Bunkerhill Health raises $55 million to expand its agent AI platform Carebricks.
Sequoia Capital, Felicis, Optum Ventures, and Y Combinator continue to participate in the closing of the company’s Series B round announced today. But total funding alone doesn’t answer the key question hospital executives want to know about healthcare AI: Will the software run inside a live hospital?
That question is one of the reasons Khosla Ventures was named to the deal. Healthcare organizations have poured money into machine learning pilots that work well in research environments but never touch actual patient records.
Bunker Hill’s argument to investors, and the health systems that are already paying for it, is that Carebricks bridges the space between models that work in a sandbox and those that run against live, facility-wide clinical data.
The background to this is the issue of spending and staffing. U.S. health care spending will reach $5.3 trillion in 2024, according to the Centers for Medicare and Medicaid Services, and workforce shortages continue to strain health care providers across the country.
Bunker Hill frames the opportunity around the gap between what health systems want to do for patients and the time their staff can do it. Decades of investment have been made in documentation systems aimed at reducing the burden on clinicians. Bunker Hill’s bet is that the next technology spending will be on software that works on ideas clinicians already have, rather than just recording.
Nishit Khandwala, co-founder and CEO of Bunker Hill Health, explained that “health care is advancing faster than the operational capacity of the health care system.” “Every major health system has many opportunities to improve patient outcomes beyond what employees can address, and we believe AI agents can help make more of these ideas a reality.”
Carebricks allows hospitals to build their own agents rather than purchasing off-the-shelf fixed products. Some agents examine cardiac images for early signs of heart disease and flag patients who need follow-up care. Some handle pre-approvals and keep registry data up to date. This extends to administrative tasks that are rarely highlighted in AI pitches but consume staff time each week. The platform is currently operated by Cleveland Clinic, University of Texas Medical Branch, and Intermountain Health.
Vinod Khosla, founder of Khosla Ventures, said, “The bottleneck for healthcare AI has never been the technology, it’s been getting the health systems to actually run it. Bunker Hill has bridged that gap, making it much easier to deploy AI and already gaining traction within critical health systems that would take most companies years to gain.”
20 AI agents running within one hospital system
UTMB provides the clearest indication of what “running” is like once the pilot label is removed. According to Dr. Peter McCaffrey, UTMB’s chief AI officer, the system currently has more than 20 agents resident on Carebricks, spanning clinical care, operations, and management.
During the first month at UTMB, a coronary artery calcium detector built on an FDA-approved algorithm flagged patients at immediate risk of heart attack. Cardiology confirmed the risk and performed triple bypass surgery.
The care team at UTMB believes early detection saved the patient’s life. This was a single case, not a controlled trial, and Bunker Hill has not published data on how often the drug causes false positives or how it performs over time in a broader patient population.
UTMB’s other numbers come from Bunker Hill and the health system itself, rather than independent audits. Nephrology triage staff will now prioritize patients according to severity, escalating urgent cases and directing others to telemedicine. UTMB reports that this has reduced the average waiting time for specialists by more than 50%.
Pulmonary nodule personnel follow up from incidental findings on CT scans to appropriate follow-up. UTMB claims 80% faster response to emergency cases, less manual work for coordinators, and twice as much follow-up according to guidelines.
These are operational results reported by health systems in real production environments and are not meaningful synthetic benchmark scores. It also means that the numbers reflect one facility’s data situation and staffing levels and do not guarantee that another hospital will see the same curve.
“We are already making a huge impact on patient care, but we are only at the beginning when health systems can operate with agent AI at this scale,” McCaffrey said.
Of course, this does not eliminate the work that health systems will still need to do themselves. Bunker Hill says it will use the new funding to expand Carebricks into a broader range of clinical and operational use cases, while building out governance, oversight and safety measures.
A platform that allows nephrology departments to build their own triage agents also means that the department owns the results of how that agent is calibrated. Health system boards considering a Carebricks-style implementation need answers about assignment of responsibilities, frequency of monitoring, and what happens when agent judgment and clinician judgment do not match before approving scale.
UTMB’s 20-agent footprint now makes Bunker Hill a reference case that few competitors can match. Whether this number acts as a signal to the rest of the industry will depend on how UTMB and the other systems currently running Carebricks handle the governance aspects as the number of agents increases.
Want to learn more about AI and big data from industry leaders? Check out the AI & Big Data Expos in Amsterdam, California, and London. This comprehensive event is part of TechEx and co-located with other major technology events such as Cyber Security & Cloud Expo. Click here for more information.
AI News is brought to you by TechForge Media. Learn about other upcoming enterprise technology events and webinars.
