In the high-stakes world of modern medicine, executive leadership is often tested not by the complexity of the clinical challenges faced, but by the resolve required to pivot institutional culture. For Marc Harrison, M.D., a veteran healthcare executive and Chair of the TowerBrook Healthcare Institute, the lesson was learned early: institutional change does not happen through whiteboard diagrams or pilot programs—it happens when the CEO treats a goal as a top-tier operational priority.
Today, that priority is Artificial Intelligence. As the $5.3 trillion healthcare industry faces a "perfect storm" of workforce shortages, administrative bloat, and reimbursement compression, AI stands as the only cross-cutting lever capable of simultaneously addressing cost, quality, and access. However, according to Dr. Harrison, the window for health systems to seize this advantage is closing, and the cost of inaction is no longer just a fiscal concern—it is a threat to the sustainability of the entire American healthcare infrastructure.
The Core Mandate: From Pilot Programs to Enterprise Reality
For many health systems, AI remains trapped in the "innovation theater" phase: a collection of disparate pilots, slide decks, and experimental software that rarely gains traction. Harrison argues that this failure is not technological; it is a failure of leadership.
"The team doesn’t follow the plan on the whiteboard," Harrison notes. "They follow whatever the leader is paying attention to."
In an era where healthcare spending is hurtling toward 20% of the U.S. GDP by 2033, the integration of AI is not merely a digital upgrade—it is a fiscal and national security imperative. As federal deficits grow and corporate competitiveness is eroded by the ballooning costs of employee health benefits, hospitals that fail to automate are essentially choosing to let third-party tech entities extract the value that should be fueling their own missions.
Chronology of a Shift: Why Now?
The transition toward AI-driven healthcare is moving at an exponential pace. While healthcare innovation historically moves in cycles of years, the advancement of AI models is now measured in months.
- The Era of Stagnation: For decades, health systems relied on massive administrative overhead—which now consumes one-third of all healthcare spending—to manage complexity.
- The Breaking Point: The post-pandemic landscape introduced severe labor shortages and record-high burnout rates, making the status quo operationally untenable.
- The AI Inflection: With the arrival of scalable, high-utility generative and predictive AI, the tools to automate, predict, and optimize care are finally mature.
- The Current Moment: We are now in the "mobilization phase." The gap between organizations that have integrated AI into their governance and those that are still "evaluating" is widening into a chasm that will prove impossible for laggards to bridge.
The Strategic Framework: Six Imperatives for CEOs
Dr. Harrison posits that if AI is not among the CEO’s top three priorities, the organization is effectively choosing decline. To move from passive observation to active transformation, leadership must adhere to a strict set of imperatives.
1. Own the Transformation Personally
This is not a task to be delegated to the Chief Information Officer (CIO) or a committee. The CEO must possess enough technical fluency to make capital allocation decisions and must be the one to adjudicate resource conflicts. By tying the CEO’s own performance evaluations and compensation to AI-driven outcomes, accountability cascades down to the executive leadership team (ELT).
2. Govern with Speed, Not Consensus
Traditional health system governance is often a "senate" of stakeholders where the approval process for a project takes longer than the development of the technology itself. A lean, cross-functional unit—comprising clinical, financial, operational, legal, and technological leaders—must report directly to the CEO and the Board.
3. Build a Transformation Unit, Not an Innovation Center
Innovation centers often become silos of "play" that never reach the core business. A "Transformation Unit," conversely, is staffed by operators who understand where the money is currently being lost. Their mandate is to map the cost structure, identify automatable work, and execute a build-buy-partner strategy.
4. Start Non-Clinical to Build Trust
To gain the confidence of the clinical staff, systems must first prove value in areas where outcomes are easily measured: revenue cycle, supply chain, IT, and call centers. Success in these high-friction, non-clinical areas creates the necessary momentum and "proof of concept" before moving into the complexities of bedside care.
5. Extend into Clinical Domains with Evidence Gates
Clinical AI must be approached with rigorous evidence, regulatory mapping, and a plan for workforce transition. Leaders should prioritize areas where AI already exceeds human baselines—such as diagnostic imaging, ambient clinical documentation, and predictive risk stratification—ensuring transparency at every step.
6. Measure Everything and Scale What Works
"If it isn’t measured, it doesn’t exist." Every AI initiative must have a predetermined path to enterprise-wide scale. Successful pilots should not be celebrated in isolation; they must be expanded, while the leaders who fail to meet AI-enabled benchmarks must face consequences in their KPIs and compensation.
The "Build-Buy-Partner" Philosophy
One of the most critical arguments presented by Dr. Harrison is the rejection of the "do-it-all-yourself" mentality. Many health systems suffer from the delusion that they can transform into software development powerhouses overnight.
"Partnership is not weakness," Harrison asserts. "It is a recognition that this transition’s scale and speed exceed what any single institution can execute alone."
Health systems should focus on their core mission: the patient-provider relationship. For non-core services, they should contract with expert partners who possess the engineering scale and data breadth that an individual hospital system could never replicate. Attempting to build everything internally introduces unnecessary execution risk that should be borne by partners who underwrite the performance and cost-deflation outcomes.
Implications: The Future of the Health System
The structural advantages that have historically protected incumbent health systems—patient trust, regulatory standing, and community roots—are currently depreciating assets. As the cost of AI capability drops, well-capitalized technology giants are finding it easier to enter the healthcare space. If health systems do not use AI to secure their own efficiency, they will be relegated to the role of a commodity provider, with the value-capture occurring in the tech layers above them.
The next three to five years will serve as a sorting mechanism for the industry. Some organizations will be reshaped by AI, effectively losing control of their future. Others will redefine the very concept of a health system, using AI to solve the systemic crises of cost and access while preserving the human element of care.
Ultimately, the choice lies with the CEO. The technology is ready, the financial imperative is clear, and the path is well-defined. The only missing component in many boardrooms today is the courage to make AI not just a project, but the central, non-negotiable priority of the institution.
About the Contributor
Marc Harrison, M.D., is a global healthcare leader and physician executive. As the former President & CEO of Intermountain Healthcare and founding CEO of the Health Assurance Transformation Company (HATCo), he has been at the forefront of digital transformation in medicine. Currently, he serves as Chair of the TowerBrook Healthcare Institute and is a strategic advisor to General Catalyst, continuing his work to harmonize clinical excellence with operational innovation.
