ORCHESTRATING BETTER OUTCOMES
“Luma’s agents make sure that patients don’t fall through the cracks… our providers noticed a significant jump in their scheduled appointments, and patients are getting care sooner.”
Ibrahim Arabie
Technology Implementation Specialist
CLS Health
4.4M
care-related next steps handled annually.
5M+
AI engagements annually.
1.5M hours
saved on inbound calls with AI in 12 months.
AI ORCHESTRATION METHODS
Orchestration to fit each need
Luma’s AI is powered by our full platform, 11+ years of healthcare experience, and industry-leading EHR integration. It builds on your workflows and patient context to make the right outcome happen, whether in a deterministic workflow or a long-running agentic job.
Deterministic AI workflows
Define each step with Luma’s Workflow Builder and Agent Library, where you have control over when, where, what, and how agents are used.
Example: Ask a scheduling agent to call a patient after a no-show to reschedule.
Agentic jobs
Choose an outcome, and let Luma handle the rest. Specialized agents handle individual tasks to make sure the outcome is accomplished.
Example: “get patients ready for hernia surgery” → scheduling, intake, and outreach agents work together.
Headless AI workflows (MCP)
Connect your AI tools like Claude, Microsoft Copilot, or ChatGPT to the Luma platform via MCP server to power your own “headless” workflows.
Example: Ask Claude to find patients due for eye exams from a population health tool, then reach out with Luma agents.
We try to remove as many barriers as possible to closing care gaps with a great consumer experience. Luma’s agents helped us guide parents and caregivers to schedule pediatric checkups without needing to call.”
Mary-Lou Milabu
Director of Quality, SJ Health
The Luma difference
With a foundation of 11+ years of patient interactions and workflow-level integration depth, we understand what “done” looks like. Agents coordinate across every touchpoint, with real-time EHR integration, to make sure the desired outcome happens.
Agentic orchestration in action
Luma’s agents guide the patient through complex, multi-step workflows. Right-on-time context, tasks, resources, and answers help patients get to their care informed and ready.
In this example, a patient needs to prepare for a colonoscopy. Below, you can explore how different agents make sure that every step is done.
GI care agent
The GI care agent’s job is to ensure patients due for a colonoscopy receive one.
It reaches out to encourage Cathy, age 45 and due for her first colonoscopy, to book her appointment and doesn’t consider its job done until Cathy’s procedure is completed.
GLP-1 agent
The GLP-1 agent’s job is to ensure that Cathy receives extra prep related to her GLP-1. It guides her to start prep earlier and to discontinue her GLP-1, as recommended by her provider, before the procedure.
If Cathy has clinical questions about her GLP-1, the agent escalates to her provider.
Patient prep agent
The patient prep agent’s job is to guide Cathy’s prep until her procedure. Its job is done when Cathy gets her colonoscopy.
The agent reminds Cathy to fast, complete bowel prep, and complete forms. Cathy can ask questions like “do I need to fast?” or “where should I park?” and the agent will provide answers or approved education, escalating to staff as needed.
Concierge agent
The concierge agent’s job is to triage and hand off any unrelated requests or needs Cathy asks about. Cathy can ask about anything related to her care and get an answer, either from a specialized agent or staff.
For example, Cathy asks if she can book her annual wellness visit, and the concierge agent hands off to a scheduling agent.
AI trust and governance
Luma agents work inside limits your team sets. You define what each agent can act on, and when an agent reaches one of those limits, it stops and hands the work to a person. Every engagement leaves a record you can read, with the full message history and every action the agent took.