like, take me back now help me understand the story here of how our turtle came into existence?

John-Isaac “jC” Clark 6:06
Yeah, absolutely happy to show at you know, one of the really interesting things about our Turo and having at one of my past startups got to work very closely with Google when Google purchased a little company called keyhole, which is now the product we all know and love is Google Earth. And I had this really fortunate opportunity of working with Google’s geo team around directions and Street View and Google Earth and maps for iOS and things like that. And it really blew my mind that there was this amazing amount of learning we all got as consumers from these little things that are hand telling us where to go and what the front of a building look like that we were trying to find well, driving, etc. But from a consumer perspective, we were transformed in our personal lives. But from an enterprise perspective, they’re kind of still in the stone age’s. Right, they didn’t get the same kind of awesome thing that we all got is individuals, of course, as individuals work in enterprises. And so you know, oftentimes I hear like, Why can’t my business like have similar style of information? Now Google got, you know, had billions of dollars of investment literally, to to launch Google Earth, Google Maps, etc. And it’s a very successful property for them as a product today. And so what I really, really thought there was, this would be great if like some industry could really be disrupted by machine learning remotely since data of properties, which is, of course, how Google built Google Earth and Google Maps, right? But maybe not have, to this extent, a couple of billion dollars. So I was looking for, in my career, an opportunity to, you know, say Where can I apply the similar style approach, but to an industry that’s ripe for disruption. And while I was at Digital globe as a commercial product, looking for a place to be disruptive, American family insurance approached me and said, Hey, we’re using one of your machine learning products on satellite imagery, we’d love to show you what we’re doing. And ultimately, I was just blown away by what they had done, which is they had developed about two years of research and development around taking satellite aerial even ground level imagery. And in seconds understanding of property, there’s a little limited time is about 12, or 15 kind of characteristics about a property. But it was one of those things. I was like the light bulb moment, I’m like, this is what I’ve been looking for. So really enthused when they said, Listen, we would like to spin this technology out. We think that as huge market applicability to other insurers, we’re already using it, we’re already finding value. And, you know, there’s other industries that probably could be disrupted by creating this type of property information so accurately and quickly, like lindian, like real estate, etc. So ultimately, I joined all those three years ago, August 6, just shy of a month from lb three years. And we spun the technology out and we to your question about the machine learning and training data. We also spun out years of American Family Policy data and insurance history data and claims data and inspection data where those inspectors go to the properties, right? And why is that important? Because it gave us this huge springboard to build really, really accurate machine learning models on all sorts of instances of files of property, and the types of things you might see that otherwise, you’re guessing, right, you’re looking at a Google Maps image or some you’re gonna think that roof doesn’t look great. But whereas we had information like this roof has this issue right now, and there was a person on this date and time, so we can go back to the point in time near that, and the imagery libraries go Yep, there it is. Use those instances to train machine learning models. And that’s actually been really, really awesome because our current customers see the value that was created by American families data, helping improve and make our models better. And then also said, you know, what, we’d like to help improve as well. And so a lot of our customers work very closely with us in some form of bringing their data, whether it be knowledge of what’s happening inside an image that their claims adjusters have or their appraisers have or their underwriters have, or is it time somewhere like here is training data like here’s a bunch of hail instances on roofs like this really happened that helps us like, you know, train models that detect that Say hail damage and solar panels or skylights.

Alexander Ferguson 10:00
Coming out of Max our technologies, if I understand correctly, correct, and I’m trying to understand the connection also to family, American family, American family, it’s like you already had all this access with like a startup out of nowhere wouldn’t have access to all that. So it’s like, immediately you’re off to the races so much further. Is that like, is that was that the cases like you had all the right pieces the right time? And right, now we can turn it into a new spin out and be able to launch it from there?

