Automation

OpenAI Deploys Computer-Use Agents in Healthcare with CVS, Cigna

AI systems now navigate EHRs and payer portals directly, automating referrals and prior authorizations while keeping clinicians in control.

Omega Editorial · November 16, 2026 · 3 min read

Key takeaways

  • OpenAI's computer-use agents navigate EHRs and payer portals by viewing screens and clicking buttons like humans, moving beyond AI copilots that only draft content.
  • GPT-6 Astra achieved 99% clinical safety ratings across 325 evaluated tasks and was rated very or extremely helpful by physicians in 93% of assessments.
  • CVS Health used an agent to clear a months-long document backlog over one weekend and now processes tens of thousands of daily consults with the technology.
  • Cigna Healthcare is aggregating member information from more than 10 systems to reduce case manager search time, while Abridge is building agents for inpatient care coordination.
  • The technology addresses healthcare's fragmented software environment where APIs don't exist or fail to connect complete workflows between systems.

OpenAI brings autonomous agents to healthcare operations

OpenAI is deploying computer-use agents—AI systems that interact with software by viewing screens, clicking buttons, and typing text—to automate fragmented clinical and administrative workflows in healthcare. The company announced partnerships with CVS Health, Cigna Healthcare, Abridge, and Ambience to apply the technology to referrals, prior authorizations, discharge planning, and care coordination.

Karan Singhal, head of Health AI and Artificial General Intelligence at OpenAI, told Fierce Healthcare that research teams began exploring computer-use agents for healthcare about six months ago, betting the technology would become reliable enough to assist with workflows while keeping humans in control.

Computer-use agents differ from existing AI copilots by actually navigating electronic health record systems, payer portals, and other applications on behalf of users, rather than simply drafting or summarizing information. The approach addresses healthcare's fragmented software environment, particularly where APIs and integrations don't exist or fail to connect entire workflows.

Why it matters

Healthcare organizations spend enormous resources on manual data entry between disconnected systems—work that delays patient care and burns out staff. Computer-use agents represent a fundamental shift from AI that assists to AI that executes, potentially eliminating bottlenecks in authorization, referrals, and care coordination without requiring expensive custom integrations across legacy systems.

Early results show high safety ratings, real-world impact

OpenAI evaluated its GPT-6 Astra model on multi-step EHR tasks shaped by physician input, including reviewing patient records, drafting messages and notes, preparing medication refills, and proposing orders. In the evaluation, 99% of 325 reviewed tasks were rated clinically safe, while physicians rated the model "very helpful" or "extremely helpful" in 93% of 646 assessments.

On HealthAdminBench, an independent benchmark from Stanford measuring performance on administrative workflows, GPT-6 Astra earned 94.3% of available rubric points and passed every required check in 60.7% of task attempts.

Singhal acknowledged the models aren't perfect. The 1% of tasks rated not clinically safe involved cases where the model might overreach and take action without proper authorization, though researchers identified no specific, plausible risks of patient harm.

Organizations deploy agents across production workflows

CVS Health is using the technology to automate repetitive work across clinical systems. In one production workflow, an agent cleared a months-long document-sending backlog over a weekend. Another agent handles system steps for tens of thousands of consults daily.

Cigna Healthcare is using OpenAI's models to pull together member information from more than 10 clinical and administrative systems, aiming to reduce time case managers spend searching across systems.

Abridge is building computer-use agents to help clinicians execute care plans across applications, including those without APIs. The company plans an inpatient use case that adds EHR information and clinical context to coordinate next steps, with agents working concurrently on medication access and home health coordination.

OpenAI designed the agents with configurable limits on system access and actions, plus audit logs that provide clear records of agent actions and approvals. Care teams, developers, and individual clinicians can access computer use through ChatGPT for Healthcare, ChatGPT for Clinicians, and the OpenAI API.

The details were first reported by Fierce Healthcare, with Singhal sharing evaluation results in an exclusive interview at the HLTH 2026 conference in Las Vegas.

  • #healthcare ai agents
  • #openai healthcare
  • #prior authorization automation
  • #ehr automation
  • #agentic ai
  • #clinical workflows

This is an original analysis by the Omega editorial team. Source reporting: AI Watch.

Want systems like this working for your business?

Book a Call

More in Automation