OpenAI Builds Computer-Use Agents With CVS, Cigna
OpenAI's new AI agents work across existing healthcare software to handle referrals, prior authorizations and discharge planning.
Updated on Nov. 16, 2026 in Healthcare

Live Poll
Do you think AI handling healthcare paperwork will make your care better or worse?
OpenAI built computer-use AI agents with CVS Health, Cigna Healthcare, Abridge and Ambience Healthcare that read screens and operate across EHRs and payer portals. The company also added more than 20 healthcare plugins to ChatGPT for Healthcare.
Why it matters
The agents can run tasks inside systems health organizations already use, without requiring new software builds. Organizations control which applications the agents reach and which steps require human review, with every action recorded in audit logs.
Cigna reported moving from concept to deployment in under eight weeks, while OpenAI's GPT-6 Luna model costs 8.5 cents per task, 98% less than Astra. Performance figures come from company-reported assessments, and independent validation was not included in the source.
The players
OpenAI
OpenAI is an artificial intelligence research and deployment company known for developing the ChatGPT family of models and products.
CVS Health
CVS Health is a major American healthcare company that operates pharmacies, a pharmacy benefit manager and health insurance services.
Cigna Healthcare
Cigna Healthcare is the health insurance and care-management arm of the Cigna Group, one of the largest health service companies in the United States.
The details
The agents read screens, click, type and move between applications such as EHRs and payer portals, handling referrals, prior authorizations and discharge planning. Health systems can access them through ChatGPT for Healthcare, the free ChatGPT for Clinicians and the OpenAI API.
Timeline
Nov. 16, 2026: OpenAI announced the computer-use healthcare AI agents.
Industry Dynamics
The rollout extends OpenAI's Agents SDK into production healthcare, where CVS has open-sourced its Agent Kit built on that SDK. The move places computer-use agents inside legacy payer and clinical systems rather than replacing them.
Patients and clinicians may see faster referral handling and prior authorization turnaround as these agents run across existing systems. Health systems retain control over which applications the agents can reach and which steps need human sign-off.
The takeaway
Health organizations evaluating AI agents should weigh the audit-log and human-review controls alongside any productivity gains. Independent validation of the reported benchmarks would strengthen confidence in these results.
Further reading
For related coverage, explore Healthcare on the business desk.
Live Poll
Do you think AI handling healthcare paperwork will make your care better or worse?