John-Isaac “jC” Clark 10:29
Yeah, I think I think so I think, you know, one of the unique advantages are true has, instead of, you know, being a group of guys and gals in a garage going, wouldn’t it be cool if we tried to solve this problem, which I’ve been one of, at least the guys in the garage doing at some point in my career? You know, this was a business, a very large fortune 500 entity, one of the oldest, most respected insurers united states that was looking to innovate. They were their question was, you know, this data is old, it’s outdated. It’s often inaccurate. It’s expensive. If we have to send a human. How could insurance change if we had accurate up to date information? That was the business problem that really spawned this? So they said, Well, okay, how could we get information that’s highly accurate up to date, when they’re like, oh, machine learning, it’s this new thing, because this really technology activity started. I joined in 2018, it had almost been three years in existence inside American family from the original inception. Right? So it’s been three years working on this. And machine learning, obviously, in 2014, you know, Geoffrey Hinton’s work, etc, kind of started really to take off cloud. providers like Amazon, and Google, you know, Microsoft all started making things like Kubernetes, and machine learning services available. So there’s a lot of bardi, a body of work that was able to help propel that. Plus, they’re like, well, where would we get all this training data, we would need to do this while we have it, we’re unsure. So it was a really unique thing for the business to have, be inside a customer, right? Or be a customer when it was founded. Because that really led us to, are we really solving a problem? No one in an r&d capability is going to be told, you’re not solving any business problems, we don’t see it working, we’re gonna keep throwing money at it, right? People have accountability to shareholders and to board members, etc. So it was really working, though. And then it was like, well, we have something really differentiated here. What do we do with this? Do we just keep it as an internal secret thing for us, or we’re already using it, we could probably on some of the business, which ampamp does have an equity interest, and Arturo noncontrolling for those of your your viewers and listeners, but you know, it’s an opportunity for them to both get the benefit there early first adopter owns some of the business. But also, let’s say American family decides they want to go into a whole new market in I don’t know, LinkedIn, or let’s say another geographic mark in the US, because they don’t write in every state. We will then have worked with other customers who were probably around those use cases, and they can still get the benefits of the product, innovation, etc. versus if it was just internal, it would only be around where they focus in what they do. So ultimately, I think our customers benefit greatly from their background, but in fact, benefits from the innovation we do with other customers in the market as well.

Alexander Ferguson 13:07
You paint a picture that makes it seem just all the dots are connecting, everything’s lining up, everything’s perfect. Was it really just a smooth ride the last three years? Or were there any major challenges you had to overcome?

John-Isaac “jC” Clark 13:23
So, gosh, you know, I think the proverbial joke, it proverbial joke is the duck on the water, right? Looks like it’s just gliding along and everything’s smooth, but underneath its feet or go like crazy, right? No, you know, it was really interesting and created certainly some first time crew challenges for me to look at how do you spend electrical property and a bunch of proprietary internal data out of a large insurer on into a startup? Right? I never had to tackle that type of threat transfer, intellectual property issues, etc. There was a course then, you know, you know, how do we go from being an enterprise capability inside the enterprise, to being a SAS capability that serves multiple enterprises concurrently with SLA and uptime, and all those things? Oh, and authentication and tracking? Because when you’re an internal capability, you know, you’re you’re just an internal service, right? You’re not held to that now, I’m going to show our product to another major insurer, right? There’s expectations on security, etc. So both of those two things were certainly unique challenges. The good news is, you know, the the great team that I’m humbled to leave here at Arturo has an immense background both in insurance and imagery and machine learning. We’ve got some of the top talent in the market, I couldn’t be proud of our team. And they’ve really been, you know, really instrumental in helping overcome and work through those things. So yeah, I would say up until our series a there was a lot of focus on the tech transfer on spinning up a team that helped build this into the business right into America, from American family, largely from the data science and analytics lab or DSL into the some of the business. But those are some of the you know, joys even If you’ve come from a fortune 500, to do a spin out versus a startup, you still are gonna have some fun and some challenges. It was just a different set of them than I would typically had typically encountered. Like going from two people in a garage trying to like build something was like, Oh, we got nine people, we’ve got big customers. But now there’s like all these lawyers, intellectual property, things, we got to figure out, how do we like harden this enterprise. So they were fun to get through. In retrospect, at the time, they were obviously like any challenge, lots of frustrations, but we got through them and ultimately think the market the industry is better for it. If you had to give us a word of advice or lesson learned, for for spin out of the unique challenges that accompany that could take like a tactic that you found that worked well, looking backwards, what comes to your mind. two terms, patience and impatience. And so I think the real reality is, internally, you know, you have to recognize that when you’re working in a spin out environment, with a large corporation, it’s not going to go as fast as a true startup would go. And actually, it’s one of the other reasons American family wanted to spin it out was because the innovation rate they thought would increase if it was outside, versus like, you know, in multiple customers versus just them as a customer and their claims underwriting, you know, pointed quote, people having to be the sole inputs for information. But impatience is important too, as you’re spending something now, because of the different cadence at which you know, a very large fortune 500 operates, not that they’re slow and stodgy. But just, they’re designed to keep doing the thing over and over really, really well. And so that means process of policy, etc. and a startup which is trying to like, go as fast as you can to like, retain customers prove value, etc. And so representing the impatience of like, we need to get this out, we need to get these things done, that you might not in a role inside that fortune 500 is really, really key. Because, you know, with anything, you don’t push on it, it’ll sit there and squeaky wheel gets the grease, but internally, you have to remember to be patient, because you are not just a startup, right? You are dealing with a large enterprise, they do have real legitimate policies and processes and procedures that they’ve got governance and accountability around, you can’t just say let’s Serkan this whole thing on intellectual property. And then someone comes back years later and said it got done wrong, you know, like that wouldn’t be defensible while raising capital, or to let our customers know, we truly now own this intellectual property. It’s can’t go back. It’s, it’ll be inside our turf forever. Those are really important things to get. Right. So patients, you know, internally, knowing that it’s gonna take more time in patients externally going, Yeah, we need to get this done as fast as we can.

Alexander Ferguson 17:42
I love it that you mentioned earlier that you had your series A, you also had a series.

John-Isaac “jC” Clark 17:50
Yeah. Yeah, we were really, really fortunate to announce just a few months back, that we raised a pretty substantial Series B round led by Atlantic bridge, which has as their name might suggest, if you think about it, I focus on helping bridge the Atlantic for tech companies in enterprise tech companies, the us also to Europe. In so as our business has expanded into other parts of the world very, very quickly, Europe as a as a next target market for us. We also have team members in Munich, Germany. So them as a partner Atlantic bridge is a partner was was really exciting for us. Also, as part of that RPS rock, pepper, rock, paper, scissors, ventures, yes, there is a richer company named Rock, paper, scissors. And I know Samantha Wang, one of the principals there assures me that is not how they make investment decisions with the partnership. But they actually have a really big focus on Asia. And a large number of LPs from the Asian tech community, community and American Asian Tech Community show, you know, we see as we’ve worked in Australia, New Zealand, a path into Singapore, Japan and other locations through their partnership, so couldn’t be happier to have both of them. They’re amazing assets to the business and, and obviously big fans,

Alexander Ferguson 19:08
what would you say is is contributed the most to your growth? Like if there was one particular tactic or element in the past few years that has really allowed you to grow faster? What would that be?

John-Isaac “jC” Clark 19:22
You know, just candidly, phenomenal customers and, and I see that for a reason right to um, I’ve jokingly say, I’ve never done sales and all I’ve done in sales, even though I have a technical background, because I think it would just happen like, you know, I enjoy talking to people and I was a technologist. I joke I’m a recovering software engineer. Fortunately, no lines of code at Arturo have ever been touched by me, which our CTO dr Tuttle is adamant stays the case. But you know, in my 20 plus career, your 20 year plus career history before Turo, I engage a lot with customers and wanting to Things that in various roles I observed was, you know, there’s there’s their sayings, like, you know, 20% of your customers will take up 80% of your time, right, and things like that. And so one of the things we worked really, really hard at Arturo was to try to work with great customers, ones that we could truly partner with, that we were going to spend a lot of time with. But we were going to accomplish a lot together, right, versus the customer that may be not truly ready to buy or not truly interested and not willing to invest the time into like, really how this can be transformative. And they may be no fault of their own, they may have a lot of other priorities or initiatives going on. But when we would engage with potential customers early on the business, one of the fundamental things, I was also doing the sales, predominantly, I would ask my team after the engagement is do we really think they’re ready to work with us, because if they’re not, we’re just pushing them to do something that’s not in their best interest, or it’s going to take away from other priorities, and we’re not actually going to help them. And we’re gonna get frustrated, because we really want to make them successful around it. And show we would say, if we couldn’t say that we would not tell them what to work with them. But we would just back away, right? And the ones that were really interested would come chase us and say, Hey, we really want to do this. And the ones that really are interested, we’re like, we didn’t have to chase because they were like, we’re both coming together at solving the problem. So I think that ended up creating a scenario where nearly every one of our current customers were really intimately involved in what are their key goals and trying to accomplish? Are they trying to reduce costs? Are they trying to increase their conversion rate when a customer comes to their website or an agent to get a quote that they’re going to win, the more of them as we’ve made it simple. As it got, we’re going to reduce their loss ratios, insurance, basically a simple, take more premiums, and then you pay out claims, right? It’s a very simple rubric. So that means if you can more accurately gauge the premium for a property versus the risk, it’s going to present your hopefully should decrease the you know, the claim and increase the premium, or overall, what’s your data gathering? So we would always work really closely to say, what are those key things you’re focused on, and can Arturo’s technology help you do it. And then the last thing I’ll say is, now that we actually have a sales team, which I’m thrilled, we have this phrase, which is get to value quickly, or eliminate the potential to get to value quickly, meaning, if we’re not the right thing, that’s all what a customer is trying to accomplish, we want to tell them, we want to find out as quickly as we can, and then help point them to another, you know, tech company that can or just let them go find it instead of trying to like become a solution for something we’re not really great with. And I think that commitment really resonates with those potential customers, that if our true wasn’t right, for the thing they’re trying to solve, now, they’re gonna remember that, you know, we help get them to where they needed to be. And hopefully, when they have a problem we can solve they’ll come to us first,

Alexander Ferguson 22:42
helping people understand your product, the technology in it, is it easy? I mean, like, is it? Is it very straightforward, they get it right away? Or how have you made it easier? In the messaging?

John-Isaac “jC” Clark 22:56
Um, you know, that’s a really good question. I would say that, anecdotally. It’s easy for them to understand that we are somehow machine learning, which is like a big black box. And you know, even as a technologist talking through like, all the frameworks, every single thing we use per model is something that’s reserved for like our head of AI. Dr. Moody. She’s amazing, right? You to a customer, I think it’s kind of like, you know, he, you know, you’re looking at Google Earth, you’re looking at Google Maps, we’ve all seen images of her home or something. Imagine if just is if, you know, you can stare at it and see that’s a building machine learning blackbox model. It’s like, Oh, it’s a building. It’s exactly their size. And it’s a rubber remember, people can comprehend I think with things like Siri, you know, Google’s voice assistant, you know, they see video cat videos on YouTube, like, how is it that like you might search for cats always comes up? Well, they know that some machine learning is happening there to classify what’s in the videos, think there’s enough awareness, at least in enterprises that machine learning can make sense of data. And so Arturo has connected, really smart machine learning models to make sense of imagery data about property, and just sucking it out and turn it into a spreadsheet, turn it into a CSV file, that’s your structured data response, I can now plug into a business system and say, where Alexander said, his riff type is asphalt shingle, and that’s a asphalt shingle comes in, I can map that to our true asphalt shingle coming in. And now I don’t have to ask Alexander and people that clicks, I do think the true nuances of how machine learning works and, you know, huge pipelines to do this, like on demand or, you know, we did all of Australia’s residential property and 48 hours last July, just like crazy, right? You know, imagine saying you’ve got every image of every property Australia and you have, you know, less than a second give me 35 attributes and in doing that, like over 95% accuracy as a human would be just like, absolutely impossible. So I think the more we we put out there, the more The market is aware of how machine learning problems, the easier it is for the story to be told. But certainly inside how the, as they say the sausage is made is is our company, the good news, our customers don’t need to worry about that. They can see how accurate we are, we always say, Hey, give me like a bunch of inspections you’ve done recently. So we know that like, you know, what was there because a human climbed up and around the property. Now let us run those. And then you compare, and you know, almost without fail or customer say it’s accurate or sometimes more accurate, then the humans and certainly are cheaper and quicker.

