Black Book Research has released results from its AI in the EHR Adoption Readiness Poll, a survey of 507 healthcare executives, informatics leaders and clinical operations staff fielded across the second and third quarters of 2026. The headline finding: artificial intelligence has genuinely moved out of pilot programs and into daily electronic health record workflows at U.S. hospitals, but the operating discipline needed to run it safely at scale is still missing almost everywhere.
A Readiness Score Stuck in ‘Emerging’
Black Book built a 200-point composite benchmark, the AI in the EHR Readiness Score, scored across six domains: AI strategy and governance, ambient AI documentation, clinical decision support and generative AI, workforce readiness and change management, infrastructure and EHR integration, and organizational risk tolerance. The average score across all 507 respondents came in at just 94.9 out of 200, placing the industry as a whole in Black Book’s ‘Emerging’ band rather than ‘Advancing’ or ‘Leading.’
Most Hospitals Are Behind, a Handful Are Ahead
The distribution is lopsided: 284 of the 507 respondents, or 56 percent, scored below the Advancing threshold, meaning more than half the industry has barely begun building the governance structures AI at scale requires. Only 30 respondents, under 6 percent, reached the top ‘Leading’ band. That gap suggests a small group of well-resourced health systems are pulling meaningfully ahead of a large trailing majority still running AI pilots without a formal governance model behind them.
Where AI Is Actually Working
Ambient AI documentation, tools that listen to a clinician-patient conversation and auto-generate visit notes, emerged as the domain advancing fastest in the poll, consistent with a broader wave of hospital systems adopting ambient scribes over the past two years to cut physician burnout and after-hours charting. Clinical decision support and generative AI features embedded directly in the EHR are also gaining ground, according to Black Book, but more unevenly across organizations of different sizes and budgets.
The Barriers That Won’t Move
Governance, clinician trust, workforce readiness and post-go-live monitoring were flagged as the persistent barriers to enterprise-scale adoption. That finding echoes other 2026 industry surveys: a separate poll found EHR vendor reliance cited by 74 percent of health system leaders as an obstacle to executing their AI strategy, while workflow integration difficulty was named by 72 percent of respondents across clinical and IT roles. Only about 4 percent of IT leaders in related surveys say they have scaled an AI deployment with measurable, reportable outcomes.
Why the Gap Persists
Health IT executives interviewed around the poll’s release argue that hospitals have been quick to buy AI point solutions from vendors but slow to build the underlying governance committees, monitoring dashboards and change-management programs that let those tools run safely once deployed to thousands of clinicians. Vendors, for their part, argue that health systems often lack the data infrastructure and standardized workflows to fully exploit tools that could otherwise reduce documentation burden and catch clinical errors earlier.
What It Means Going Forward
Black Book’s poll suggests the health AI market is entering a harder, less glamorous phase: the tools already work well enough to deploy, but hospitals now have to build the boring infrastructure, governance boards, audit trails, ongoing performance monitoring, that turns a promising pilot into something a regulator or malpractice attorney would call safe at scale. Expect the next year of health-system AI spending to shift from buying new models toward hiring the governance and informatics staff needed to actually run the ones already installed.
