Clinical Expertise Emerges as Critical Factor in Healthcare AI Adoption

Carta Healthcare, the leader in enterprise clinical data management, released findings from a national survey of U.S. healthcare leaders showing that AI is proving its worth but failing to expand, and that leaders will not trust it at scale without deep clinical expertise. In the survey's most lopsided result, 92% said deep healthcare domain expertise is critical when evaluating an AI vendor, with half of all respondents assigning it the maximum possible score. Even where AI has delivered measurable value, 71% of organizations are not expanding it at pace.

The findings reframe a debate that has focused on whether or not AI belongs in clinical settings. The data suggests that question has largely been answered, though adoption stalls for structural reasons. Organizations validate AI in a pilot, see results, and then fail to expand it.

When asked what slows AI adoption, respondents pointed to operational friction. Difficulty integrating with the electronic health record (EHR) was the top barrier at 44%, followed by lack of executive sponsorship or budget (37%) and competing organizational priorities (33%). Concerns like clinician trust, ongoing cost, and regulatory uncertainty tied far lower at 26% each.

Nearly half of respondents (48%) said a solution that fits inside the EHR would most help accelerate adoption, ahead of peer case studies with measurable outcomes (36%) and followed by vendors willing to assume performance risk (28%). Taken together, the findings indicate that healthcare buyers want proof, not slideware.

"For years the conversation has been about whether healthcare AI is good enough. This data tells us that's the wrong question," said Brent Dover, CEO of Carta Healthcare. "Organizations are seeing value and still can't scale it. The biggest barrier to adoption is whether the technology fits the way their teams already work. When a solution lives inside the EHR and proves itself against measurable outcomes, adoption follows."