Alexander Ferguson 25:34
So you’re saying people don’t take care what machine learning is they just have enough of an understanding of analogies of similar things, that that base knowledge is fine.

John-Isaac “jC” Clark 25:44
They don’t they don’t. It’s like the App Store for you know, Google Play or Apple App Store. Do you know what software development is? Probably not. It’s totally something but it makes these apps and I know that these apps bring me value, and I can use them to do things I need to do. So I don’t need an understand. And in fact, you know, there’s a, an interesting expression that I have, which is Arthur C. Clarke, you know, the famous science fiction author had a interesting quote than to any technology. sufficiently advanced is indistinguishable from magic. But I think that in the in that era, that was an interesting statement, right? Wow. It’s like magic. However, technology is so prevalent our lives today that, you know, john Isaac Clarke said, any technology sufficiently advanced is indistinguishable as technology. Because I don’t think people really think like, wow, this is an amazing, this is my phone. It’s like almost just as pizza now, but it does magic for me right now, when you think about it, magically, we’re like, we can talk to people on the other side of the globe, we can view places, you know, far away instantly, we can get directions from here to like Tallahassee. And second, nobody thinks that that is magical anymore. It’s they don’t even think of this technology. So I think, as technology is so prevalent, people just accept that something I do not understand whatsoever, does something I can’t understand. And that brings me value. And I’m gonna use it because it makes my life better for my business. But

Alexander Ferguson 27:07
I love the simplicity of that. Does it just make their life better? That’s all that really, really matters. For you, what can you paint the future of where you guys are headed? Both? Let’s start with this the future of our tomorrow and the roadmap of where you guys are?

