The Center for Data Innovation spoke with Ben Schmidt, PhD, co-founder and CEO of RoadBotics, a Pittsburgh-based startup that transforms visual infrastructure data into meaningful maps using AI. Schmidt discussed how local governments can automate road inspections and make better decisions about road maintenance.
This interview has been edited.
Daniel Castro: What was your goal in founding RoadBotics?
Ben Schmidt: There were a couple. One was purely on the technical side. In terms of my background, I did a PhD in bioengineering, so I’ve always been interested in the more technical side of things. I took a little foray into energy, but stayed in the data science/machine learning/machine vision world. RoadBotics is a spinout of Carnegie Mellon University. There was a lot of autonomous vehicles (AV) research at the time, and we were using the same technology, but to understand something totally different, which was infrastructure.
I’ve also always been curious and interested in how local government operates. When I was growing up, my dad was the mayor of our town and a board of trustee of our town. For a few summers, I worked with our public works team. So I was very much aware of how that all worked and operated.
I was really curious in the fusion of those two things. You can use some really advanced technology and apply it to an area domain that doesn’t see a lot of very advanced technology and really do some magical things.
And I think we have in the last five years. We’ve now worked with around 300 different governments. And really advanced technology in an area that largely doesn’t see these kinds of deployments can make some rapid progress and really change things.
Castro: How does automating the collection and analysis of roadways and other public infrastructure help governments?
Schmidt: A lot of the infrastructure discussion is at the federal and state level. It’s highways, it’s bridges, it’s really big things that cost lots of money. Those are always going to be the sexy, interesting, exciting things that everyone’s talking about. But the practical reality is that about 80 percent of the road infrastructure in the United States is owned and managed not at that level, but by local towns, cities, counties, et cetera. Most of the infrastructure that you use is not in any way reliant on the federal government or state government.
This means that most of the decisions for that infrastructure—”where are we going to fill potholes?” “where are we going to repave the road?” “when are we going to fix water main breaks?”—all that kind of stuff is decided at a very local level. And then similarly, your tax dollars that are going to repave that road to fix those potholes are similarly spent at a very local level. Much of the decision-making—the sort of decisions that determine why America’s infrastructure is in such a poor state—is actually a local problem.
There are 19,000 municipal governments in the United States. It’s a very distributed problem. And so the typical way that all of that’s taken care of today is you jump in a truck, you grab your notepad and your pen, and you drive around and look at what the roads look like and make some decisions. And then spend a million dollars to repave that road.
Despite what people might wish to believe, that’s the reality at every level of our government. That’s how infrastructure is really done. It’s all subjective.
Why is one town’s roads in better condition than another? Funding helps. But the other part is that maybe that town just has a little bit better process than the other one.
The most exciting part with using a data-driven, technology-driven approach is that by creating this objective data about roads there’s no bias and subjectivity in it.
You can really level set across all these different environments, all these tiers of government, and create a place where it really does come down to budget. It’s how much are you spending rather than how skilled you are, how much you’re paying attention, what kind of policies you have, et cetera.
A data-driven approach can really create a big shift in how we think about infrastructure.
Consider budget. Where to spend your dollars is the most effective way to make them useful. One of the great ways that we squander a lot of money in this country is that we go and pave a road and then the water utility or the gas utility comes in and it’s all wasted. All that money went right out the door.
Another is communication. A data-driven approach changes the way governments operate. The typical process is jumping in a truck, collecting some data, and then asking how to spend the $2 million budget. People say things like “I think the road on my street is terrible” or “this person told me yesterday while I was at a grocery store that this road’s in terrible shape.” All of a sudden that entire process is translated over to a color-coded, objective map. Now they can have an open discussion in front of a map saying “Here are the roads we’re going to fix next. And then in year two, we’re going to pick these. Why? Because they’re the worst.”
So now citizens are calling into those governments and saying, “Hey, my road is in terrible condition. Why didn’t you fix it?” And they can say “Yes, your road is terrible. Let me show you one thing though. Your road here is not nearly as bad as this road over here. And that’s why we picked this one.”
Objective data really elevates a lot of that discussion. It’s a huge change in how they approach a lot of these problems.
Castro: Let’s talk about accessibility. How can cities and municipalities work with RoadBotics to monitor and improve sidewalks to address mobility concerns?
Schmidt: Most of what I’ve talked about so far is our roadway product, which is what we’ve had since the beginning. It’s all about assessing road conditions. Last may we launched an Agile Mapper product, which is what we call our second product line. That looks at infrastructure in a broader sense, such as sidewalks, curbs, traffic lights, streetlights, all the other things governments care about.
