Despite its increasing use, artificial intelligence integration is lagging at healthcare systems and organizations across the U.S., according to a new survey from Arcadia.
Fifty-two percent of respondents report that AI can fundamentally transform healthcare in correct applications, though about half (53%) of AI insights are “only partially embedded” into decision-making processes and only 14% report full integration at “key decision points.”
The healthcare platform’s survey drew insights from 281 healthcare leaders across provider, payer and service organizations at the March HIMSS26 conference.
“Healthcare leaders are increasingly aligned on AI's potential to improve care and drive measurable value, but many organizations are still working to operationalize those capabilities,” said Michael Meucci, Arcadia president and CEO, in a statement. “The next wave of AI value won't come from better models. It will come from better execution. Organizations that can successfully embed AI insights into everyday decisions and workflows will be best positioned to realize meaningful clinical, operational, and financial impact.”
Twenty-one percent of respondents report that AI is most valuable in specific scenarios, with another 21% reporting that solutions deliver the greatest value in conjunction with “strong human oversight.” Only 6% of respondents report viewing the technology as more risky than valuable or overhyped.
Daily decision-making was named as a challenge for 31% of respondents when scaling AI responsibly. Other needs for responsible adoption included education (27%), strengthening data foundations (22%) and measuring impact (20%).
“As healthcare leaders move beyond experimentation, success will increasingly depend on their ability to embed AI into decisions and workflows that improve care, reduce costs, and deliver measurable outcomes,” Meucci said.
Healthcare leaders are looking for AI to deliver measurable business outcomes, citing cost savings (33%), reduced workforce turnover (27%) and improved financial forecasting (21%) as the most important results.