John-Isaac “jC” Clark 27:24
So I think, you know, a couple key areas, obviously, you know, our investors ask us these questions is we’re like raising more money. And there’s, there’s kind of three or four key areas I’ll start with the first three is their near term. And the fourth is the aspirational vision, which we hope to achieve over time. But you know, the first is geographic expansion, right? We’re one of the few companies in this space that has proven that we’re not only can we do this in the US very effectively, but we can do in other parts of the world. Right? We’re obviously starting with like first world countries, country level, single family, residential homes, which are some types of cultures, even first of all countries like Europe, like Italy, will a kind of non major urban modern city in Italy will have homes that are all built connected to each other. Right, you know what I mean? Not suburbia, USA style. But you know, there are single family homes in Italy and Germany and other places around the world. So we’re kind of beginning to expand our reach to offer insurers and these other parts of the world, the same types of capabilities. Often, that means different architectural styles that don’t exist in Australia, New Zealand, US candidates cetera. And that’s really kind of our first goal is to be inexperienced bring this value to really all insurers who deal with residential property. The second is commercial properties, right? There are obviously a lot of buildings out there that are commercial. Now interesting stat, there’s roughly 155 million structures in the United States, of that 125 million are residential in nature. So That only leaves about 30 million properties that are non residential, you have what are called habitable residential meaning like a hotel or a condo, it’s not a single family home, but people live there, or or stay there. And then you’ve got office space, etc. So we really kind of want to move towards being able to again, offer our capabilities to commercial insurers understand those commercial properties, so that they can understand who’s there, what’s their structure, condition, risks, etc, just like our residential insurers do. And then of course, lending, right, if you lend on this property, if you’ve given them the loan for it, what you care about what’s actually there, how it’s changing, you know, if suddenly, the property is in pretty bad disrepair, and no one seems to be at it, and there’s a loan on it. Maybe that could be a sign that something bad might happen from a default perspective. So there’s a number of lenders who have reached out to us in the past few months, both in the US and abroad, saying, Wow, how could we use the same type of information for baking making better loan decisions for understanding what’s happening in our loan portfolio for maybe optimizing streamline the loan process and with the real estate market, what it is right now here in July anywhere in the US. And no no stoppage in sight to the fraud. That’s really important, right? Being able to get that house, get that loan in or purchase that property to inventory so low, etc. So we’re excited about being able to expand la geographically but into both commercial insurance in the lending side of both residential and commercial as well. So those are kind of the key three areas that our team is really focused on for the next probably a good two years. I’ll add the fourth though. And I think it’s an interesting one to mention. You know, we fundamentally believe that outer Turo, we want to create greater understanding of physical places and spaces to enrich lives, right? That might mean you are rich, because you don’t have to answer any questions, maybe your life didn’t get majorly better. But that Saturday morning with your coffee, not answering any questions, it made it better to some degree. But in time, our real vision is we go geographically as we do different types of property is that we can go beyond things like first world countries, and potentially be able to facilitate offering financial insurance or lending products into other parts of the world that don’t have access to that. One thing that most people don’t may not be aware of is that lending and insurance really started when there was cadastral. Mapping property maps like this is where Alexander owns a piece of land. And here is what he hasn’t started from taxation, largely in kind of the UK and, you know, hundreds of years ago, and without that, banks and lenders, you know, and insurers don’t know like, how do I What am I insuring? Where is it out? Well, now with capabilities like Arturo, if we think about expanding into like, you know, parts of Latin America, or Africa or Asia and underdeveloped areas, could we Serkan navigate the need, just like Africa went straight to cellular instead of telephone lines? Could we circumnavigating need, or being able to have people go out and you survey property, but instead use these remotely sensed images to understand what is there and then enable maybe micro loans, micro insurance products in our parts of the world. So ultimately, I think our hope and goal is for moving beyond where we are today in order to really, truly enrich lives and parts of the world that aren’t there yet. That’s aspirational, it’s years away. But it’s something that we think about a lot here as a bigger vision for our business and the team than just helping lenders and insurers, like be quicker. So their consumers are happier.

Alexander Ferguson 32:19
Are you JC like starting being in multiple companies that you found it or being part of those startups and being in this technology world? What’s driving you? Like, like, what gets you up in the morning, and that just keeps you going? And you’re excited about coming from where you’ve been? And where you’re headed? What’s driving it?

John-Isaac “jC” Clark 32:38
Yeah, that’s a great question. Um, you know, I think two key things. It’s always hard, but getting to work with phenomenal people. I just take a lot of joy and pride in and, you know, having such an amazing team and the privilege to lead them, it’s, it’s awesome, right? This is my first time as CEO, the first time that I’ve made plenty of mistakes, and the first time it’ll make a lot more, but working with, you know, a great team, solving those problems, overcoming those mistakes together is really great. Same goes for our customers, right? We are got world class customers that have some phenomenal people to work with. And that just makes it fun. I’d say, you know, the other thing, though, is being able to do something that’s really transformative and disruptive, right, I was really privileged to like, be exposed to the Google Earth and Maps team, you know, Ken, gosh, you know, more than 10, almost 15 years ago, I can’t believe I’m saying that, like, wow, you know, what I was kind of in my really early 30s, late 20s. And I was just blown away by like it was, you know, our lives change, like Uber doordash, you know, so many things would not exist without the ability to say, here’s an address, get me there, or route me around the, you know, in just our awareness of our planet through kind of the geo literacy that I looked at Google geo team helped create. But I see a lot of industries that just don’t have a comparable amount of groundbreaking change and helps them and so doing something like what I’m doing in our terone on the team here is what is working around, is what really exciting is truly changing the way an industry can function from all outdated information to long questionnaires, or checkboxes informed literally still exists, or people crawling around on roofs, which is dangerous is time consuming is with COVID. Lots of people didn’t want people coming in and around their homes, right. And then all the pandemic right. So I think just doing it different. And enabling a way to do it different is what you know, I really am motivated by a lot.

