The Practice

Most AI pilots in health systems fail not at the model layer — but at the infrastructure layer beneath it. The order matters because patients are downstream.

Principles drawn from a decade of implementation across U.S. academic and integrated health systems.

01

Interoperability as a foundation

Connected data is the prerequisite. Lab-to-EHR consolidations, FHIR APIs, MCP server architecture — every meaningful AI deployment rests on whether systems can speak to one another in a governed, secure way.

02

Analytics as a network effect

Enterprise analytics scale when infrastructure is shared. A platform introduced into one operational setting at scale tends to spread organically — an early demonstration of the network effects that well-sequenced infrastructure decisions produce.

03

Agentic AI as thought partner

The third phase is not another tool. It is action taken on top of connected data — AI agents operating across clinical and operational workflows with the context required to deliver value rather than noise.

Operational Footprint

Gatwiri Mwiti is a healthcare infrastructure and operations leader with over a decade of experience across UC San Diego Medical Center and Kaiser Permanente. She holds a Master of Applied Science in Population Health Management from the Johns Hopkins Bloomberg School of Public Health. She writes Perspectives biweekly on the operational layer of healthcare, emerging science and technology, and the public conversations that haven't yet happened.