AI agents are starting to fill out medical forms and file insurance requests in place of human staff

The Decoder
Read full story· 4 min
OpenAI has introduced GPT-6 Astra and GPT-6 Luna agents that automate administrative healthcare tasks by interacting with existing software interfaces, reducing costs and improving efficiency. These agents draft medical orders and insurance requests, with clinical staff reviewing outputs before finalization. Major US healthcare companies like CVS Health are already deploying these agents to manage high volumes of referrals and consultations.

What happened

OpenAI has launched two AI agents, GPT-6 Astra and GPT-6 Luna, designed to automate administrative tasks in healthcare settings. These agents interact directly with existing clinic and insurance software to draft medical orders and insurance requests. Clinical and administrative staff review the agents’ outputs before they are finalized. Major healthcare companies in the US, including CVS Health, have started using these agents to handle large volumes of referrals and consultations.

Key details

  • GPT-6 Luna processes administrative tasks for about 8.5 cents per task, which is 98% less than Astra.
  • GPT-6 Astra scored 94.3% and GPT-6 Luna scored 91.8% on the HealthAdminBench benchmark from Stanford University.
  • Physician reviewers rated 99% of 325 GPT-6 Astra tasks as clinically safe.
  • CVS Health uses these agents to process referrals and consultations, clearing backlogs and handling tens of thousands of cases daily.
  • For HIPAA-compliant workflows, organizations must sign a Business Associate Agreement with OpenAI and configure security settings.

Why it matters

Healthcare providers and insurers in the US can now automate routine administrative work at a much lower cost, potentially reducing the need for human clerks for these tasks. This change allows clinics and insurers to process higher volumes of paperwork more quickly, but they must also ensure data privacy and compliance with regulations like HIPAA. Staff must review AI-generated outputs, and organizations are responsible for configuring and monitoring the agents to maintain safety and privacy standards.

More on this story


More in Healthcare