Alexander Ferguson 34:33
I love it. Curious for you, as as you said this first time being a CEO button that you’ve led teams before. If you had to think of maybe one of the most difficult times as a leader that you were able to learn from a What can you share his insights as a leader, whether it’s another CEO or leader, that you gain some insight out of that trouble?

John-Isaac “jC” Clark 34:55
Yeah, I mean, there’s a laundry list log in It was a bunch of other probably filler. I think probably one of the one of the best things I can share. And this has been on my mind a little bit more recently, as our team has grown so substantially, we were, I think around 30 people at the beginning of this year, and we’re like 75, today, just barely six months into the year. So it’s like, Wow, so many great people. And I think one of the things that any leader needs to do more and more of is if you’re hiring other great people in the business to truly enable them. Like I used to think my job at the beginning of Stanford company was to say, this is what we need to do today, this is what you need to do like this, what we get accomplished. So it’s more directive, right? As you grow and scale a business, you’re bringing in other great people, they know like what Customer Success is supposed to be our VP of customer success is amazing. Our VP of sales is amazing. Like, I don’t need to say like, this is what we need to do for customer success today, they know more than I will ever on how to do that function. So the role of the CEO I see as becoming more of an advocate, an advocate for our customers, or future customers, and advocate to remove blockers or hurdles from my team, that I that only I could potentially write to enable them to go achieve phenomenal customer success or bringing on through sales and marketing, you know, fantastic new customers that we can truly drive value for. So I would say that that’s a transition, they’re going from like, people are looking at you as the person to rent to know what they shouldn’t be doing to the people in the room know what they should be doing. They’re executing it, and you’re there to help them right into enable them, obviously, still hold them accountable. That’s part of the role skill, but it’s not directed anymore. If you set the goals, the vision, the goals that will help us get to the vision, and then you’re enabling these people to do their best work versus saying this is the work you need to go do. And that’s that’s that’s a that’s a sudden change, especially as you double your headcount so quickly. But it’s awesome, because you get to sit back and see kind of people make far better decisions and come up with cool ideas than I ever could. So

Alexander Ferguson 37:02
how does that feel? Like? How do you overcome that maybe in the ability where, in some ways you’re losing, not losing control, because you are managing and leading them, but that’s a change in your own shift of how you see your role itself? How did you manage that? How did you feel about that?

John-Isaac “jC” Clark 37:16
Um, you know, I think it wasn’t necessarily a conscious decision as much as well, I think that factors, a few points, or one or two of my key leaders, you know, I’ve worked with very long here. So my original co founders, like, you gotta just let us do this, like, don’t try this, like, my CTO is like you What do you want to like, try to solve this, I know, your former engineer was like, Look, we got way we got great people, right. And sometimes there is that natural inclination, especially me having to start a background for two decades of you know, you all just dive in, you want to solve the problem. And sometimes you’ve got to let go and let the team solve the problem. Because it’s going to take me forever, they’ll know what they know about our technology stack or what the problem is. But it’s hard to sometimes tamp that down. So I’ve learned to when there’s an issue or something to really kind of just trust in the team, sure, press the deadline or the timeline that we need to like be aware of. And by doing that too much, we’ll never actually motivate people, we’ll just add one more distraction to them when they’re already stressed and trying to solve a problem. So I have had some say, my leaders say, Hey, you know, we got Yes, like you don’t say, Okay, I’m gonna back off, you know, I’ll wait for your update in eight hours. But I think another thing is, you know, when you see the great work that the team begins to do, when they’re empowered, they understand the goal, the vision that we’re trying to pick, you accomplish these goals to achieve, they come up with things that you never would I like to joke though, at the end of the day, being a CEO, when we do something great. It is the team that helped us do that deliver that when nothing is wrong, and it’s my fault. So, you know, you still have to be able to look at when a challenge emerges. And okay, this is how are we going to come together and solve this ultimately, right? Our board our customers, our shareholders do look to me to make sure that we’re working through those things. But that’s just part of the job. So it is,

Alexander Ferguson 39:07
love. Thank you for sharing your insight on that leadership. I’ve been kind of closing here on a last question of technology. What innovation technology innovation in this industry, do you predict we’ll see in the near term or long term?

