Bunkerhill Health has raised $55 million to scale its agentic AI platform, Carebricks.
The closing of the corporate’s Collection B spherical, introduced immediately, folds in continued participation from Sequoia Capital, Felicis, Optum Ventures, and Y Combinator. Nonetheless, a funding complete doesn’t reply the important thing query any hospital government needs to learn about healthcare AI: does the software program run inside a working hospital?
That query is among the many the reason why Khosla Ventures put its title on this deal. Healthcare organisations have put no scarcity of funding behind machine studying pilots that carry out nicely in a analysis setting after which by no means contact a reside affected person chart.
Bunkerhill’s argument to buyers, and to the well being methods already paying for it, is that Carebricks closes the area between a mannequin that works in a sandbox and one which runs towards reside medical information at institutional scale.
The backdrop is a spending quantity and a staffing drawback. US healthcare spending hit $5.3 trillion in 2024, based on the Centers for Medicare & Medicaid Services, and labour shortages proceed to pressure suppliers nationwide.
Bunkerhill frames the chance round a spot between what well being methods need to do for sufferers and what their workers have time to execute. A long time of funding went into documentation methods meant to ease the burden on clinicians. Bunkerhill’s guess is that the subsequent spherical of know-how spending goes towards software program that acts on concepts clinicians have already got, reasonably than simply recording them.
“Drugs has superior quicker than our healthcare system’s means to operationalise it,” defined Nishith Khandwala, Co-Founder and CEO of Bunkerhill Well being. “Each main well being system has extra alternatives to enhance affected person outcomes than its workforce has capability to deal with. We imagine AI brokers will help them flip extra of these concepts into actuality.”
Carebricks lets hospitals construct their very own brokers reasonably than shopping for a hard and fast product off the shelf. Some brokers overview cardiology imaging for early indicators of coronary heart illness and flag sufferers who want follow-up care. Others deal with prior authorisations or preserve registry information present, and the vary extends into administrative work that hardly ever will get consideration in AI pitches however consumes workers hours each week. Cleveland Clinic, the College of Texas Medical Department, and Intermountain Well being all run the platform immediately.
Vinod Khosla, Founding father of Khosla Ventures, stated: “The bottleneck in healthcare AI was by no means the know-how, it was getting a well being system to truly run it. Bunkerhill closed that hole. They made it a lot, a lot simpler to undertake AI and have already got traction inside essential well being methods that may take most corporations years to earn.”
Twenty AI brokers operating inside one hospital system
UTMB presents the clearest image of what “operating it” appears like as soon as the pilot label comes off. The system now has greater than 20 brokers reside on Carebricks, spanning medical care, operations, and administration, based on Dr. Peter McCaffrey, UTMB’s Chief AI Officer.
In its first month operating at UTMB, a coronary calcium detection agent constructed on an FDA-cleared algorithm flagged a affected person as being at imminent threat of a coronary heart assault. Cardiology confirmed the danger and carried out a triple bypass.
UTMB’s care group credit the early detection with saving the affected person’s life. It’s a single case, not a managed trial, and Bunkerhill has printed no information on how typically the agent produces false positives or the way it performs throughout a broader affected person inhabitants over time.
Different figures at UTMB come from Bunkerhill and the well being system itself reasonably than unbiased audit. A nephrology triage agent now prioritises sufferers by severity, escalating pressing instances and routing others to telemedicine, and UTMB stories this has minimize common specialist wait instances by greater than 50 %.
A lung nodule agent tracks incidental findings on CT scans by way of to the proper follow-up, with UTMB citing an 80 % quicker response on pressing instances and a doubling of guideline-concordant follow-up alongside a drop in handbook coordinator work.
These are well being system-reported operational outcomes from reside manufacturing use, not artificial benchmark scores, which issues. It additionally means the numbers replicate one establishment’s information circumstances and staffing setup, not a assure that one other hospital will see the identical curve.
“We’ve already seen large influence on affected person care, and we’re solely firstly of what turns into doable when a well being system can function with agentic AI at this scale,” McCaffrey stated.
In fact, none of this removes the work well being methods nonetheless should do themselves. Bunkerhill says it is going to use the brand new funding to broaden Carebricks right into a wider vary of medical and operational use instances whereas constructing out governance, monitoring, and safeguards.
A platform that lets a nephrology division construct its personal triage agent additionally signifies that division owns the results of how that agent is tuned. Well being system boards weighing a Carebricks-style deployment want solutions on legal responsibility task, monitoring cadence, and what occurs when an agent’s judgment and a clinician’s judgment disagree, earlier than signing off on scale.
UTMB’s 20-agent footprint offers Bunkerhill a reference case few opponents can at present match. Whether or not that quantity holds up as a sign for the remainder of the business is determined by how UTMB, and the opposite methods now operating Carebricks, deal with the governance facet because the agent rely climbs.
Wish to study extra about AI and large information from business leaders? Take a look at AI & Big Data Expo happening in Amsterdam, California, and London. The great occasion is a part of TechEx and is co-located with different main know-how occasions together with the Cyber Security & Cloud Expo. Click on here for extra info.
AI Information is powered by TechForge Media. Discover different upcoming enterprise know-how occasions and webinars here.
