AdventHealth's recent integration of OpenAI's ChatGPT into its operations is transforming the way healthcare is delivered, effectively giving back valuable time to clinicians and enhancing patient experiences. By automating extensive documentation and administrative tasks, the hospital system is improving operational efficiency and expanding its capacity to serve an increasing number of patients across nine states.
Addressing Operational Challenges
Operating with tight margins and growing demands, AdventHealth faces significant administrative complexities. The burden on staff often results in lengthy processes that detract from patient care. For example, physician advisors reviewing cases for utilization management previously spent around 10 minutes on each case, sifting through charts and drafting rationales. These inefficiencies accumulate quickly, creating ongoing operational strain.
To address these challenges, AdventHealth recognized a rising interest in AI among its workforce, despite restrictive policies on chatbot use. “We had folks who were eager to start, but there were a very large number of people who were on the sidelines,” said Rob Purinton, Chief AI Officer at AdventHealth. The decision to maximize AI adoption rather than running isolated pilots set the stage for scalable change.
Strategic Adoption of AI
A key decision in the AI rollout was to frame its purpose as alleviating administrative pressure rather than simply automating tasks. “We don’t talk about AI as automation. We talk about time back,” Purinton noted. By focusing on reclaiming time from mundane tasks, AdventHealth aimed to encourage consistent and safe AI usage across its teams.
To support this, peer-based groups were created within departments, allowing teams to share workflows and best practices tailored to their specific needs. This method proved more effective than large, centralized training programs, enabling staff to engage with AI in a relatable context.
Collaboration with OpenAI
As AdventHealth moved from experimental phases to broader enterprise deployment, selecting a technology partner became key. The organization chose OpenAI for its ability to meet healthcare's strict privacy and governance requirements. “We chose OpenAI because we weren’t looking for a demo. We were looking for enterprise infrastructure,” Purinton explained. The structured outputs and governance controls provided by OpenAI assured AdventHealth of responsible AI scalability across its operations.
AdventHealth's use of ChatGPT Enterprise and ChatGPT for Healthcare has already led to measurable improvements in clinical workflows. For instance, physician advisors can now generate structured summaries of patient charts more efficiently, reducing the time spent on information assembly while maintaining clinical quality.
Measuring Impact and Future Directions
The organization has established a framework for assessing AI's impact through two primary lenses: adoption and workflow performance. Daily usage metrics are monitored for accountability, while workflow performance is evaluated through turnaround times and throughput metrics. These assessments have shown significant reductions in time spent on repetitive tasks, faster turnaround on internal workflows, and fewer revisions due to improved consistency in outputs.
AdventHealth's commitment to AI aligns with its mission of whole-person care, which requires clinicians to have the time to engage meaningfully with patients. Success stories from AI integration, such as a physician being able to leave work on time instead of completing documentation late into the night, highlight the positive impact on work-life balance.
Looking ahead, AdventHealth plans to expand AI applications beyond administrative tasks, focusing on areas like patient access and clinical decision support while upholding rigorous governance and trust standards. Leadership at AdventHealth recognizes that the successful scaling of AI depends not only on the technology itself but also on how it is introduced and adopted throughout the organization. “Adoption is not ‘go use the product.’ It’s ‘change leadership,’” Purinton concluded, emphasizing the need for trust and measurable outcomes to transition from pilot programs to widespread implementation.
