AdventHealth and OpenAI: Scaling AI to Reduce Clinical Administrative Burden

AdventHealth and OpenAI: Scaling AI to Reduce Clinical Administrative Burden

AdventHealth has deployed ChatGPT for Healthcare to reduce administrative burden and streamline clinical workflows across its system. By automating time-intensive documentation and support tasks, the health system is expanding clinical capacity and improving the patient experience by allowing clinicians to focus more directly on patients.

Enterprise Deployment of ChatGPT for Healthcare

AdventHealth transitioned from isolated pilots to an enterprise-scale deployment of ChatGPT Enterprise and subsequently ChatGPT for Healthcare. The organization prioritized enterprise infrastructure over simple productivity software, selecting OpenAI based on its reasoning capabilities, structured outputs, and governance controls necessary for regulated healthcare environments.

Key infrastructure choices included:

  • Compliance and Privacy: The use of ChatGPT for Healthcare provides specific data protections and compliance support required for healthcare operations.
  • Governance: The system emphasizes reliability and governance to ensure the tool can be responsibly scaled across a large health system.

Workflow Redesign and Clinical Applications

AdventHealth has integrated AI into specific workflows to reduce "cycle times" and eliminate the blank-page problem for staff. The primary focus is on reducing the time spent on repetitive documentation and review tasks.

Utilization Management

In utilization management, physician advisors use ChatGPT for Healthcare to:

  • Generate structured summaries of patient charts.
  • Surface relevant clinical details.
  • Surface relevant clinical details and draft initial rationales.

While the clinician remains the final decision-maker, the AI reduces the time spent assembling information. AdventHealth measures the impact of these changes using system-level data and timestamps in electronic health records to ensure improvements are statistically significant.

Operational and Administrative Efficiency

Beyond clinical roles, AI is used across finance, HR, and IT departments to:

  • Create first-pass drafts of documents and plans.
  • Convert policies and communications into structured formats.
  • Summarize unstructured information into actionable steps.

Adoption Strategy: "Adoption as the Product"

AdventHealth treats AI adoption as a measurable operational metric rather than a technical rollout. The organization frames the technology not as "automation" but as "time back"—the goal being to return capacity to clinicians and staff.

To drive consistent and safe usage at scale, AdventHealth employs the following strategies:

  • KPI Tracking: The organization tracks "messages per user per business day" (excluding weekends and holidays) as a primary KPI to monitor usage trends.
  • Peer-Based Learning: Instead of centralized training, the system relies on domain-based peer groups (e.g., finance teams working with finance teams) to share prompts and best practices.
  • Change Leadership: Leadership views scaling AI as a matter of change leadership rather than product deployment, focusing on proving value through measurement and trust.

Measurable Outcomes and Impact

AdventHealth evaluates AI impact through adoption metrics and workflow performance, specifically focusing on throughput metrics such as time per task, turnaround time, and volume handled.

Reported gains include:

  • Reduced Documentation Time: A decrease in repetitive documentation tasks, which has reduced "pajama time" (work completed by physicians in the evenings).
  • Increased Capacity: By compressing tasks—such as reducing a 10-minute review to two minutes—the organization increases overall clinical capacity without increasing staffing levels.
  • Operational Consistency: Fewer rework cycles due to more consistent first drafts of internal communications and policies.

Future Directions

AdventHealth is leveraging its current success in administrative reduction as a foundation to expand AI applications into patient access, clinical decision support, and new care delivery models.

Sources