Your hospital’s moat is not in Epic.

In healthcare, the system of intelligence lives at the engagement-revenue boundary.

Healthcare’s systems of record are well known. Epic, Oracle Health, MEDITECH. The patient engagement platforms (Phreesia, Luma, Artera). The revenue cycle platforms (Waystar, Availity, R1). The eligibility services, the population health tools, the practice management systems. These platforms are entrenched, and they will remain so. They are not going to be replaced.

What healthcare does not yet have is the system of intelligence that sits above them. The layer that reads across the EHR, the portal, the RCM stack, the call center, the scheduling system, the population health platform. The layer that orchestrates context, decides what matters in a given moment, writes back into the appropriate system, and increasingly takes action in the world.

In practice, this layer sends the message to the patient who is missing a follow-up. It drafts the documentation prompt for the clinician about to undercode an encounter. It routes the 3am family call to the right nurse. It identifies the claim about to be denied for a fixable reason. It surfaces the population health gap and assigns the action.

Geographically, the healthcare system of intelligence lives at the engagement-revenue boundary. Patient Engagement is the front face. It is the orchestration that determines whether the right patient gets the right contact at the right time, in a form they can use, before small problems become bigger ones. Revenue Cycle is the back face. It is the orchestration that determines whether the right work gets captured for the right payer, with the right documentation, and denials prevented rather than chased downstream. AI capability is the engine that makes the integration practical now.

The two are inseparable. McKinsey’s January 2026 analysis “Agentic AI: The race to a touchless revenue cycle” projects that AI enablement of healthcare revenue cycle could cut cost to collect by 30 to 60 percent. For a 6 billion dollar health system, where revenue cycle costs typically run 2 to 3 percent of revenue, the same analysis projects 60 to 120 million dollars in annual savings. The financial gravity of the back face is now visible to every CFO.

But the back face only works when it reads from the front face. Documentation that supports clean claims comes from clinical workflow. Denial prevention rules come from payer plan rules. Both depend on reading across the same set of underlying systems and writing back into them. This is what the system of intelligence does. It is what the existing vendor stack cannot do on its own.

Patient engagement is the front face of your system of intelligence. Revenue cycle is the back face. They are two halves of one strategic asset.

The technology stack is a commodity. The institution’s system of intelligence is not.

Every comparable hospital runs the same EHR. Every comparable hospital has the same vendor-supplied modules, the same patient engagement platform candidates, the same revenue cycle software pool. The technology stack is, increasingly, a commodity.

What is not a commodity is the system of intelligence the institution builds on top of it. That layer encodes everything specific to this hospital: how it actually delivers care, how its payer relationships actually work, how its providers actually document, how its patients actually behave, how its escalation paths actually run. It accumulates institutional knowledge no vendor can extract and no competitor can replicate. It is the one piece of a hospital’s technology stack that is genuinely its own.

In the language of the Andreessen Horowitz piece, “institutional memory becomes something a company can actually ship.” For a hospital, that institutional memory is the most valuable strategic asset on the balance sheet.

Hospitals that build a strong system of intelligence early will compound. Care quality improves because the intelligence layer acts on richer context. Financial performance improves because documentation is captured upstream and denials are prevented rather than chased. Patient satisfaction rises because the institution acts on what it knows. Staff burnout falls because the right work is routed to the right people. Each gain feeds the next.

Hospitals that delay will keep buying point solutions, wondering why none of them stick, and watching the AI-era advantage migrate to the institutions that figured out where it actually lived.

The Bain and KLAS Research 2025 healthcare AI study of 228 provider and payer executives makes the point in numbers. 70 percent of providers and 80 percent of payers now have an AI strategy in place or in development. Providers’ top investment priorities are revenue cycle management and clinical workflow. The strategies are forming. The investments are happening. The American Medical Association’s 2026 Physician Survey confirmed the same inflection from the clinician side: 81 percent of US physicians now use AI in practice, more than double the rate from 2023.

The question is no longer whether to build the system of intelligence. The question is whether the institution builds it, or rents it from someone else.

Your hospital’s moat is not in Epic. It is in the layer above Epic.

In May 2026, OpenAI and Anthropic paid roughly 15 billion dollars to back the same operating model.

This is an architectural decision, not a vendor decision. The CEO, COO, CMIO, value-based care architect, and integrated population health leader have already noticed that the engagement face and the revenue face are inseparable. The question is whether the institution builds the layer that ties them together, or keeps buying point solutions that paper over the gap.

In May 2026, the two largest AI companies in the world made the same bet, with billions of dollars of committed capital.

On May 4, Anthropic announced a joint venture with Blackstone, Hellman & Friedman, and Goldman Sachs to build “a new AI-native enterprise services firm that will work with companies to rapidly bring Claude into their core business operations.” The new firm is “a standalone entity with Anthropic engineering and partnership resources embedded directly within its team.” The consortium also includes General Atlantic, Leonard Green, Apollo Global Management, GIC, and Sequoia Capital. Healthcare is named explicitly as a target sector. Reported committed capital is approximately 1.5 billion dollars.

On May 11, OpenAI launched the OpenAI Deployment Company, majority-owned and controlled by OpenAI, with more than 4 billion dollars of initial investment (reportedly valued at 14 billion dollars by partner firms). The 19-firm consortium is led by TPG, with Advent, Bain Capital, and Brookfield as co-lead founding partners, and includes B Capital, BBVA, Emergence Capital, Goanna, Goldman Sachs, SoftBank, Warburg Pincus, and WCAS, plus consulting integrators Bain & Company, Capgemini, and McKinsey. The venture’s stated purpose, in OpenAI’s own words, is to “embed engineers specialized in frontier AI deployment, known as Forward Deployed Engineers, or FDEs, into organizations working on complex problems in demanding environments.” To accelerate from day one, OpenAI simultaneously acquired Tomoro, inheriting approximately 150 Forward Deployed Engineers and Deployment Specialists.

Both companies adopted the operating model Palantir invented and documented: embed senior engineers inside the customer’s operational environment, build production-ready workflows on the company’s stack, stay accountable until the system runs in production. Palantir CEO Alexander Karp made the architectural argument explicit in his February 2025 book The Technological Republic, an instant New York Times number-one bestseller.

The Forward Deployed Engineer model is not new. What is new is that the two most well-capitalized AI companies in the world just paid billions of dollars to validate it as the right way to deploy frontier AI into the enterprise.

The operating model is settled. The remaining question is who deploys it on behalf of the institution, and on what commercial terms.

In May 2026, OpenAI and Anthropic paid roughly 15 billion dollars to validate the operating model Digital Scientists has been running in healthcare since 2007.