• Arturo, a spinout of American Family Insurance, uses artificial intelligence (AI) and deep learning to evaluate images of buildings and assess risk factors.
  • The technology makes the underwriting process faster and more accurate.
  • Google Earth and Google Maps offer untapped value for businesses that need information about a certain area and don’t want to lose time and money by making in-person visits.


Humans have been insuring things ever since we had things to lose. But not all insurance companies are stuck in the past.

American Family Insurance opened its doors in 1927. But the company isn’t letting its legacy business stop it from nurturing new technology. Spinout Arturo takes something we’re all familiar with — satellite images used by the likes of Google Maps — and combines it with emerging technology. 

“We take images from space and from the ground and use deep learning to understand better and faster than a human could what’s at that property, what’s the risk, what are the damages, what are the attributes?” CEO John-Isaac Clark (known as jC) explains. Getting a machine to make the tough calculations means a quicker and more accurate underwriting process.

While the technology is powerful on its own, its capabilities are dramatically increased when used in conjunction with American Family’s decades of data and experience. 

jC joined the podcast to talk about committing the founder’s cardinal sin, what brought him and American Family together and why a future where drones fly past your window every 30 seconds could actually be a good thing.

John-Isaac Clark, CEO, Arturo (Sam Fiske/Technori)

Interview Highlights — John-Isaac Clark from Arturo

How committing entrepreneurial heresy paid off

I did a bunch of tech stuff and ultimately started a company with my brother in 2006 right after Google bought Keyhole. We worked extensively with Google. 

This is gonna sound like heresy as an entrepreneur, but I was like, I’ve done so much of the startup stuff: my entire career has been nothing besides startups. I want to work in a bigger environment. What’s it like to have a $25 million budget that you just get — that you don’t have to raise? 

The other thing I was really passionate about was seeing Google starting to use deep learning and machine learning on a decade’s worth of data to create really interesting products for consumers. I was like, what if you could do this for businesses? 

Ultimately, I got the opportunity at DigitalGlobe (now Maxar), the world’s largest commercial satellite imaging company. They said, what if you had the opportunity to build out a data science and machine learning platform on 17 years’ worth of images of the surface of the planet as head of product? It was applying the thing that I thought would be the next step forward, plus it fit my desire to see what it was like to work in a big corporation. I ended up as head of commercial product at DigitalGlobe. Incredible team, awesome experience. 

American Family was a customer of that platform. They loved the fact that I’d had a startup career, had worked extensively in location-based analytics and at Google. But also, they’re a Fortune 500 company, and because I’d spent a couple of years at a big corporation, I knew what a profit and loss statement (P&L) was, what a balance sheet was, knew about revenue projections and had worked with intellectual property (IP) and licensing issues. They said, let us show you what we’re doing. I was blown away.

Where you see the roof of a house, this technology sees a risk factor

When an insurer prices or underwrites a property, they are sometimes staring at Google Earth or Google Maps or looking at a house. The problem is, people’s time is expensive. Also, humans aren’t that scalable. 

What we do is take those same types of images from space, from the air, from the ground and use deep learning to understand better and faster than a human could what’s at that property, what’s the risk, what are the damages, what are the attributes? People put in pools, they build sheds, they add an addition. They also don’t take care of things sometimes. 

We’ve taken anything that consumers or business people might look at on Google Earth or Google Maps and turned it into something that a business system can turn into a price, a risk, an underwriting. 

Historical experience + new tech is where things get interesting

The world’s evolving very rapidly. In five to 10 years, we won’t even care about satellites because there’ll be drones buzzing by your window constantly. We’ll be able to look at your roof every 30 seconds if we care to, because someone will be passing it. There’ll be autonomous vehicles going by.

We take all of this new and emergent data, plus unlock data inside our insurance customers. It turns out history is more useful for creating better predictive models about what’s really going to happen, what’s the risk? When you fuse those together, you can start doing really novel and unique things. 

It’s creating an opportunity to operate on a smaller margin and to be highly segmented. You know exactly what the customer that’s really valuable looks like so you’re going to find them.

Move carefully and build things

Being a startup spin out from a corporate entity is extremely different. The benefits are that you have tools at your disposal I could’ve only dreamt about at startups I’ve done before. It’s amazing to have the backing of a Fortune 500 and their financial access and brand recognition. 

The challenge is startups love to move fast and break things, but as a Fortune 500 you care about your reputational integrity. You care about your brand, about how you use your data and how that’s perceived. And so that is a balance: how do you still move fast and get all these things done? 

To do anything in 18 months as a Fortune 500 is like lightspeed. August 6, 2018 was my first day and we had zero employees. By February 14, 2020 we’d closed our Series A with strong revenue growth, a great pipeline and a team of almost 25 people. That’s insane. Very few startups can say that. So when I think it didn’t go as fast as I wanted, it actually went really fast.

Find this business interesting? Check out other interviews with similar companies: