Luma Health launches 59 clinic workflow actions in ChatGPT
Luma Health’s November 16, 2026 ChatGPT plugin offers 59 EHR-connected workflow actions, from appointment outreach to payments, with broad write access and controls to govern it.
By Grace Kim
4 min read

What's News
- Luma Health announced 59 ChatGPT clinic workflow actions on November 16, 2026.
- The plugin’s 59 actions read from and write back to electronic health records under existing health-system rules and permissions.
- OpenAI included Luma among more than 20 new ChatGPT for Healthcare plugins on November 16, 2026.
- Balances of $100 or less receive a payment task; larger balances receive a financial-assistance task.
- Responses expose only birth year, omit insurance member and medical record numbers, and mask phone and email details.
Luma Health launched a ChatGPT plugin with 59 clinic workflow actions on November 16, 2026, allowing authorized staff to start administrative processes inside the chatbot rather than move through successive electronic health record screens.
In its announcement, Luma said the plugin moves ChatGPT beyond answering questions and into work that can send patient messages, create referrals, prepare procedures, assign payment tasks and update records.
What can the 59 workflow actions do?
Luma says work that previously required a report request, a spreadsheet and hours of phone calls can start with one prompt.
One end-of-day workflow identifies patients who have not confirmed appointments for the following day and those with incomplete checklists. It then messages them to arrive 10 minutes early and finish paperwork.
Another workflow finds patients with unpaid charges and upcoming appointments, adding a payment task for balances of $100 or less and a financial-assistance task for larger amounts.
The 59 actions cover:
- Scheduling and waitlists
- Referral creation and updates
- Patient messaging in preferred languages
- Appointment checklists
- Procedure-preparation enrollment and progress checks
- Payments and staff tasks
If a provider moves clinics, a leader can request a list of that provider’s scheduled patients and waitlist entries, then offer sooner appointments with another provider for the same appointment types.
Luma’s immediate claim is simple: ChatGPT can now initiate clinic work, not merely answer questions.
How does the plugin connect to health records?
Luma’s connector reads from and writes back to the electronic health record (EHR), which remains the system of record. Every action follows the health system’s existing rules and permissions.
David Beam, Luma’s VP of Technical Delivery, used the term “‘headless’ AI workflows” for that design.
Leaders can retrieve data and act on it in ChatGPT without moving through successive EHR screens, according to Luma.
Under Luma’s design, the EHR stays authoritative while ChatGPT serves as the interface for permitted changes.
What did OpenAI announce for healthcare?
OpenAI’s November 16, 2026 healthcare announcement included Luma among more than 20 new plugins available in ChatGPT for Healthcare. Access is subject to workspace settings and administrator enablement.
OpenAI’s care-coordinator example paired Luma’s list of unconfirmed appointments for the next day with TriZetto’s insurance-eligibility checks, producing a call list of items to resolve before each visit.
OpenAI says ChatGPT for Healthcare offers customers a Business Associate Agreement (BAA) to support HIPAA-compliant workflows. The company also says content shared with the product is not used to train models.
Organizations must use eligible services covered by their BAA, apply required security and retention settings, and extend protections to connected systems.
OpenAI’s release places Luma inside a broader healthcare plugin ecosystem whose availability remains under customer control.
What access and audit controls apply?
OpenAI’s Luma Health plugin listing requires a Luma account and sign-in through the connected organization. Available actions depend on the connected account and server tool set, while appointment and record changes require explicit user intent. The integration targets authorized healthcare staff managing administrative patient-access workflows and does not provide diagnosis or treatment advice.
Luma’s Model Context Protocol (MCP) documentation labels the server early access. The hosted remote flow uses browser-based OAuth sign-in. Consent covers the signed-in account’s available tools, including write actions; it does not create a read-only role or narrower per-tool scope.
The standard ChatGPT connector may expose only the default tool set, while workflow-specific tools require additional request metadata. Luma’s two Agent Launchpad tools list configured workflows and enrollments but do not themselves enroll patients or start workflows.
Create, update and cancel operations use the authenticated account’s existing Luma API permissions. Luma logs tool calls with timestamps, resource IDs and session context. Individual calls attach to tokens rather than named human users.
Responses replace phone numbers and email addresses with availability flags, expose only birth year, and strip insurance member numbers and medical record numbers. Luma says the redaction is best effort, though responses still return patient names where operationally required.
Organizations also need appropriate PHI controls and a Business Associate Agreement with their LLM provider, according to Luma’s documentation.
Luma’s documentation makes the trade-off clear: tool calls are auditable, but attribution remains at the token level and write consent is broad.
How will health systems govern the agents?
On its AI governance page, Luma says customers define which workflows agents handle, which EHR actions they can take, their operating hours and permitted information sources. Agents stop and escalate to staff at those limits and always honor Do-Not-Contact flags.
Staff can review complete message histories, agent actions, post-call summaries and changes to agent behavior.
Luma positions the integration as a conversational front end for record changes and patient outreach. Its useful scope depends on the controls around it.
Going forward, Luma says the usable scope will depend on four controls: which accounts can act, how explicit intent is established, where staff take over and how each action is audited.
Source: unite.ai
More from Grace Kim
Show full bio
Market editor covering people, launches and day-to-day practice in entrepreneurship, startups and venture funding at Business Bearings.
2,178 articles