Khosla-Backed Health AI Startup Raises $27.5 Million To Supercharge Primary Care

^Telemedicine makes doctors more accessible, but it doesn’t necessarily make them more productive. Seeing patients remotely still involves a lot of time and data entry into electronic health records. Neal Khosla wants to change that by using artificial intelligence to turbocharge primary care and make telemedicine better for both doctors and patients. “We think that AI can be a superpower, where we can make the cost of care lower,” says Khosla, cofounder of San Francisco-based Curai, which announced a $27.5 million Series B financing round on Wednesday. 

The company has built a chat-based platform that essentially uses AI as a virtual member of the patient care team alongside two humans—a clinical assistant and physician—to help them make better, faster decisions.  “At the end of the day, technology as a standalone is not valuable,” says Khosla. “It has got to be integrated into human processes and it’s also got to change the way that people work.” 

If his last name sounds familiar, that’s because Neal is the son of billionaire investor Vinod Khosla, whose firm Khosla Ventures is an investor in Curai. The Series B round was led by Morningside Group, joined by Khosla Ventures and General Catalyst. The company has raised $57 million to date. 

While the Covid-19 pandemic spurred the U.S. healthcare system to quickly adopt telehealth, which was seen by those in the industry as a historic victory, it also underscored just how sluggish and change averse the system can be. But Curai is anything but change-averse. It doesn’t want to knock a few bucks off the price of a telemedicine visit. The company is looking to radically reduce existing costs by a factor of 10, meaning primary care could be accessible to the uninsured, gig workers, minimum-wage workers and people with high-deductible plans. The goal is for the chat-based visit to “be cheaper than the cost of taking a bus to your physician,” says Vinod Khosla.   

At the same time, Curai is ironically something of a throwback to the era of having a the local family physician who knew the intimate details of every patient’s history. Its technology can be applied to scan through medical records and organize key points in seconds. “The AI can do all those things that a physician doesn’t have time for and then the physician can make the final decision,” says Vinod, who famously coined the phrase “Dr. Algorithm” a decade ago. “That’s the way you will really reduce the cost and dramatically improve care quality and continuity.” 

“That’s what we’re really building towards … this alternate safety net for health care access.”

^What sets Curai apart is the underlying technology, which was developed by Neal and cofounder and CTO Xavier Amatriain, who is best known for building the algorithms behind the personal recommendations you get if you use online streaming service Netflix. After graduating with a master’s in computer science from Stanford, Neal was looking for a way to make a bigger impact. A family health scare a few years earlier had made him realize the precarious nature of insurance and access to healthcare. Amatriain, who left Netflix to work at question and answer site Quora, was similarly looking for a new challenge. “We believed AI could be the great equalizer and help amplify [physicians],” says Neal. “But there was a lot of discovery to be done about how we could do it.”

In 2017, the duo embarked on what would become a two-year journey that would generate not only machine learning algorithms, but also an electronic health record built from scratch. The Curai platform ingests information as the patient answers a series of questions, and then helps organize and recommend next steps in conjunction with a human clinical assistant and doctor. The AI is dynamic in that it responds to what the patient is telling it, rather than following a rules-based system.

Patients are able to do real-time messaging for urgent issues. For less urgent matters they are sent a response later. The most important component of the product is the “learning loop,” says Amatriain. “All the data that is generated by our system, every day, goes back into training those models.” He gives the example of Covid-19. A year ago, the virus didn’t exist, but now each Covid-19 patient that uses Curai helps retrain the model and improve its accuracy. “The AI doesn’t make the final decision,” Amatriain adds, but notes “it’s a very important part of the team.” In 2019, Curai named Dr. Sylvan Waller, a former executive at telemedicine company MDLIVE, to the role of chief medical officer. 

While the Curai team has been focused on perfecting its technology, Gerald Chan, the cofounder of Morningside Group, which led the investment round, is excited to get it deployed on a larger scale. “The system gets better with usage,” says Chan. “We very much would like to see validation by having more users and more learning.” More than 350,000 patient visits on the Curai platform to date have led the AI to achieve a nearly 90 percent diagnosis accuracy rate, according to the company.  

Right now Curai is only offered in California, but it’s planning to roll out in half of the U.S. by mid-year and in all 50 states by 2021. Neal says he expects the total staff to double to around 60 employees over the same period. Curai previously sold on a direct-to-consumer subscription model, with a free 6-month trial and $7.99 monthly after that. With this funding round, the company will begin offering its platform to employers, insurers and public sector groups. 

The company’s ultimate vision is to expand beyond primary care and help customers navigate other aspects of the healthcare system, like drug pricing and lab tests. “That’s what we’re really building towards over the next one or two years, this alternate safety net for health care access,” says Neal. “It’s centered around a relationship with a physician who can help you access care and navigate those choices.”