It’s never easy to raise capital, RoadBotics CEO Mark DeSantis says, but it’s been fairly straightforward for his Carnegie Mellon spinout.

“We are fortunate in that we have a great team,” DeSantis says. “They’re solving a real problem and we’re getting customers as proof of that. If you have those things, raising money isn’t nearly the challenge that people think it is.”

The Pittsburgh-based company — which grew out of CMU research into standardizing road assessment using artificial intelligence — closed a $7.5 million Series A round this month. That brought the company to around $11.4 million raised since launching in December 2016.

RoadBotics has more than 150 customers in 22 states and 11 countries. Even with revenue coming in, the company needs capital to solve a problem at the greatest possible scale, DeSantis says

“That’s the end game: How can I solve this problem on a large scale?” he says. “And that typically requires money. You can bootstrap, right? People have done that, but that’s hard to do.”

DeSantis, who is on his fourth machine learning startup, shared his experience building RoadBotics from the ground up, including rapid two-year growth since it got its first customers.

Find the nails first

DeSantis helped co-found RoadBotics after his previous company kWantera, and its subsidiary kWantix, was acquired by Moody Aldrich Partners. CMU’s Christoph Mertz had the idea while working on autonomous vehicle sensors.

He wondered if same image processing and AI technologies could be used to evaluate the conditions of the roads on which they travel? And, could a simple smartphone do it? Several years later, Mertz developed deep learning algorithms that could evaluate road images captured by a smartphone.

“We’ve grown very fast and it’s mainly because the problem that we’re trying to solve, which is to give people who maintain roads actionable information about the status of those roads,” DeSantis says. “It’s a hard thing to do and it’s also very expensive.”

In the U.S. alone, there are more than 4 million miles of paved road that cost more than $100 billion to maintain each year.

RoadBotics’ app can be downloaded, and a cellphone camera pointed through the windshield to collect data on cracks and other pavement and asphalt features. As soon as the phone finds a friendly Wi-Fi connection, it sends the data to the cloud. RoadBotics’ AI platform analyzes the road surface, meter by meter, just like a pavement engineer would, and generates a rating that goes to customers in the form of a multicolored map.

The system is objective and gets smarter every time it’s used, which is useful for borderline calls.

“They will probably agree on the extreme values,” he says. “This is a perfectly brand new road. This is a terrible road. The problem is our customers need to see the nuance between those two states, and that’s where the disagreement is.”

The idea required front-end validation, even before the RoadBotics team went out selling the value of its tool. DeSantis recommends all startup entrepreneurs avoid being a hammer looking for nails. Once you find the nails you can craft your hammer.

“Don’t worry about your technology or your product,” he says. “You need to understand the problem. You need to have a Ph.D. in the problem first.”

Because RoadBotics’ technology provides value to the people who own or manage a road network, its list of customers is growing rapidly. So, is its investor list, which includes BMW, Hyperplane Venture Capital, the Wharton Alumni Angels, Radical Ventures, and Pittsburgh’s Innovation Works.

Built to last

Managing RoadBotics’ future growth isn’t a small challenge for DeSantis and the other 38 employees, but getting the right early adopters was critical.

DeSantis says you want early stakeholders who challenge you to develop a product that makes a difference for them. Those early customers also need to be different enough that you’ll think broadly about the product.

“Getting those right early customers can make all the difference in the world, that can really be the difference between winning and losing, period,” he says.

But your job isn’t done at that point, you can’t assume that once you have those early adopters and they like the product, you have a business.

“You then have to ask yourself: Is this a big enough market and are these people truly being objective or are they just cheerleaders?” DeSantis says.

Founders sometimes get ahead of themselves, thinking they’ve arrived.

“The answer is, no you haven’t,” he says. “You’ve got a long way to go. You’re not there. You think you are; you’re not. So, you start drinking your own Kool-Aid and that’s when the problems start.”

In addition to making sure you always challenge yourself and your team, DeSantis says nothing happens in the time frame you expect, so you have to plan for that. Some things are going to take longer than you expect.

You also are not doing it alone; bring talented people along with you. “You’re fooling yourself if you think that you’re going to do it all yourself,” he says.

Lastly, once you have people around you who can handle working in a startup — because not everyone’s cut out for it — you need to give people room and an opportunity to do their job.

“Your job is to architect and build the company, not over-the-shoulder second-guess people,” DeSantis says. “If you get the right talent that’s motivated and incentivized, then give them the freedom to do what they need to do, up to and including making mistakes.”