The pace of AI investment in healthcare has few modern parallels. Health systems are signing enterprise agreements, standing up governance committees and announcing AI strategies with a frequency that would have been unimaginable five years ago. The tools are more powerful, the vendors are more sophisticated and the pressure from boards and executives to demonstrate AI capability has rarely been more acute.
Yet the clinical leaders responsible for making AI work inside these organizations are sounding alarms. In responses gathered by Becker’s, more than two dozen CMIOs, CMOs, CNOs and health IT executives described what they believe is going wrong — not with the technology itself, but with the potential strategic and organizational decisions being made around it. Health systems moving too fast on AI risk getting the fundamentals wrong.
“The most dangerous thing happening in health IT is the assumption that technology adoption equals transformation,” said Hetal Rupani, senior director of business intelligence and analytics at Johns Hopkins Medicine in Baltimore. “It doesn’t.”
The strategic issue most often named is speed without structure. Organizations are under enormous pressure to implement AI quickly, and that pressure may lead to tools going live before the governance structures, workflow redesigns and clinician training needed to make them work are in place. The result is technical implementation that fails to reach value.
“In healthcare, successful digital transformation is not measured by implementation alone; it is measured by whether technology improves care delivery, reduces burden on clinicians, and drives better patient outcomes,” said Ms. Rupani.
The urgency driving that speed is understandable. New AI capabilities are emerging weekly, and organizations that move slowly risk falling behind competitors who do not. But several respondents argued that health systems run the risk of selecting tools because the tools are available, compelling and on-trend versus solving the real problem.
“Hospitals and leaders are being overwhelmed by AI and other technology vendors knocking on our doors every day, each promising to improve quality, efficiency, and revenue cycle,” said Mohamed Rami Nakeshbandi, MD, vice president and CMO at Downstate Health Sciences University in New York City. “But we have very limited standardized ways to objectively evaluate which solutions truly meet our clinical need, deliver meaningful value, and have the real return on investment.”
Nipa Shah, MD, chair of community health and family medicine at the University of Florida College of Medicine in Jacksonville, sees the opposite as detrimental as well. When the landscape shifts faster than decision-making cycles, leaders become hesitant to commit.
“By the time a decision is made, the landscape may have already shifted,” Dr. Shah said. “While caution is understandable, the greater risk may be stagnation — delayed decisions, missed opportunities, and an inability to realize the benefits of innovation.”
The strategic problem compounds when AI is layered onto processes that were never examined before automation. Enitza George, MD, associate professor and chief population health officer at SUNY Downstate Health Sciences University, argued that this is where the deepest risks live in what the tools are being asked to do.
“Technology scales whatever system we give it, including its inefficiencies and bad assumptions,” Dr. George said. “So before asking whether a process can be automated, we need to ask whether that process should exist in its current form at all. Once you automate a flawed process, you don’t just preserve the flaw, you scale it. You make it faster, harder to see, and potentially much harder to undo.”
In healthcare, the consequence of that failure is not confined to wasted capital. It can move directly into clinical decision-making, patient access and care equity. The scale and speed that make AI valuable in theory are the same properties that make a poorly designed AI deployment dangerous in practice.
The issues become more acute as the technology evolves. Several respondents flagged the transition from generative AI — tools that summarize, suggest and generate content — to agentic AI, tools that act autonomously inside clinical workflows, as a threshold moment for health system strategy.
“We are moving from AI that summarizes and suggests to AI that can increasingly act: preparing orders, managing inbox work, coordinating care, and executing multi-step workflows inside the EHR,” said Tomi Kolade, MBBS, assistant CMIO at UTHealth Houston. “That shift changes the stakes entirely.”
The organizational response to that escalation, several respondents argued, has not kept pace. Governance frameworks built for generative AI are not necessarily adequate for agentic AI, and the clinician involvement required to design agents safely is a higher bar than reviewing a documentation tool after the fact.
“As we enter the agentic era, physicians cannot simply be end users brought in at the end to validate what has already been built,” Dr. Kolade said. “They need to be embedded in the design and validation loop from day one, helping define what the agent should do, when it should act, when it should stop, and when human judgment must take over.”
The question of who is at the table when AI is designed came up repeatedly across respondents. Terry McDonnell, DNP, RN, senior vice president and chief nurse executive at Duke University Health System in Durham, N.C., argued that the gap between clinical advisory boards and genuine co-development is where strategic AI investment most frequently fails.
“The best solutions are driven by the people who are closest to the work,” Ms. McDonnell said. “Having clinical advisors or an advisory board is different from co-development. When executives make decisions to purchase and implement solutions that have not had the patient and clinician involvement to ensure that the right solutions and changes are taking place, that creates waste and we can’t afford to introduce waste into our already complex systems.”
That waste is more than financial; it is the erosion of clinician trust that happens when AI tools fail to deliver on their promises, and that has strategic consequences for every subsequent AI launch.
Deepa Velayadikot, MD, medical director of care coordination and regional medical director of hospital medicine at Cooper University Hospital in Camden, N.J., described a pattern of adoption driven by trend rather than need. Health systems may see substantial AI investments producing little measurable return, increased clinician burden and a loss of trust when tools underdeliver.
“Without thoughtful evaluation, strong governance, and alignment with organizational goals, AI risks becoming an expensive buzzword rather than a meaningful tool for healthcare transformation,” Dr. Velayadikot said.
What separates the health systems getting AI right from those accumulating expensive, underperforming deployments is a question of the organizational discipline. The technology is not the differentiator. The strategy around it is.
“The organizations that succeed won’t be the ones moving fastest,” said Stephen Bardoczi, vice president of planning and post acute care emeritus at Witham Health Services in Lebanon, Ind. “They’ll be the ones disciplined enough to fix the foundation first, simplify where they can, and require every new capability to earn its place before it’s trusted with patient care.”
At the Becker's 11th Annual IT + Revenue Cycle Conference: The Future of AI & Digital Health, taking place September 14–17 in Chicago, healthcare executives and digital leaders from across the country will come together to explore how AI, interoperability, cybersecurity, and revenue cycle innovation are transforming care delivery, strengthening financial performance, and driving the next era of digital health. Apply for complimentary registration now.