John-Isaac “jC” Clark 39:24
You know, I think it’s, um, it’s interesting and tell you which ones I wish we would see in the near term. Um, you know, machine learning is a phenomenal technology. I’ve been through my career, both the advent of client server at the beginning of my career when people were going from mainframes to like desktop. I’ve been through the web, you know, kind of one data to data. I’ve been through the mobile explosion I did a lot in mobile in my career, from a technology perspective. And I really think this is kind of truly, you know, one of the most transformative technologies I’ve ever I’ve ever had the privilege to work with, and I think more transformative, potentially at once. came before it, though couldn’t have done without the ones before. But I do think you know, the challenges of doing things like one shot learning, right where you don’t have to come up with 1000 examples across all geographic areas and different types of things. In order for the machine learning, mom’s just talking about what we do, to know that, you know, that benefit of having American family data, having customers give us access to their data to make our products consistently better is because when you add more machine learning training data, the model gets better, right? Like, that’s just basically how it works. There’s no secret machine learning scientists sitting somewhere tweaking convolutional neural net params with zero training data, and it magically now can tell you what type of roof you have, that just doesn’t work that way. That said, One shot, deep learning that would enable us to be able to have less of those and models, able to kind of detect them would be amazing, because we will just do a lot more for our customers. I think the industry in machine learning feels very similar about that. I think the second thing and then I think another thing we’re kind of passionate about here is when we think about today, we provide physical property characteristics. Like it is a single family home, it is a one story it is a asphalt shingle roof or, you know, concrete tile are now going into like conditional, okay, the roof is concrete tile, but 22% of the tiles are chipped or damaged, right? to then go. The roof is this quality, which is kind of subjective versus objective, like objective, it’s a rough objective isn’t to conquer subjective. And it’s quality is this is something that, you know, we’re working with our customers a lot on, everyone has a different way to define good or bad or okay, right. And so you have to be able to kind of create framework. But I think ultimately what the market really wants, isn’t data points, if they make decisions, or they want to be told, like, let me know, anytime a home becomes bad, or is trending back, right? I don’t want to miss a count of the data points, whatever I just want to know be told like this property or these properties have issues emerging. I think that is kind of an interesting technological advancement from a machine learning perspective where we could apply higher level machine learning style models to underlying history about like, when a home ended up having an issue, what were the things that happened before it, can the machine learning models, learn to detect those things and then tell us, hey, we’re seeing times that are indicating This home is going towards this type of issue that I think can be really, really exciting. But I do think is still a year to offer, at least for us to be able to implement but um, you know, you got to look, you always got to look to the future, in order to understand kind of how you’ll take advantage of those types of technological improvements when they begin to emerge.

Alexander Ferguson 42:40
Thank you, JC for paying the future of what it could be where it will be also taking us all the way back from from the beginning of where you guys have been and where you’re headed. For those that want to learn more you can go over to that’s Thank you so much for your time, JC this great conversation.

John-Isaac “jC” Clark 43:00
Yeah, Alexander is a real pleasure being here. Thanks so much for having me. Always happy to be back a future lead. But um, thanks so much for the time today.

Alexander Ferguson 43:07
Absolutely. And we’ll see you all on the next episode of UpTech Report. Have you seen a company using AI machine learning or other technology to transform the way we live, work and do business? Go to UpTech And let us know


YouTube | LinkedIn | TwitterPodcast

Fostering Relationships, Forecasting Financials with Brandon Metcalf from Place Technology

Using AI for Better BI with Ajay Khanna from Tellius