Agentic AI at Work Inside Health Systems: Lessons from UTMB and Cleveland Clinic

How two leading health systems turned more than a hundred clinical ideas into AI agents that finish the work, and what they learned about scaling, governance, and trust.

In a recent webinar with the Scottsdale Institute, leaders from two of Bunkerhill’s deepest deployments sat down to talk about what happens when a health system stops buying a point solution for every problem and instead runs many AI agents on our platform, Carebricks. Carebricks is live at more than a dozen health systems. Two of them, the University of Texas Medical Branch (UTMB) and Cleveland Clinic, shared how they got from a handful of pilots to agents working alongside their staff every day.

The panel featured:

• Dr. Peter McCaffrey, Chief Digital and AI Officer at UTMB, a pathologist by training who oversees the system's AI intake and deployment.
• Dr. Salim Hayek, Chair of the Department of Internal Medicine at UTMB, a cardiologist and physician-scientist.
• Dr. Po-Hao “Howard” Chen, Vice Chair of Artificial Intelligence for the Diagnostics Institute at Cleveland Clinic, a musculoskeletal radiologist.

A live poll set the stage. Asked where their organization stands with agentic AI today, roughly 23% of attendees said they were still exploring, 50% were piloting one or two use cases, and 27% were already scaling across the enterprise. It is a snapshot of an industry that has moved from curiosity to real adoption in under a year.

How does a health system scale from one AI pilot to twenty agents?

The short answer: treat agents as an extension of your workforce, and build a repeatable path from idea to production instead of a new project every time. At UTMB, that path turned more than 100 submitted ideas into over 20 live agents, with Dr. McCaffrey expecting that number to reach 40 or more within a year.

His central point is that most work in a health system does not fail because it is hard. It fails because there is simply too much of it for people to get to. When a patient “falls through the cracks,” that is usually a polite way of saying a job needed doing and no one had the hours to do it.

Rather than reinvent a technology stack for each use case, UTMB reuses common building blocks: the way an agent reasons, how it looks and feels, how it reads an HL7 feed or writes back into Epic. Every integration becomes a permanent capability of the platform, so a new agent is an iteration, not a fresh procurement cycle. That is what lets aortic-aneurysm detection, infectious-disease triage, and case-mix-index optimization live on the same foundation.

One example made the stakes concrete. UTMB's first deployments included screening imaging for coronary artery calcification, a finding often noticed on scans ordered for something else. In a case later published in Texas Monthly, an agent surfaced a patient with a heavy calcification burden and little prior cardiology contact. The follow-up that resulted led to a CT angiogram and, ultimately, a three-vessel bypass. Not every case ends that way, but without an agent paying attention, no one was paying attention to that patient at all.

Can AI fix patient access without hiring more physicians?

Yes: by routing the right patient to the right physician through the right modality, instead of scheduling first-come, first-served. Dr. Hayek framed patient access as the most common problem shared across health systems, and one that hiring alone cannot solve.

He had seen the old playbook up close. At the health system where he previously worked, wait times to see a cardiologist stretched to six months. The fixes were familiar ones: Saturday clinics, after-hours slots, telehealth, and physician incentives. They worked, but they were expensive, and they still funneled low-complexity cases into specialty clinics. By some estimates, roughly 30% of specialty care is low-value and could be handled elsewhere.

The deeper flaw was triage, or the lack of it. When scheduling runs first-come, first-served, urgency is ignored entirely.

UTMB's answer is a clinician-driven agentic workflow that reads each incoming referral and assesses three things: complexity, urgency, and the best modality to see the patient, whether in person, by telehealth, or by e-consult. Crucially, this was not bought off the shelf. Every specialty built its own playbook, each division assigned an AI champion, and those champions review a random sample of charts every month to validate the agent and refine the rules.

The outcomes, across several subspecialties in internal medicine:

• Nephrology: high-complexity patients were seen roughly twice as early as before, while 10 to 15 e-consults a week resolved lower-complexity questions in 48 to 72 hours instead of a months-long wait.
• Endocrinology: a service that looked like it needed three to four more endocrinologists began meeting its full demand with the existing team, simply by shuttling the right patients to the right modality.
• Colonoscopy: evidence-based risk scoring prioritized patients at higher risk of colorectal cancer, surfacing a meaningful number of adenomas and interventions earlier than the queue would have.

The principle underneath all three: when you pull a simple question out of the queue, you move a high-urgency, high-complexity patient to the front.

