No matter where you are in the world, repairing infrastructure, especially roads, is a constant battle. If
only there was a way to use technology to repair infrastructure in a more efficient way. Well, that day is
finally here thanks to RoadBotics, a Pittsburgh-based startup that works with governments all over the world
to make digital maps of their infrastructure so that they can use tax dollars more effectively when it’s
time to repair roads and other forms of infrastructure.
RoadBotics co-founder and CEO Ben Schmidt joined the Midstage Startup Momentum Podcast with Roland Siebelink
this week. He shared all of the details of RoadBotics, including its product, its journey, and key lessons
learned along the way.
- How RoadBotics spun out of Carnegie Mellon University.
- The unique challenges of working exclusively with governments.
- The surprising benefits of being a company that sells exclusively to governments.
- Why GovTech is a viable option for startups despite its bad reputation.
- The key to finding early adopters in the GovTech space.
- How a small company like RoadBotics was able to expand internationally.
Roland Siebelink: Hello and welcome to the Midstage Startup Momentum
Podcast. My name is Roland Siebelink and I'm an ally and coach and advisor to
many fast-growing startups around the world. One of them is in our studio today.
It's RoadBotics and we have Ben Schmidt, the CEO and co-founder here today.
Hello, Ben. Thank you for dialing in today.
Ben Schmidt: Hello and thanks for having me today, Roland.
Roland Siebelink: You're in Pittsburgh, Pennsylvania today is that correct?
Ben Schmidt: That's correct.
Roland Siebelink: Let's talk about RoadBotics. For those of the audience
that haven't heard of RoadBotics yet, what is it that you do? Who do you serve
and what difference do you make in the world?
Ben Schmidt: Great questions. We are based here in Pittsburgh because we are
a spin-out of Carnegie Mellon University. What we've been doing for the last
five years and change is working with municipalities, governments around the
world to better understand their roads and their infrastructure. The
state-of-the-art today for most governments around the world of managing these
things is pen and paper. We help those governments to create digital maps of
their infrastructure using a combination of low-cost sensors - smart phones - as
well as machine learning and AI.
Roland Siebelink: Okay. Is the goal or the ultimate benefit that you provide
to these governments that there's a higher efficiency, better usage of their
assets? What are typically the benefits you sell to them?
Ben Schmidt: There's a couple of different ones. You've hit on several of
them. One is just more objective understanding. Rather than having the
subjective judgment of which road is in which condition, we create algorithms
and pipelines that do that assessment for them. The other one - as with
digitizing anything - you then have a record of it. You can see it over time. We
can go back and review it, see what changed, and even investigate why did things
change? Why did this road degrade faster than that road? Why is the sign
missing? Things like that.
There's this whole host of benefits of using these kinds of technologies to
further the government's mission of managing and creating services for its
people. And we're at the pointy end of that when it comes to
Roland Siebelink: Okay. When you refer to that pointy end,
infrastructure-related, are you looking at specific kinds of infrastructure that
are more important than others in the overall assets of those local governments?
Ben Schmidt: I've given you two answers. From the inception where we
started - hence our name, RoadBotics - we focus a lot on pavement. For most communities,
if you think about cities, smaller suburbs, rural communities, the majority of
their tax dollars when it comes to infrastructure are going to go to roads.
Everything from basic maintenance, filling potholes, cracks, things like that,
all the way to full reconstructions, which can become very expensive. You're
talking somewhere north of about $1 million per mile of road. That's where we
focus a lot of our time, was on helping them to manage that infrastructure, the
road part of it. That's what I mean when I'm saying that.
Over time, as clients gave us feedback, as we engaged with more governments
across the United States, across the world, we got a lot of feedback about the
other things that governments are managing, the other kinds of infrastructure
that governments manage. And very much a similar flavor. It's a lot of manual,
it's a lot of subjective judgment calls. Can we use some of the same techniques
around mapping, data collection, AI machine learning, to actually further these
other kinds of infrastructure.
What I mean is initially roadway furniture is one way to describe it. Traffic
lights, street lights, signage, things like that. But then even branching out
into other things that governments manage. That could be sidewalks, curbs - if
you walk around any city, any town, and you really pay attention for a little
bit, there's a lot of stuff - where is it? What condition is it in? Does it need
to be repaired? Does it need to be replaced? When does it need to be repaired?
When does it need to be replaced? All of those are massive challenges. And then
when you think about the sizes of infrastructure, you're talking hundreds of
miles, thousands of miles. It really becomes this overwhelming challenge for
governments to manage. And that's ultimately where we help.
Roland Siebelink: Excellent. The whole world is already opening up for me as
I indeed imagine walking across the street and seeing all that street furniture,
as you call it, and then the curbs and the sidewalks. It must be a massive
number of assets or millions of dollars of assets that are just in the street
that people are charged to take care of.
Ben Schmidt: Absolutely. And there's two parts to it. There's the common
refraining of infrastructure is crumbling. If you go into your typical suburban
town, it might have somewhere around 100 miles of roadway. A hundred total miles
of road that they manage through mostly your local tax dollars. As I said
before, it costs about $1 million to repave one mile of road. Roads last, if
you're lucky, maybe 20 years.
What you're talking about is you have a community, 100 miles, 20 years that the
road lasts, you're talking about five miles, six miles of road that have to be
repaired every single year. That's $5 or $6 million. Your average small town,
suburban budget is not that high, probably closer to a few hundred thousand
dollars. This is how we end up in our predicament, which is as those roads
degrade, we just don't have the funding for it. And then even worse than that to
compound it, where we particularly help, is do you make the right choices? Are
you picking the right mile or two miles of road? And that's where these
decision-support technologies come in. What we like to think of ourselves as. We
don't make the decision. We help government officials, policymakers to make the
Roland Siebelink: You already mentioned that the origin story of RoadBotics
was at Carnegie Mellon. Can you talk a little bit more about that? How did it
actually get initially founded? Where did the idea come from? And how did the
people get together initially?
Ben Schmidt: I don't think we have enough time for all of that. One of my
co-founders, Christoph Mertz, he's been with CMU for a good while now. It was
really his original idea. His first insight here was how could you use a
smartphone - cheap, ubiquitous sensor; they've got great camera technology, they
have access to the internet, they've got all sorts of additional sensors on
them - how do you use that to actually understand something about roads and
The first thing that he tackled was around crack detection. Mounting a
smartphone up in the windshield of a car, like a dash cam, could you use some
advanced, sophisticated, machine learning processes to extract where are the
cracks on that road? Pavement usually starts to degrade and creates those
cracks. How do you identify those in the image and then ultimately generate a
map out of that?
That was his original idea at CMU. We ended up taking that - myself, Christoph,
and two others, so there were four original founders - took that idea, spun it
out of Carnegie Mellon, formed the RoadBotics entity, the company, and then
spent the next six months driving a real commercialized product out of that,
that not only looked at cracks but all the surface distresses that a pavement
engineer would care about. Potholes would be a big one. Cracks are important.
But then there's other things like graveling and other kinds of distresses that
happen on roads that mean something about the life of that road.
That's how we did it. It was a six-month sprint from spin out to commercial
product. We had our first clients right around that same time. They were
communities right outside the Pittsburgh region. For the last five years, that's
what we've been doing. We're up to probably over 300 governments today, work
mostly in the United States, but we also have some in the UK, we have some in
Australia, Canada. Been very successful in terms of roads are roads anywhere in
the world. There's this massive opportunity to help.
Roland Siebelink: Whether people drive on the left or the right, it doesn't
really matter. That's very good. But I did want to hone in a little bit on
selling to governments. We had some other podcast guests already who are selling
a form of gov tech - not competitive with you guys. But tell us a little bit
about that. Like you said, five years ago, you had your first client and then
you've grown it to 300 governments. How do you sell to governments? What is
different about it than any other experience you may have in selling to more
commercial entities? And what's some tips that other other founders in gov tech
might actually benefit from?
Ben Schmidt: The easiest way to put it is it is not easy. All the rumors are
true. Gov tech is hard. We've been very successful. I think we have a product
that definitely stands out and is unique, which helps a lot. That creates some
opportunities that you might not have in a more competitive environment. That's
one. But product I don't think wins necessarily. I think we all know that in the
I think for selling to gov tech, the hardest parts are really each government is
very unique. And I would say that in two respects. Both, they are unique and
they believe they are unique. Everyone wants their own unique concept around it.
Much of the procurement world - I guess I'll be a little harsh and say - is
generally broken. Its original purpose of creating this fair environment has
probably done quite the opposite. It's created a very unfair environment. Trying
to navigate those challenges - there's a lot to navigate. But it's definitely
doable. And I will say - that was all the negative; I started with the negative.
Positive for gov tech, governments are very loyal. Once you start working with
them, they continue to work with you, which is huge on many dimensions. It
offers referrals and renewal opportunities and things like that. That is
terrific. The other part is that for the most part, governments are not price
agnostic, but if they find something that they need, they will go and get it. If
you can find a solution that works, you can have great success with that.
When you first get into the room with the government, the first thing they'll
tell you is "No, we have it. We got this. Everything's fine. We already have
this on lockdown." It's the usual standard, get the people out the door because
we're about to be bombarded with terrible questions. It's just the knee jerk
response of government. Totally get it. It's like a scab. It's defense. The
fascinating part is that if you can get past that wall and you have these open
and honest conversations with governments, without fail - and I've sat across
from hundreds of governments at this point - without fail, every single one of
them are so excited to have any new tool in their arsenal that they can get at.
Opportunities to use technology, to be a little bit more efficient, to leverage
modern whatevers to help them, they are ecstatic about it.
For the exact same reason that gov tech is hard, there aren't many innovators in
the space. Once you do finally get past that barrier, governments are extremely
excited to try to adopt these things. They still have a lot of problems. They
still have challenges trying to get that across the line. But they are some of
the most excited and excitable clients that I've ever gotten to work with.
Roland Siebelink: Excellent. But as you say, there is always the fear that
having more data will also then cause a deluge of citizen requests to get things
fixed or alternative requests. How do you break through that initial fear? You
said people have developed almost scabs against that. How do you get beyond
Ben Schmidt: That's the interesting part. Those comments around governments
don't want to know. It's never the government that tells you that. It's the
external, the consultants, the VCs that come in and tell you, you don't want to
do this because of this. I've never heard a government say, "I don't want to
know this information."
Roland Siebelink: You're saying it's the government skeptics out there that
surmise in a way.
Ben Schmidt: Absolutely. All of us gov tech companies that have the real
scabs here that are burned and scarred. I would definitely say that it's not the
governments that are interested in hiding. They want to know. And I'll even say
that in two parts. You have the staff, the government officials that work for
the government, that's what they do for a living. And then you have the elected
officials. The elected officials want to know about what's happening because
they have a message around it. Here's what's happening. And here's what's
exciting. Here's how I'm changing it.
The staff want to know so that they can go tell the elected officials what's
happening, so they can get the funds to do what they're supposed to do. It's
very funny - all of us external people are saying, "No, they don't want to know
about their potholes." But it turns out when you get into the government side of
it, they're like, "I want to know everything because I can use it in these
different ways." I think it's very characteristic of gov tech in general, is
that it's filled with this mysticism about what you can do and what you can't do
and what people will believe. And then you finally get in a room with a town
manager or something like that, and it's a totally different experience.
Roland Siebelink: I was wondering, by definition, governments are
monopolistic and non-competitive, but since you sell into local governments, I
can imagine there's a little bit of competitiveness going on with the next-door
government or municipality or council, whatever you want to call it. Do people
actually look at what the neighbors are doing and try to keep up with them?
Ben Schmidt: That is a very good question. If I were to offer one tip, it's
that. It's that if you go to almost any geography, anywhere in the world, there
will be that one town that is 20 miles, 50 miles away, or it's next door, the
one town, they're the innovator. You talk about an early adopter. There's always
one government anywhere, they're the early adopter. If you can get them
convinced, I'll buy it.
It's fascinating. There's all sorts of local societies. There's local
professional chapters. And they're very local. They have some national
conferences. But public works - we work a lot with public works. Every single
head of public works will know every other neighboring towns' head of public
works, socially, professionally, they will know each other.
And it makes sense, right? If you get snowstorms or the fire departments need to
share equipment. They'll just share it. They know each other very well. If you
can find that early adopter, get them to be convinced - and by the way, every
government will tell you who that is. They'll say "Hmm, this is interesting. I
don't know if I would use this, but talk to them. They're the ones that usually
use all this new fangled stuff, and if they like it, I'll jump on board."
Roland Siebelink: Is there a pattern you've discovered? I'm just curious. Is
it better funded governments that are the early adopters or younger citizens?
What's the pattern if you've discovered any?
