Artificial intelligence is a big deal for the likes of Amazon, Microsoft and Google — but what about a small business that can’t afford to have a data scientist on staff?
That’s the niche that Bellevue, Wash.-based Kurvv plans to fill, with a service that takes a company’s data and fits it into a pre-trained AI model that may not be perfect, but is good enough to address the problem that needs solving.
“We’re targeting companies that have data, but don’t have the knowledge or the resources to hire data scientists and can’t bring in a consultant,” Kurvv’s CEO, Ryan Lee, told GeekWire.
Lee left his post as a data science program manager at Microsoft in May to focus on getting Kurvv off the ground, drawing upon more than 15 years of experience in product management. One of Lee’s fellow co-founders is Vince Roche, who co-founded Boost Media, a San Francisco-based ad optimization venture, and now serves as Kurvv’s chief technology officer.
Lee said the investment will help his fledgling startup flesh out its business strategy and develop tools for its “data science as a service” offering, or DSaaS. He said the company is already working in alpha mode with customers in eight different verticals — including e-commerce, digital marketing and nonprofit groups. (Lee declined to name names, however.)
Kurvv’s business model capitalizes on the concept that not everyone needs a custom-trained deep-learning model, just as not everyone needs a custom-fitted business suit.
“The fact is that most business problems don’t require advanced models to solve,” Lee said.
If you’re dealing with medical applications or self-driving cars, you probably wouldn’t want a pre-trained model that fits the data to a mere 60% or 80% level of accuracy. But that level of precision could be just fine for helping a boutique hotel figure out which high-value customers to target for an email sales promotion. The hotel scenario was one of Kurvv’s first case studies. Kurvv said the AI-assisted campaign got results within 24 hours and ended up driving $35,000 in incremental revenue.
That kind of payoff may not sound like much to the big players in the AI game, but it’s enough to get small businesses interested in artificial intelligence — assuming that the price is right.
“Based on instances where we were able to compare, we’ve found that not only were our costs 10 times less than what other well-established AI companies would charge, but also 10 times faster in terms of start to results,” Lee said.
Right now, Kurvv is powered by Lee, Roche and a third co-founder who’s staying behind the scenes for now, plus about seven people who are working on a contract basis. Lee said he’d like to hire at least one additional “high-caliber” data scientist to help the venture get to the next level.
Kurvv isn’t the only startup aiming to offer low-cost AI solutions. Seattle-based Xnor, for example, is fielding a self-service software platform called AI2GO that lets users select and download pre-packaged AI applications and deep-learning models. Another Seattle startup, Amperity, uses machine learning to leverage customer data, attracting a high level of industry buzz in the process. And looking beyond the Seattle area, there are dozens of companies jumping onto the DSaaS bandwagon.
Will Kurvv get ahead of the curve? That’s what Lee is counting on, and what the $1 million in seed funding is going toward.