Health systems shift from piecemeal AI pilots to enterprise-wide transformation strategy

Healthcare's AI adoption problem was never really about the models. It's about the org chart.

That's the thesis emerging from UCI Health, where Dr. Deepti Pandita serves as vice president of clinical informatics and chief medical informatics and AI officer. Ahead of her fireside session at the HIMSS AI in Healthcare Forum in San Diego on October 22, Pandita is making a case that should concern any executive still running AI as a series of departmental pilots: the return on investment simply isn't materializing under that model.

Health systems have spent years greenlighting isolated deployments. A chatbot here, a documentation tool there, an analytics dashboard for one unit. The logic made sense on paper. Smaller scope, faster wins, lower risk. But Pandita's read is blunt: that approach has hit a ceiling, and the interconnected nature of modern hospital operations is the reason why.

"An enterprise approach has become increasingly important because healthcare has grown too interconnected for isolated solutions to succeed," she said. A change in one department ripples through many others. AI, digital workflows, patient engagement tools, analytics and EHR modernization all cross traditional organizational lines. Treat them as siloed department projects, and the result is fragmentation: duplicated spend, inconsistent governance, and systems that can't scale past their original use case.

That's a capital allocation problem as much as a technology one. Every duplicated investment is money spent twice for a fraction of the value. Every inconsistent governance framework is latent compliance risk. For CFOs and boards evaluating digital transformation budgets, the piecemeal model carries a hidden cost structure that rarely shows up until the fifth or sixth pilot fails to scale.

Why alignment beats architecture

Pandita's central claim is uncomfortable for technology-first thinking: the hardest part of enterprise transformation is not the technology.

"The hard work is not usually the technology itself, it's creating alignment," she said. That means bringing clinical leaders, operations, IT, finance, compliance, quality, nursing, physicians and executives around a shared vision and shared measures of success. It requires governance, trust and transparency. It also requires something harder to institutionalize: a willingness to make decisions that serve the enterprise even when they aren't optimal for any single department.

That's a nontrivial ask. Departmental leaders are typically incentivized, formally or informally, to optimize for their own budget lines and outcomes. Asking a department to accept a suboptimal tool because it serves system-wide interoperability is a governance challenge, not an engineering one. Health systems making progress, according to Pandita, are the ones treating digital transformation as an organizational capability rather than a stack of individual projects. They are investing as heavily in change management, leadership alignment and culture as they are in the underlying software.

That investment mix matters for anyone tracking where health system dollars actually go. A hospital that allocates its digital transformation budget entirely toward licensing and infrastructure, with little left for change management or governance structures, is arguably underinvesting in the piece most likely to determine whether the technology gets used at all. Culture, in Pandita's framing, becomes the true scaling constraint, not the software itself.

This tracks with a broader pattern showing up across the sector. Related commentary from health system CIOs has stressed that digital transformation depends on strong foundations and unity of purpose, and that CIOs need to fix data and governance foundations before scaling AI further. The throughline across these perspectives is consistent: infrastructure and intent have to be aligned before technology investment pays off, not after.

Mass General Brigham's experience embedding AI across its enterprise offers a useful data point in the same direction. Rather than parceling out AI tools department by department, the system's approach has emphasized coordination across the organization, a structural choice that mirrors what Pandita is describing at UCI Health. The pattern suggests this isn't one system's idiosyncratic preference. It's becoming a recognizable strategic posture among health systems that have moved past the initial pilot phase.

For investors and operators watching the health IT vendor landscape, this shift has direct implications. Vendors selling narrow, department-specific point solutions may find themselves competing against a buyer preference that has moved toward platforms and governance frameworks capable of operating across an entire enterprise. That's a different sales motion, a longer procurement cycle, and a different buyer at the table. Point solution vendors that can't demonstrate enterprise interoperability or governance compatibility risk becoming harder sells as health systems consolidate their AI strategy under a single organizational umbrella.

Pandita frames the stakes plainly. "The future winners in healthcare digital transformation may not be the organizations that implement the most technology, they may be the organizations that are best at creating trust, alignment and organizational readiness around technology," she said. Those capabilities, she added, ultimately determine whether innovation stays a promising pilot or becomes meaningful enterprise transformation.

The Bottom Line

The evidence here is qualitative rather than quantitative. No ROI percentages or cost figures accompany Pandita's remarks, and this is a preview of a conference session rather than a peer-reviewed study. Readers should treat the claims as directional, an experienced practitioner's assessment rather than a benchmarked outcome.

Still, the signal is worth tracking. If enterprise-wide governance and alignment genuinely determine which health systems extract value from AI investment, that reshapes how vendors should be evaluated and how health system leadership should be structured. The rise of dedicated chief AI officer roles, Pandita's own title among them, is itself a governance response to this problem. Watch for whether more systems formalize similar roles, and whether vendors begin marketing governance and interoperability capabilities as aggressively as they market model performance. That shift, if it happens, will be the clearest confirmation that Pandita's thesis has taken hold beyond one conference stage.