Robert Lee, founder SubterraAI

It’s immediately apparent that Robert Lee is not a native of Cincinnati. From the east coast of Australia, Lee had been coming to Cincy on and off for about three years when, during his last visit, COVID kept him here for longer than he’d planned. Now he is putting down roots, starting a family and a business here.

Lee’s company, SubterraAI, digitizes underground infrastructure so it can be monitored over time. Cincinnati Future spoke with him about what the company does and his plans for the future.

Tell us what your company does.

Lee: Traditionally, vendors or suppliers sell very expensive inspection systems to municipalities who then get stuck with a $100,000 inspection unit that doesn’t get upgraded. We’ve gone the opposite [direction]. We use consumer technology in the form of GoPro cameras. The municipalities can use 10-20 of them to equip their field teams. They use our platform to then turn that video footage into something much more valuable and quantifiable. We create a 3D representation of that asset. We tag, measure and compare the conditions over time.

You can foresee certain degradations as well?

Lee: Yes. Municipalities are very reactive. They go out and clean and inspect based on problems that are occurring now. They’re not very proactive, because they don’t have the analytical tools to predict what’s going to happen. We use historical data from previous inspections. We introduce low-cost inspection to get more data, and then we use machine learning to fill in the gaps. We can predict, for example, if there is a high probability that a pipe could fail.

How did you get into this field?

Lee: My background is in geoscience; I’m a geologist by trade. I used to work in underground coal mines. I struggled with technical writing, so I took a camera with me everywhere I went. I documented work sites with images, and I put them in my reports. People began to like the image-oriented way of reporting. That’s how I explained what I knew.

From there, people started asking what I could measure from these images, if I could extract more information. I utilized photogrammetry (the science of using a set of images to reconstruct the scene in 3D). Give me a camera, and put me on the ground to take a number of images. I put them through software to then extract a 3D representation of that area.

I was also mapping roadways in 3D to see where convergence was happening. Then I went into train tunnels. Someone introduced me to sewers, and I actually physically walked waist deep in raw sewage to collect data. Then I created an untethered floating drone — I call it the Super Scout — that I can just drop in the sewer to collect the data for me. Now I don’t have to go in the sewer anymore. I always say, if you look at the roadways and see how bad they are, just imagine what’s going on under the ground.

How does the device work?

Lee: Cities are made up of miles and miles of sewers. When it gets to the downstream, tens of thousands of people’s sewers all converge into a major trunk and interceptor service. The device removes the need to have people walk through sewers, and it replaces the outdated technology. Our device can capture one mile of data in 45 minutes, replacing traditional means that [previously] took over two weeks to collect. We were rapidly capturing data at 20, 30 and even 40 times more than the current means they have now, without exposing people to that environment.

So why build the company in Cincinnati?

Lee: Cincinnati and cities around it — like Columbus, Indianapolis and Louisville — are all similar in terms of sewer infrastructure. We knew we could also employ our tech in those nearby cities. We’re located two to four hours away from all of these cities. There’s a high density of the population here in the Midwest, so it helps us scale our solution more rapidly when compared to being located in a place where we’re not very close to large populations.

What are your plans for the future?

Lee: We’re looking to build a mapping solution. This will involve doing geospatial analysis and machine learning to start predicting what conditions will exist 5, 10 and 20 years out. We’re winning contracts now. There’s a lot of interest from South America and the U.K., so we’re looking to build the resources to deploy [our tech] in more places. We’re looking to partner with engineering and inspection firms to get the technology out there. It’s all about growing.