When the National Marrow Donor Program needs to match a recipient up with a donor, the first step is to take a sample of the recipient's DNA and run it through a database containing more than 42 million pieces of data. For years, that process would take up to 48 hours. Now it takes less than a minute.

That's because the organization recently started working with artificial intelligence company phData. The Minneapolis-based startup builds machine learning and analytics platforms for companies and organizations that aren't equipped to do so themselves. It had $18 million in 2019 revenue, and has more than 60 clients ranging from health insurance companies to law enforcement agencies to medical device makers. 

"These companies are looking to build platforms that require hundreds or thousands of machine learning models," CEO Ryan Bosshart says. "That's something that Facebook and Google and those kinds of folks are used to, but traditional companies aren't. So we do it for them."

Bosshart and phData's co-founders, Adam Fokken and brothers Brock and Mac Noland, met while building software for the financial and legal industries at Thomson Reuters. As they observed the cost of creating A.I. platforms declining, they realized there might be a business opportunity in building them for companies that might not have the expertise to do so on their own. Fokken and the Nolands, all engineers, left their jobs to found phData in 2014. Bosshart followed soon after and was installed as the company's CEO. 

PhData made $9.8 million in revenue in 2018, up from $170,000 in 2016. That earned it the top spot on this year's Inc. 5000 Series: Midwest--a list of the fastest-growing private companies in the region. The company prices its projects based on size, duration, and complexity, with the most simple machine learning systems starting at $50,000. 

Some projects are relatively basic, like the air-filter company that wanted to create a better user experience on its website. Others are far more involved, like a machine-vision platform that performs quality control for a major fast-food chain. (Bosshart declined to name the client.) Cameras placed in the kitchen examine food as it cooks to ensure it meets the company's standards, while additional cameras over the counter identify the items on a food tray and check them against the customer's order to ensure all is correct. 

"I know it's small," Bosshart says, "but being able to solve some of these problems is pretty cool. It's going to be a crazy and exciting new world out there."

 inline image

PhData is far from the only company building A.I. as a third-party solution--and for good reason. "A.I. is becoming a differentiator and a way of future-proofing, but most companies don't know where to begin," says Ahmer Inam, chief A.I. officer at Redmond, Washington-based Pactera Edge, which has built systems for more than 100 Fortune Global 500 companies. "The desires of upper management don't match up with the company's actual capabilities." Other competitors include Fractal Analytics, a multinational company that last year raised $200 million. And while Bosshart says some current and prospective clients have decided to pause projects during the coronavirus crisis, he's optimistic that demand will return and there won't be a lasting impact on the business. "Much of the work we're doing with our customers is highly strategic and long-term," he says. "These analytic and machine learning platforms are part of their 10- or 20-year plans."

Bosshart says phData separates itself by paying extra attention to customer service. After a project is live, the company helps the client operate it and keeps staff on call 24 hours a day to deal with any potential issues. PhData doesn't do a lot of marketing, instead relying mostly on word of mouth, and only takes on one or two new customers each month. The extra care for each existing client, Bosshart says, has been a key to success--as has staying humble.

"This area of machine learning and analytics is still pretty new, so there's no template for it," he says. "And so coming in with a lot of humility and listening to our employees, listening to our customers, and being willing to change and to take advice, have all been extraordinarily helpful for us."