Luma Health has expanded its Navigator AI agent to include outbound voice and SMS workflows that rebook missed appointments, schedule referrals, and close care gaps directly within a health system's EHR. The move turns cancellations and no-shows from notification events into resolved outcomes, aiming to recover revenue and reduce manual scheduling work for staff.
How the agentic workflows function
When a patient cancels or misses an appointment, Navigator contacts them with available times matching the original visit type, provider, and location. The platform also reaches patients due for preventive services - such as annual wellness visits or diabetic eye exams - and those with active referrals who have not yet booked. Health systems configure the outreach criteria, including which providers, facilities, appointment types, and patient attributes trigger the workflow.
"The shift to agents is a shift from notifying to completing," said Michael Chou, senior vice president of product at Luma Health. "The agent works a cancelled slot until it's rebooked or genuinely exhausted and comes back with a resolved outcome, rather than a delivery receipt."
The initial scope is limited to scheduling, rescheduling, and nonclinical appointment questions. When Navigator encounters unresolved scheduling matters or clinical questions, it routes them to staff for review. The system inherits the same consent, access, and audit controls the health system already applies to patient communications, and it follows communication preferences recorded in the EHR - including opt-outs made during an interaction.
Connecting Navigator to broader AI workflows
The release includes a Model Context Protocol server that lets organizations connect Navigator to AI tools such as Anthropic's Claude, Microsoft's Copilot, or OpenAI's ChatGPT. A health system could, for example, pull an outreach list from its population health management system and send it to Navigator for execution.
Access is configured according to least-privilege and minimum-necessary principles. Tool calls and agent actions are logged, and patient information is not shared among customers or used to train foundation models. Luma plans to work with participating health systems to measure patient satisfaction with AI-supported scheduling and rescheduling.
The progression from messaging to action
Luma introduced its Care Pathway Messaging chatbot in 2018 and has since raised $160 million in total funding, including a $130 million Series C in 2021. The new agentic workflows extend the platform's capability beyond delivering messages to acting on cancellations, referrals, and care gaps. In an earlier interview with Healthcare IT News, Luma cofounder and president Aditya Bansod described the philosophy behind the approach. "I don't believe that people have to be the coordination layer across healthcare operations," Bansod said. "The purpose of healthcare is not having people coordinate healthcare. The purpose of healthcare is providing care to patients."
For health systems deploying multiple AI tools, the shift toward AI Agents & Automation that complete tasks rather than simply flag them represents a practical step toward reducing administrative bottlenecks. Luma's approach keeps humans in the loop for clinical decisions while letting the agent handle the repetitive work of filling cancelled slots and chasing overdue referrals.
Why this matters for healthcare professionals
No-show rates in U.S. health systems typically range from 5% to 30% depending on specialty and patient population. An agent that works a cancelled slot until it is rebooked - rather than sending a single text and stopping - directly addresses the revenue and access problem that empty appointment slots create. For scheduling teams, the reduction in manual outreach means staff can focus on complex cases that require clinical judgment, while the AI handles the high-volume, rules-based follow-up that currently consumes hours of phone time each day.
Your membership also unlocks:
