71% of Hospitals See AI Value But Can't Scale Beyond Pilots

Most hospital AI pilots succeed in demonstrating value, yet the vast majority of healthcare organizations cannot translate those wins into enterprise-wide deployments, according to new survey data from Carta Healthcare.

The research found that 71% of healthcare organizations reporting measurable value from AI initiatives still are not expanding those projects at pace—a statistic that shifts the conversation from whether AI works to why proven concepts remain trapped in isolated demonstrations.

Integration replaces trust as the primary obstacle

EHR integration difficulty emerged as the leading barrier to AI adoption, cited by 44% of survey respondents. That figure substantially exceeded concerns about clinician trust and regulatory issues, each identified by only 26% of participants.

Brent Dover, CEO of Carta Healthcare, which conducted the survey, said the finding reflects where AI ultimately succeeds or fails: inside clinicians' existing workflows. Tools that require clinicians to step outside their normal documentation and decision-making processes add friction that quietly kills adoption, regardless of technical accuracy.

Integration extends beyond technical interfaces to encompass the contextual nature of clinical documentation. Extracting meaningful information requires understanding clinical intent rather than simply moving data between systems, Dover noted in an interview with Healthcare IT News, which first reported the survey results.

Clinical leaders take ownership while governance gaps persist

The survey revealed that clinical leaders now most frequently own AI strategy, surpassing both IT organizations and executive leadership. However, 26% of respondents reported having no clearly defined AI owner at all.

Dover characterized these findings as two sides of the same challenge. When clinical leadership owns AI strategy, it signals that these initiatives are viewed as clinical decisions with clinical consequences rather than technology experiments. Yet the absence of clear ownership in one-quarter of organizations means promising projects lose momentum once initial pilot enthusiasm fades.

The research suggests successful governance models anchor responsibility with clinical leadership while IT handles integration, security, and infrastructure and executive leadership provides funding and oversight.

Why it matters

Healthcare has moved past the proof-of-concept phase for AI. The operational barriers now blocking scale—workflow integration, governance structure, and vendor accountability—are organizational challenges that require leadership decisions rather than better algorithms. Hospitals that solve these implementation problems will gain competitive advantage through consistent clinical and operational improvements, while those that continue launching disconnected pilots will see diminishing returns on AI investment.

Vendors face higher scrutiny

Ninety-two percent of survey respondents said deep clinical domain expertise is critical when evaluating AI vendors, reflecting lessons learned through implementation experience. Healthcare organizations increasingly demand evidence of measurable outcomes at peer institutions and expect vendors to demonstrate results under real operational conditions rather than controlled pilot environments.

Dover recommends that health systems require vendors to prove integration within their specific environment, using their data and workflows, before contract signing. Organizations should favor vendors capable of delivering information directly into existing workflows rather than requiring internal teams to absorb additional complexity.

The survey findings indicate healthcare buyers have become more disciplined, shifting attention from technical capability alone toward long-term operational performance and workflow fit.

The survey details were first reported by Bill Siwicki at Healthcare IT News.