A new survey from Carta Healthcare finds that 71% of healthcare organizations with successful AI pilots still are not expanding those initiatives at pace, and EHR integration has overtaken clinician trust as the leading barrier to adoption. The results suggest the industry has moved past proving AI's potential and is now wrestling with the operational realities that determine whether AI becomes part of everyday clinical practice.
Pilots succeed, scaling stalls
The survey found that 71% of healthcare organizations reporting measurable success from pilot projects still are not expanding those initiatives at pace. Brent Dover, CEO of Carta Healthcare, which organized the survey, said the statistic changes the conversation.
"The most useful thing about this finding is what it rules out," he said. "The pilots proved that it does. What stalls is everything that comes after the proof."
Pilots succeed because they are intentionally narrow, Dover said. They typically have a committed champion, a clearly defined workflow and concentrated organizational attention. As organizations attempt to spread those projects across departments, those favorable conditions disappear.
"Success in a controlled setting tells you the model is capable," he said. "It does not tell you an organization is ready to operationalize it."
Health systems instead encounter competing priorities, budget pressures, inconsistent workflows and unclear accountability. Those are operational challenges rather than AI shortcomings, Dover argued.
EHR integration is the top barrier
The survey found that EHR integration remains the leading barrier to AI adoption, cited by 44% of respondents. That placed it well ahead of clinician trust and regulatory concerns, each identified by 26%.
"A tool can be accurate in a demo and still be useful in real time the moment a clinician has to step out of the workflow to use it," Dover said. "Anything that sits outside the workflow adds steps, and added steps are where adoption quietly dies."
Integration extends beyond technical interfaces. Extracting meaningful data requires understanding clinical intent rather than simply moving data between systems. The findings reflect a broader shift in AI for Healthcare adoption, where the focus has moved from proving value to managing integration and workflow.
Dover recommends treating integration as a core purchase requirement, not an afterthought. "Ask vendors to prove integration in your environment, with your data and your workflows, before you commit," he said. Health systems should also evaluate who carries the integration burden, favoring vendors capable of delivering trustworthy information directly into existing workflows rather than requiring internal teams to absorb added complexity.
Clinical ownership is emerging
Clinical leaders now are most frequently the owners of AI strategy, surpassing IT and executive leadership. At the same time, 26% of respondents reported having no designated AI owner.
"A pilot without an owner is a permanent pilot," Dover said. "There is no path to scale."
He advocates a governance model that places clinical leadership at the center, with IT leading integration and technical considerations and financial leadership supporting execution. One clinical leader should remain responsible for adoption and results, he said. For AI for Executives & Strategy decisions, the survey suggests governance and ownership deserve as much attention as the technology itself.
Vendor evaluation
Nearly all respondents - 92% - said deep clinical domain expertise is critical when evaluating AI vendors.
"Generic AI, however impressive, does not automatically understand the difference between a well-documented case and a clinically ambiguous one," Dover said.
He suggests health systems should reference providers with demonstrated outcomes, ask for references in comparable settings, and require evidence of successful go-live experiences.
"The investments that do scale are evaluated on different questions: Does the vendor understand the clinical workflow well enough to be trusted with it?" he asked. "Will they show case outcomes and not just promises?"
Dover said the survey also indicates that healthcare providers are becoming more disciplined in their approach to AI. They are looking for measurable outcomes, documented performance and integration confidence, rather than just potential.
Why this matters for Executives and Strategy
For executives, the survey's findings point to a shift in how AI investments should be evaluated. The question is no longer whether AI can work - it already does. The real work is in turning strong pilots into software that the entire hospital trusts.
Dover believes the departments that are succeeding treat AI as a portfolio, not just a pilot. They allocate resources for ongoing integration, evaluate both clinical and operational outcomes, and establish clear ownership across regular line management. That means clear accountability, proven vendor partners, and a genuine focus on how AI fits into the clinician's actual routine.
Your membership also unlocks:
