Zoi Capital is re-imagining how to invest in healthcare AI with its debut fund, which it began raising in April. The New York-based firm is targeting $200 million for Health AI Fund I.

Dr Ronald Razmi, co-founder and managing director of Zoi, said the idea for the firm came about after he was approached by the former head of state of a G7 nation who was looking for help in evaluating how AI can increase access to high-quality healthcare. The conversation clarified the need for experts with extensive health AI expertise to identify the best use cases for funding.

“A lot of investment has gone into use cases that so far haven’t really shown any benefit,” Razmi told Venture Capital Journal. “Because of the risk to the patients, because of the convoluted nature of the industry where the buyer is not the user… you have to navigate a bigger maze in order to bring innovations into healthcare.”

All of that makes healthcare tech a sector that requires both clinical and operating experience to choose the right companies targeting the best use cases to invest in, he said.

Even companies that have developed an excellent algorithm that solves a real healthcare problem have faced zero adoption rates. With limited annual technology budgets, drug manufacturers and medical centers “have to focus on certain use cases that are going to move the needle in terms of their business models, that improve revenues, that lower costs, that make it easier to attract and retain key staff like doctors and nurses,” Razmi noted. That means a technology that still requires a degree of human supervision may not be at the top of their list.

A former cardiologist who practiced medicine at the Mayo Clinic, Razmi entered the business world in 2007 as a McKinsey consultant, where he helped funds with healthcare investments. In 2011 he founded and became CEO of Acupera, whose software helps medical centers proactively manage their patient populations. Acupera merged with another company in 2017. Having added AI to Acupera’s platform, Razmi said he’s familiar with the various barriers faced by health AI entrepreneurs.

“I faced every one of them, including data access, the complexity of the sales cycle, adoption issues with the users,” he said. “So that kind of frontline experience, both as a clinician but also as an operator, is invaluable when you’re assessing these investment opportunities.”

Razmi’s co-founder, Brian Beller, was previously a corporate officer at several healthcare companies, most recently at Horizon Therapeutics, the Ireland-based biotech company that Amgen acquired last year for $27.8 billion. Beller was also on the management team of other companies that have had successful exits, such as Schwarz Pharma, which was acquired by UCB for $5.6 billion in 2006, and medical device company Fenwal Holdings, which Fresenius Kabi acquired for $1.1 billion in 2012.

Razmi and Beller shared their fundraising goals in an announcement, but they have not yet filed a formal fundraising document with the SEC.

Zoi Capital plans to invest in 10 to 15 companies out of Fund I, leading Series A or B rounds with checks between $6 million and $12 million. A portion of the fund will be reserved for follow-on investments in portfolio companies that reach critical milestones. Anticipating a lot of time spent on optimizing product plans and commercialization strategies, “we want to come in when the technology risk is lower, when the product market-fit has been established,” Razmi said.

Razmi has been discussing term sheets with a few US-based founders and has also looked at some European and Israeli companies.

In some countries that Razmi has visited recently while promoting his book, AI Doctor: The Rise of Artificial Intelligence in Healthcare, which was published in January, “there’s a lot of discussion about how AI can improve their healthcare systems and some of the same issues we’re having here,” including staff shortages and inefficiencies, he said.

Zoi plans to help the companies it invests in market their technologies in other parts of the world where they can improve healthcare delivery. That could be of interest to potential investors in its fund, said Razmi.

“When I was a McKinsey consultant, we would go into a country, study their healthcare system and figure out key unmet needs and what technologies could address those needs,” he said. “We have that expertise bringing those technologies into a new region and standing them up. So that’s part of our value proposition for investors.”

Over the three years Razmi spent writing his book, he and Beller developed an investment platform that includes a knowledge management system that breaks down healthcare into 18 subsectors and identifies key technologies for each. They also built a database of several thousand companies that lists their distinguishing characteristics, which problems they solve and what sectors they focus on.

“We also developed a set of score cards both for use cases and the companies to make sure we avoid a lot of the pitfalls that have plagued investors to date,” he said. Although most of the investment framework is proprietary, Razmi has shared parts of it in his book for founders and investors to consider before spending years building companies based on certain use cases.

Saving time and work

One technology category Zoi is excited about is called ambient intelligence, which uses microphones and cameras to record conversations between doctors and patients and can generate notes for future treatment, order prescriptions and make referrals. That’s a big step forward from the electronic health record, which is one of the most significant breakthroughs in digital health products but ended up creating extra work for nurses and doctors and adversely colored their views of health technology, said Razmi.

“A big focus for Zoi is technology that actually relieves work for users,” he said. “When I was a practicing doctor, I hated writing notes. It was slowing me down. It added a lot of hours to my day. So documentation seems to be an initial use case where there’s significant unmet needs. And with generative AI, the capabilities of these technologies are getting to a point where they may be good enough to create a note that the doctor doesn’t need to edit too much.”

Razmi said the chief medical informatics officer of a large medical center told him that he thinks 80 to 90 percent of medical centers will have this type of technology in place within the next two to three years.

“It solves a real problem. It’s low-risk because it doesn’t involve clinical care or clinical decision-making, and it will finally show some benefit to people in the medical community in terms of what these technologies can do rather than traumatizing them [with tools that add] more work to their already-busy schedules,” he said.

Another kind of ambient intelligence embeds sensors in walls or in a bed and uses signals to analyze a patient’s movements, whether their gait is steady or if they have fallen. It can also monitor respiratory and heart rates.

“That means it can be used to monitor people in long-term care facilities or older people in their homes. And in a hospital room, where it’s keeping an eye on a patient without the nurse constantly going in and checking on them. It can collect some of that information that a nurse usually collects,” Razmi said.

This technology has reached a level of maturity that Zoi is looking at several companies in that space and it has already offered a term sheet to a company using ambient intelligence, he added.

“This is the kind of thing that’s foundational. It moves the needle and saves people a lot of work,” he said. “The promise of AI should be doing things that we’re doing now but we don’t want to do them. So we can focus on more important stuff.”