Beyond the exam room: How top health systems use artificial intelligence to reshape nonclinical operations
Article | October 21, 2025
Reading time: 5 mins
When Stanford Health Care President and CEO David Entwistle kicked off a panel discussion at a 2025 Vizient AMC CEO Network meeting with the line, “We’re going to explore a topic no one is really talking about,” he knew it would be met with knowing laughs. After all, the subject at hand was AI — an industry gamechanger that’s top of mind for just about everyone.
But when it comes to healthcare, Entwistle argued, there needs to be more talk about AI’s application in nonclinical settings.
That’s exactly the topic Entwistle — along with Cleveland Clinic Northeast Ohio Market President Jorge Guzman — tackled. While much attention has been given to AI’s role in diagnostics and clinical care, these leading health systems are finding transformative value in administrative, operational and support service functions. From streamlining surgical scheduling and staffing to reimagining facilities maintenance, the conversation underscored how AI is quietly but profoundly reshaping the healthcare enterprise from the inside out.
Stanford: Building trust through internal assessment
Entwistle opened the session with insights from an internal survey designed to understand employee attitudes toward AI.
“Before we start to wholesale huge applications of AI, we wanted to understand within our organization what the thoughts, fears and anxieties are with its uses,” he said.
The findings showed stark generational and role-based differences in adoption, with physicians and younger workers more likely to use AI tools daily.
Stanford Health Care has already implemented dozens of AI-powered applications, including tools that help with inbox management and note generation. In some cases, Entwistle said, the benefits have been different from what was originally expected. For example, while the majority of AI-generated clinical notes need editing, clinicians report that the application saves significant cognitive effort.
Entwistle emphasized the need for systematic evaluation, citing Stanford Health Care’s “FURM” model, which assesses AI applications based on whether they are Fair, Useful and Reliable.
“There are a lot of shiny new things out there,” he said. “We always have to ask: Which ones are adding value?”
Cleveland Clinic: Boosting throughput and staffing precision
Guzman shared how the system uses AI to unlock capacity and improve resource management in a region with persistently high patient volumes.
“We are using AI to forecast OR cases with a foresight of 100 days in advance,” he said, adding that these forecasts help anticipate supply needs, block scheduling and staffing — allowing the system to react before bottlenecks occur.
Looking ahead, the system plans to use AI for procurement and is exploring agentic AI — autonomous systems capable of making decisions — to enhance efficiency in backend operations. Guzman also highlighted the system’s high hopes for ambient AI in clinical documentation.
“I have never seen doctors so excited,” he said. “They’re almost giddy.”
Guzman also emphasized that successful AI implementation hinges on culture and communication.
“Never underestimate the ‘why’ and the ‘how’,” he said. “Transparency in model performance — like showing how AI predictions match reality — helps build confidence and accelerate adoption.”
Key takeaways: How AI is revolutionizing nonclinical functions
- Administrative tasks: Reducing manual workload for scheduling, documentation and revenue cycle management.
- Operational efficiency: Forecasting OR usage, predicting patient admissions and optimizing staffing.
- Supply chain management: Enhancing procurement efficiency and ensuring the right supplies are available at the right time.
- Facilities and maintenance: Predictive analytics for equipment maintenance and infrastructure management.
Stanford Health Care uses a multi-step, multidisciplinary process that analyzes the technology and potential impact, as well as fairness, reliability and usefulness before implementing AI tools. Cleveland Clinic assesses AI solutions based on their scalability, efficiency, impact on patient care and cost-effectiveness.
Both institutions perform extensive validation studies, often through research initiatives or real-time monitoring. They measure AI accuracy against actual outcomes and ensure transparency to build user confidence.
AI implementation is driven by staff engagement, grassroots innovation challenges and automation hubs that provide technical support and resources for employees to integrate AI solutions into their workflows. It’s important to emphasize AI as a tool to enhance roles rather than replace them. Staff are encouraged to participate in AI-driven innovation to shape how AI improves their day-to-day work.
AI-driven scheduling tools optimize OR and hospital staffing, reducing inefficiencies and improving capacity planning. AI models predict patient demand, allowing better nurse and physician allocation. Additionally, AI automates repetitive tasks, such as meal delivery via robotic assistants, predictive maintenance for facility management, and streamlining linen and central service workflows.
AI helps reduce the burden of manual paperwork by automating prior authorizations, coding processes and insurance claim reviews. It has significantly cut down administrative processing times.
The biggest challenges include staff resistance to change, concerns over job displacement, ensuring AI models are validated and bias-free, and aligning AI applications with organizational goals.
- Improved ambient AI for automatic documentation and real-time patient interaction
- Enhanced predictive analytics for hospital throughput and capacity management
- Streamlined call center operations that lead to reductions in wait times for scheduling and administrative inquiries
- Revolutionized supply chain management through agentic AI that reduce inefficiencies in procurement and distribution