ipopba - stock.adobe.com
KLAS: Ambient speech remains healthcare's top AI use case
Leaders are cautious about higher stakes, care-transformative applications, a new KLAS report shows.
Healthcare organizations are increasingly adopting AI, but use remains concentrated in low-risk, efficiency-focused workflows rather than advanced clinical applications that could transform care delivery, according to a new KLAS Research report.
The report, Healthcare AI Update 2025: What Use Cases Are Adopted the Most?, draws on responses from 3,370 individuals across 1,742 unique healthcare organizations.
According to respondents, financial pressure, staffing shortages and the need to demonstrate a clear return on investment (ROI) are shaping how and where AI is deployed. Nearly all interviewed organizations reported piloting or using some form of AI, yet few have expanded adoption broadly across departments.
Ambient speech and imaging lead clinical AI adoption
Ambient speech remains the most widely adopted clinical AI use case, with 79% of organizations using it -- along with reports of positive ROI. Clinicians cited ambient speech leading to improvements in EHR documentation. Other clinical applications, such as chart summarization and decision support, remain far less common.
Imaging is one of the few other clinical areas where AI use is well established. Organizations reported using AI for stroke triage, mammography and cardiology, primarily to flag urgent findings and speed turnaround times. KLAS attributed imaging's broader adoption to its longer history with AI-enabled tools. It also noted that most imaging applications are designed to support prioritization rather than to replace clinical judgment.
Operational and administrative use cases continue to grow
Outside of direct clinical care, organizations reported a growing use of AI for administrative and operational tasks. Common use cases include meeting transcription, document creation and other routine office work, for which organizations often use widely available platforms such as ChatGPT or Microsoft Copilot.
Revenue cycle management is also drawing increased attention. Organizations reported using AI for claims adjudication, coding automation and denials management. KLAS found that both payer and provider organizations are prioritizing revenue cycle AI as reimbursement pressures increase.
Vendor adoption favors established players
Organizations continue to favor AI offerings from larger vendors with existing relationships.
Microsoft, Epic and OpenAI were the most frequently used or considered vendors across AI categories. Epic users reported the broadest range of AI use cases, spanning clinical, administrative and revenue cycle workflows.
In ambient speech, Microsoft and Abridge were the most frequently considered third-party vendors. At the same time, EHR vendors, including Epic, Oracle Health, athenahealth and eClinicalWorks, are rolling out native ambient speech tools, which some organizations prefer for easier integration and less vendor complexity.
Agentic AI adoption remains limited
Despite growing vendor marketing, adoption of agentic AI remains limited. Of more than 3,000 respondents interviewed, only 17 specifically mentioned agentic AI, and just one organization reported active use.
Most respondents said they have not selected vendors for agentic AI, stating that data readiness challenges, implementation complexity and unclear ROI are barriers. KLAS found that adoption is far below the current market hype as organizations focus on stabilizing simpler AI use cases first.
KLAS' findings suggest broader AI adoption will depend less on new capabilities and more on proving ROI, strengthening governance and integrating tools into everyday workflows. For now, most organizations appear set on expanding cautiously rather than quickly.
Elizabeth Stricker, BSN, RN, comes from a nursing and healthcare leadership background, and covers health technology and leadership trends for B2B audiences.