One of the jokes that I tell a lot is that I can pick up my phone and I can find any Starbucks anywhere on the planet. But when your fire department comes out to your street and there’s a fire on the street, they have to go and manually look for all the fire hydrants.
Most towns don’t know where their fire hydrants are. They do not have a map of it. So your question about accessibility around sidewalks. The first problem most governments have is they don’t actually know which sidewalks they have. They don’t know where the curb cuts are. It’s not because they’re not paying attention—I’m sure in some city record they have the blueprint of where they put in a curb cut on the sidewalk when they built it—but it’s not comprehensive. They don’t actually have a good understanding of it.
One of the first questions that I can start answering immediately and quickly is the inventory problem: “what do I have right now?” And then as we move forward, there’s a lot more we can do. What condition is it in? Is it the right kind of asset? Is it the right kind of infrastructure? All of those layer on top, but broadly speaking, all the governments we talked to, they do not even have a map of their roads.
The first problem we solve for most of our clients is giving them a digital map of the roads they own. And then we can tell them how to survey them and get their conditions.
So to me, that’s the huge movement. We’re going to see this kind of digitizing of physical assets really unlock a lot of new capabilities.
Think about it like a small town in America, with maybe a hundred miles of road. It would take days, potentially weeks, of a full-time job to go and walk every single one of those.
The idea of figuring all this out quickly with technology really catapults what you can do with it.
And what’s exciting is there’s just so much potential.
Castro: Last year you published a report on the state of roads in 20 U.S. cities and released the accompanying data. What did you learn in this study?
Schmidt: Roads are actually in much better condition than what most people think. It’s kind of a psychology thing, right? If you drive to the grocery store and you hit a pothole, you are going to complain about the pothole to all your friends, everyone you meet everyone in the grocery store, and the checkout clerk.
If you drive to the grocery store a hundred more times and never hit a pothole, you will never once speak to anyone that you meet and say “the roads are terrific here!” You’re just not going to do it. We have this tendency to over-remember the really bad things and not so much the really good things.
The life span of a road it’s very long: 15 years, 20 years, maybe 30 years in a few locations.
You have this very long time where it’s in good condition. So certainly doing that kind of survey across so many different cities really showed that there’s this very wide spectrum.
Some are certainly in better condition than others, but in general, for almost every government in that survey, and every government we’ve ever worked with, there’s a huge chunk of roads that are okay, there’s a huge chunk of roads that are getting there, and then there are some roads that are in poor condition. And in general, governments are handling them.
Almost everyone I speak to thinks “Well, my city has the worst roads.” It’s just not true. The reality is we’ll always have some good ones, we’ll always have some mediocre ones, and we’ll always have some bad ones. You just want to shift the proportion over time. And with some data, we have a baseline, and we can be smarter about it.
Castro: Funding for public infrastructure is always a challenge. How do you think the $1.2 trillion bipartisan infrastructure bill will impact the ability of local governments to invest in new technology?
Schmidt: I think comes back to the psychological issues I was talking about before about building a new bridge or highway. People love it. Taxpayers love it. Politicians love it. It gets splashed all over the news. No one cares at all if the maintenance crew was out last week. In fact, the opposite is true. Most of the time when the maintenance crews are out, people are pissed because there’s traffic. Some of that needs to be changed.
I don’t know that we can change the psychology of it, but the funding really needs to be focused elsewhere. Infrastructure is such a long game, like 50-years long.
It’s critical that we’re not just funding new cool projects and then assuming someone else will fix it.
We really need to focus a lot more on taking care of the stuff we have. That’s the easiest way to make it all better versus just building brand new things and hoping that we can outgrow ourselves.
Second, I would really love to see more local governments have a lot more autonomy over how they can spend those dollars. Largely that trillion dollars in the infrastructure bill is going to go to the federal government, to the state governments, and then hopefully get into some of the cities, towns, et cetera.
Most of our infrastructure is in those cities and towns. The highways we’re going 70 miles an hour on are owned by the state, so it is important they are safe, but most of the infrastructure is at that really local level. And so it’s really important that we talk more about how can smaller governments get the tools, expertise, knowledge, training, and funding to make sure that they can maintain the roads they need to maintain.
That’s not the federal government and that’s not the state government. That’s your local government. I would love to see funding opportunities for how to create a nice, disciplined approach to managing that kind of infrastructure.