OpenAI computer-use agents reportedly file prior auths for CVS, Cigna
OpenAI is positioning two upcoming models, GPT-6 Astra and GPT-6 Luna, as computer-use agents that automate US healthcare administration by operating existing clinic software and insurance portals the way a human clerk would, according to The Decoder. The report says Luna handles prior authorizations — the insurer approvals required before certain treatments or prescriptions are covered — at roughly 8.5 cents per task, and that CVS Health and Cigna Healthcare are already deploying the technology. The account gives no launch date, contract size, task volume or detail on which workflows are live, and cites no on-the-record confirmation from OpenAI, CVS or Cigna.
Verified and unconfirmed
- Verified (as reported by The Decoder): OpenAI intends GPT-6 Astra and GPT-6 Luna to automate administrative healthcare tasks via computer-use agents, Luna handles prior authorizations at about 8.5 cents each, and CVS Health and Cigna Healthcare are described as already deploying the technology.
- Not yet confirmed: the GPT-6 Astra and Luna names and capabilities, the 8.5-cent figure, and the CVS and Cigna deployments all rest on this single secondary report with no cited primary statement, and the sources leave out launch timing, deployment scope and any independent test of accuracy or error rates.
Background
Computer-use agents work through the same interfaces as human staff — clicking, typing and navigating portals — instead of through software integrations, which is what allows them to operate legacy clinic and payer systems that have no API. OpenAI announced computer-use agents, alongside improved clinical answers and new healthcare plugins, for ChatGPT for Healthcare in a November 16 post; AI Daily Radar’s November 16 digest reported that the announcement gave no release date, pricing, named customers, or list of supported software. This report adds a per-task cost and named insurer deployments, but those specifics still trace back to OpenAI and its partners rather than to independent testing.
Why it matters
Because the agents drive existing portals and clinic software rather than requiring an API, the practical bottleneck for insurers and providers shifts from software integration to granting access, supervising the runs and keeping audit trails of what each agent submitted — which is where compliance review will focus if the deployments expand.
What to watch
Watch for OpenAI, CVS Health or Cigna to confirm the deployment on the record and name which workflows are live; without that, the model names and the per-task price remain single-source claims.
Comments
Comments use a GitHub account and are moderated. Comment rules