AdventHealth and OpenAI: Revolutionizing Healthcare with ChatGPT

AdventHealth and the Impact of ChatGPT on Healthcare

AdventHealth, a leading hospital network, has recently integrated ChatGPT for Health into its daily operations. This initiative aims to alleviate the administrative burden on medical staff by automating repetitive tasks such as documentation and support. With this automation, care teams can now reclaim several hours each week, allowing them to focus more on patient care. This approach promises to enhance not only operational efficiency but also expand clinical capacity, accelerate access to care, and improve the patient experience.

The Challenges of a Pressured System

AdventHealth, present in nine states and serving millions of patients each year, faces challenges similar to those encountered by many large healthcare systems. These challenges include tight financial margins, increasing demand for services, and heightened administrative complexity. Daily workflows are particularly affected. For example, medical advisors spend about 10 minutes per case reading records, identifying relevant details, verifying criteria, and drafting structured justifications. This time, multiplied by hundreds or thousands of cases, quickly adds up, creating a considerable burden.

This pressure is not limited to clinical roles. Finance, human resources, information technology, and other teams also spend significant time drafting documents, summarizing information, and preparing necessary but time-consuming materials. As a result, these teams often operate in a "constant operations mode," with limited capacity for higher-value work. Meanwhile, interest in AI is already emerging within the organization, although its use is still constrained by formal policies.

"We had people who were eager to get started, but there were a great number of people who were holding back," explains Rob Purinton, Chief AI Officer at AdventHealth. "They weren't sure how to effectively use AI in their daily work."

A Strategic Adoption of AI

AdventHealth's leadership quickly realized that to achieve meaningful change, it was not enough to conduct isolated pilot projects. The real challenge was ensuring a consistent and secure use of AI across the entire workforce. "The hardest part of AI in healthcare is making sure humans use it safely, consistently, and at scale," says Purinton. "We made the decision from the outset to consider adoption as the product."

This decision influenced how AI was deployed. Instead of presenting AI as mere automation, leaders framed it as a means to reduce administrative burden and give time back to clinicians and staff. "We're not talking about AI as automation. We're talking about time reclaimed," explains Purinton. "If we can take a 10-minute review and significantly compress it—while maintaining quality—that's capacity we can give back to our clinicians."

AdventHealth also treated adoption as a measurable operational indicator. The organization tracks the number of messages per user per business day, excluding weekends and holidays to create a consistent baseline. This metric is monitored and managed like any other KPI, with goals and trends reviewed regularly.

A Large-Scale Deployment with OpenAI

As the organization transitioned from experimentation to enterprise-wide deployment, leadership prioritized tools capable of meeting the healthcare sector's requirements for privacy, governance, and reliability. "We chose OpenAI because we weren't looking for a demo. We were looking for enterprise infrastructure," says Purinton. "The reasoning capability, structured outputs, and governance controls gave us confidence that this was not just productivity software. It was something we could deploy responsibly in a healthcare system."

AdventHealth adopted ChatGPT Enterprise and then ChatGPT for Health, which provided additional assurances for regulated environments, including data protections and compliance support. The speed of innovation and collaboration also influenced the decision. "We really appreciate being closer to what's possible," says Purinton. "And we found OpenAI very collaborative as we think about pilots, deployments, and what's next."

Restructuring Workflows

One of the earliest measurable use cases was utilization management. By using ChatGPT for Health, medical advisors can generate structured summaries of patient records, highlight relevant clinical details, and draft initial justifications. The clinician remains responsible for the final judgment, but the time spent assembling information is reduced.

The organization measures impact using system-level data, including timestamps in electronic health records, rather than self-reported estimates. "We prefer metrics that are integrated directly into the process," explains Purinton. "We can see exactly how many minutes have improved and whether that change is statistically significant."

Beyond clinical workflows, similar patterns have emerged in departments:

• Document and plan writing begins with a first draft rather than a blank page.
• Policies and communications are converted into structured, usable formats.
• Unstructured notes and information are quickly summarized into action steps.

These changes reduce cycle times, limit revisions, and improve consistency of outcomes.

Measuring the Impact of AI

AdventHealth evaluates the impact of AI along two main dimensions: adoption and workflow performance. On the adoption side, tracking daily usage has created accountability and visibility into how quickly AI is becoming an integral part of routine work. On the workflow side, pilots are evaluated using throughput metrics such as time per task, response time, and volume processed. In utilization management, the goal is to reduce review time while maintaining quality and consistency.

Across departments, teams report:

• A reduction in time spent on repetitive documentation and review tasks.
• Faster response times on internal workflows.
• Fewer rework cycles due to more consistent first drafts.
• Increased capacity without additional staff.

The organization often describes these gains as "time reclaimed," but leadership directly ties this concept to measurable outcomes. "If you take a 10-minute task and reduce it to two, and that happens a thousand times a week, that's real capacity," says Purinton. "The question is how you reinvest that capacity."

Freeing Up Time for Clinicians

For AdventHealth, the value of AI is closely linked to its mission of providing holistic care. This requires time—time for clinicians to spend with patients and families, and time for staff to focus on higher-value work. One example illustrates the impact at an individual level. A physician who previously spent evenings completing documentation, often referred to as "pajama time," was able to finish their work during regular hours after workflow changes supported by AI. "He left work at work," explains Purinton. "He could go home and be present with his family." Stories like this reinforce the organization's approach that AI is a tool to reduce administrative burden rather than replace roles.

Towards a Broader Impact

To date, most measurable gains have come from reducing time spent on existing tasks. AdventHealth views this as a starting point. The organization is now focusing on expanding into areas such as patient access, clinical decision support, and new care delivery models, while maintaining the same emphasis on governance, measurement, and trust. The central lesson, according to leadership, is that the scale of AI depends less on the technology itself and more on how it is introduced and adopted. "Adoption is not 'go use the product.' It's 'change leadership,'" says Purinton. "When you measure it, prove its value, and lead with confidence, that's when you go beyond pilots."