Physicians were right. Their hospitals are finally catching up.

For two years, the pattern has been clear: physicians are using AI heavily in practice, but most say their employers still aren’t deploying it well. In 2026, AMA data showed 81% of physicians using AI professionally, while survey reporting found 67% using it daily and 89% weekly.

So many clinicians adopted ChatGPT, Claude, Abridge, DAX, and other tools on their own, because the workflow demand was already there. The institutions moved more slowly than the people inside them.

Then this happened.

March 31: Mount Sinai announced its first enterprise-scale OpenEvidence deployment, embedded directly into Epic for physicians, nurses, and pharmacists.

April 30: Echo IQ announced EchoSolv AS was deployed across Mount Sinai for severe aortic stenosis decision support.

Two clinical AI rollouts in 30 days. Both at the point of care. Neither framed as a pilot.

This is what happens when a hospital stops treating clinical AI as a future category and starts treating it as workflow infrastructure.