Ben Schmidt: I don't know that there's one overarching pattern. I think
there's probably a couple - you named a few of them. One would definitely be
communities that tend to have more money, tend to be more "okay." The
interesting part is that communities on the other end of the spectrum are also
just as innovative by virtue of they have to. "We don't have enough budget for
this. You're telling me we can be twice as efficient now? Sure. We'll give that
And then I do think there is quite a bit, as with any small government, the one
pro of it is that it's a few people that make all the decisions. If those few
people are of that interest and tilts - they want to try new things, they want
to see push the boundaries - those are your best early adopters. They're the
ones - the whole government will get behind you immediately.
Roland Siebelink: Ben, you mentioned also that you've started to sell
internationally as well - mostly some English speaking countries, as I
understand it. What drove you to that decision and how's that been going? What
are some lessons learned from internationalizing you're offering?
Ben Schmidt: Internationally, one of the things that we've done, just given
that we're a smaller entity, we really look for a partner. We've brushed over it
so far. We tend to work directly with governments, which is what we've been
talking about a lot. We also, historically, a lot of the infrastructure
management work is done by a civil engineering consulting firm. We'll work with
those civil engineering consulting firms to offer our services through them to
Internationally, we work only that way. We find a private entity in another
country that has a good understanding of the local market, the local
jurisdictional pieces, taxes, accounting, gov tech, all that kind of stuff.
We'll work with them and then they'll work on behalf of the governments with us.
That channel relationship just makes it a lot easier for a small entity like us
to do business anywhere in the world. It all boils down to finding really good,
high-quality, trustworthy international partners; once we get that going, then
it just takes off by itself.
Roland Siebelink: Very smart. Finding the right incentives to make them
happy and make a good business for them as well. And have them deal with all the
specifics of that country or province or whatever the case may be. Excellent.
You mentioned you have about 300 government clients that you're working for at
the moment or you've worked with so far. How big could this become, Ben?
Ben Schmidt: Let's hope we have every road on the planet at some point.
Roland Siebelink: That sounds like a big, hairy, audacious goal. Every road
on the planet.
Ben Schmidt: I totally agree. I think slightly more humbly, I would say I
don't think anyone that's seen our technology will go back. I don't think
they'll go back to pen and paper. Whether it will be us or someone else, I'll
try to say the jury is still out. But I think it's very much, this is the future
of how roads and infrastructure are going to get managed. And I would even go
one step further, which is, I think you go five years from now, 10 years from
now, you have autonomous vehicles, you have lane departure cameras, you have
delivery vehicles or garbage trucks, vehicles that have sensors already on them.
If we can distill that out - create or find the information that governments
want to know about it - that's going to be the chain going forward. And I think
that's how you create this - worldwide, you have a map of every piece of
infrastructure on the planet, it's being updated constantly, and that's being
handed on really a silver platter, a cloud-based silver platter to government to
make decisions about what to do next. That would be the holy grail of
infrastructure management. I will say I'd like it to be us. But understand that
there's a lot of stuff that has to be done between then and now.
Roland Siebelink: Very cool, Ben. I find it fascinating. On your way to
serving every single road on the planet, what are some of your next milestones
and what are you looking for - maybe people listening to this podcast to help
Ben Schmidt: Great questions. I would start the second one, which is if you
are a government official, please reach out. But in lieu of that, almost
everyone here lives somewhere where the government runs the roads for them.
Always happy to have referrals and things like that. Ultimately, that's how a
lot of activity gets done in local government - somebody just raising their hand
and saying, "Hey, can we check this out?" We're always up for that.
Growing commercial efforts. We need more folks talking to governments. We need
more folks creating relationships and things like that. Hiring across the board.
Roland Siebelink: Always likes to see that a startup is investing as much in
their go-to-markets capacity as their engineering capacity. That's what makes
Very good. If people want to hear more about RoadBotics, Ben, where should they
go? What's your URL and what should they download?
Ben Schmidt: Our website is the best: roadbotics.com. You can learn all
about us there.
Roland Siebelink: Excellent. Okay. Very cool. Thank you so much for your
time. Ben Schmidt, the CEO and co-founder of RoadBotics. It was a pleasure to
Ben Schmidt: Thanks for having me.
Roland Siebelink talks all things tech startup and bring you interviews with tech cofounders
across the world.