Industry

OpenAI Expands ChatGPT for Healthcare With Computer-Use Agents and Clinical Search

OpenAI advances ChatGPT for Healthcare with autonomous agents that navigate healthcare software, improved clinical evidence search, and 20+ new plugins for clinical and administrative workflows.

Last verified:

Computer-Use Agents Automate Healthcare Workflow Coordination

Healthcare organizations operate across fragmented software ecosystems — electronic health records (EHRs), hospital scheduling systems, payer portals, and clinical documentation platforms rarely communicate directly. According to OpenAI, the company released computer-use agents for ChatGPT for Healthcare that autonomously navigate these disconnected systems, reducing manual data entry and administrative overhead. The agents read screens, click buttons, type text, and move between applications to complete multi-step workflows such as processing insurance referrals, transferring patient documents between systems, and drafting clinical orders for physician review.

Early adopters Abridge, Ambience, Cigna Healthcare, and CVS Health are deploying the agents for referral processing and care coordination, using voice-based interfaces to trigger automated sequences. In one demonstration, a clinician uses voice commands within the ChatGPT app to request follow-up on an MRI order; the agent reviews the patient chart, drafts the order, and prepares a patient-facing reply for the clinician to approve before sending.

Security, Compliance, and Human-in-the-Loop Controls

OpenAI integrated HIPAA compliance mechanisms into the agent framework. According to the company’s announcement, organizations must use eligible services covered by a Business Associate Agreement (BAA) with OpenAI, configure required security and retention settings, and ensure protections extend to all connected systems. Critically, content shared with ChatGPT for Healthcare is not used for model training.

The platform includes access controls and audit logs for secure deployments. Healthcare teams define which applications agents can access and specify which workflow steps require human review before execution — a safeguard addressing clinician concerns about autonomous action in high-stakes environments.

Enhanced Clinical Evidence Search Across Care Teams

OpenAI strengthened the clinical search capabilities embedded in both ChatGPT for Healthcare and ChatGPT for Clinicians. According to the company, the updates improve how the system finds, evaluates, and cites medical evidence in response to clinical queries. The enhancement allows clinicians to connect research findings to their specific cases and translate evidence into practical next steps, with transparent citations that clinicians can inspect and verify.

The company also introduced more than 20 new plugins to the healthcare plugin ecosystem, expanding integration options with existing clinical workflows and third-party health data platforms.

Why This Matters

The gap between clinical evidence and clinical practice has historically been bridged through manual literature review and institutional knowledge — time-intensive processes that delay care decisions. By automating both administrative task-switching and evidence retrieval, OpenAI’s updates target two distinct friction points in healthcare delivery: the operational (data moving between systems) and the epistemic (connecting research to cases). The computer-use agents address the former; improved clinical search addresses the latter.

For healthcare administrators, the agents’ ability to reduce manual referral processing and document coordination could lower administrative burden on clinical staff — a category of work that rarely involves direct patient contact. For clinicians, faster access to cited medical evidence may accelerate evidence-based decision-making, though the clinical impact will depend on adoption rates and integration with existing workflow preferences. The HIPAA-compliance infrastructure and human-in-the-loop controls are prerequisites for hospital deployment; whether these mechanisms scale to complex multi-system environments remains to be validated in practice.

Frequently Asked Questions

What can ChatGPT's computer-use agents do in healthcare settings?

According to OpenAI, the agents navigate existing healthcare software by reading screens, clicking, typing, and moving between applications to automate tasks like processing referrals, moving documents between systems, and drafting clinical orders for clinician review.

Is ChatGPT for Healthcare HIPAA compliant?

Yes, but with requirements: organizations must use eligible services covered by a Business Associate Agreement (BAA) with OpenAI, apply required security and retention settings, and extend protections to connected systems. Content shared is not used to train models.

Who is already using these agents?

According to OpenAI, early healthcare adopters deploying the agents include Abridge, Ambience, Cigna Healthcare, and CVS Health, primarily for referral processing and care coordination workflows.

What improvements were made to clinical search?

OpenAI improved how ChatGPT finds, evaluates, and cites medical evidence in both ChatGPT for Healthcare and ChatGPT for Clinicians, with updates to help clinicians connect research to clinical questions and translate findings into actionable next steps.

#healthcare #AI agents #chatgpt #clinical-decision-support #automation #hipaa