How do you clear a 60,000-patient actionable findings backlog?

Cleveland Clinic did it by putting AI agents on the slow, repetitive first pass and keeping clinicians on the judgment and the patient conversations. Dr. Chen described the problem behind every radiology scan: the diagnosis the ordering doctor was looking for is rarely the only finding on the image.

He calls each of these an actionable finding: a lung nodule spotted incidentally when a patient comes through the ER after a car crash, for instance. The trauma gets treated; the nodule, which could one day become cancer, is easy to lose track of once the patient goes home. To catch them, the Clinic had built a team of advanced practice providers who manually tracked every flagged case in the EHR. It was a genuinely good idea that could not keep pace: the backlog grew into the tens of thousands, eventually around 60,000 cases.

Carebricks agents now work through a set of rule-based questions on each actionable result. Is a follow-up actually indicated? What is the right window? Did the recommended study happen, and if not, why not? Can an MRI reasonably supersede a recommended CT? The agent resolves what it can and escalates only the cases that genuinely need a human.

Backlog reduced from ~60,000 cases to near zero · 44% more patient-facing clinics held

The benefit the team valued most was not throughput. It was getting their people back to the work they trained for. Instead of mining the record to fill out the equivalent of a spreadsheet, the providers now spend far more time as a safety net for patients, explaining why a finding matters and what to do next. Clinic volume for those conversations rose 44%, and so did job satisfaction. As Dr. Chen put it, automation “opens the possibility for folks to engage with the system, because they know there's more certainty that it will be followed up.”

What did the panel agree on about deploying agentic AI?

Beyond their individual programs, the three leaders converged on a set of hard-won principles for making agents stick inside a health system.

Let demand pull, don't push

UTMB did not recruit champions so much as watch them volunteer. Once clinicians saw an agent solve a colleague's problem, they arrived with their own. About 90% of the people driving use cases nominated themselves.

Governance is one front door, not a maze

Both systems run a single enterprise-level AI governance process. The platform is vetted once; each new use case gets its own approval, with an expedited path when it closely resembles one already live. There are no shadow deployments. Every use case is inventoried, tied to a champion, and held to a time-bound KPI it must meet to stay in production.

Clinicians own judgment; the platform owns execution

Clinical criteria always come from the physicians and staff who live the work; the forward-deployed team bakes those criteria into the workflow and the way it acts, whether through HL7, FHIR, SMS, or a write-back into the record. Dr. Hayek was blunt that a boxed product a technical team designs alone will not survive contact with patient care, and that a leader drafting the first playbook is what gets everyone moving.

Match the rigor of validation to the risk

Cleveland Clinic validates its agents with a blinded, retrospective comparison: does the final action a provider takes come out the same whether a human or an agent mined the chart first? They plan to repeat it roughly every six months to catch any drift. For lower-risk decisions like referral triage, Dr. Hayek noted, two physicians can reasonably disagree, so the bar is not perfection. It is beating how the system schedules today. “And we're beating it by far.”

Treat the vendor as an extension of the team

Both leaders described Bunkerhill's forward-deployed model as an extension of their own staff: the same people, embedded for the long haul, who understand the culture and sit in on the recurring work of tuning each agent. As Dr. McCaffrey put it, one central contract is fine, but without many touchpoints into the organization, momentum rarely lasts.

Key takeaways

1. One platform beats a dozen point solutions. Running many agents on shared infrastructure turns each new use case into an iteration, not a new procurement cycle.
2. Agents do the work no one had time for. Most “falling through the cracks” is simply undone work; agents extend the workforce to reach it.
3. Access is a triage problem, not only a staffing problem. Routing by urgency, complexity, and modality let UTMB meet endocrinology demand with its existing team and see high-complexity nephrology patients twice as early.
4. Automation gives clinicians their vocation back. Cleveland Clinic cleared a ~60,000-case backlog and increased patient volume to their actionable findings clinic by 44%, all while raising job satisfaction.
5. Clinicians lead, governance is centralized, and rigor scales with risk. Physician-owned criteria, one front door approval, time-bound KPIs, and validation matched to the stakes are what makes agents durable.

Turn your best ideas into reality

Every health system is full of ideas like these, waiting for a path forward. If you're ready to move from early pilots to agents that run across the enterprise, Bunkerhill Health can help. Request a demo at bunkerhillhealth.com.