Luma Health AI review (Oct 2026): patient engagement, waitlist and recall automation, and EHR-connected messaging for practices

What this Luma Health page covers

This page is for a practice manager, access director, scheduling lead, referral coordinator, or IT lead at a medical practice, community health center, or health system using or considering Luma. It is not medical or legal advice.

Luma calls itself “Operational AI for Healthcare.” Beyond rules-based reminders, its 15 September 2026 post describes agents that follow up on care gaps and on cancellations and no-shows until the patient rebooks, with agentic referral scheduling marked “coming soon.” Luma says outreach at one customer, SVCHS, rebooked more than four times as many canceled appointments as in the same two months of 2025 — a customer figure.

Ground rules: the practice manager approves every automated outreach template and checks that recalls name the right visit type; the scheduling lead samples rebooked visits; the referral coordinator checks fax matches; clinical questions always go to staff; and Luma’s figures are its claims, not our measurement.

Major features (with how someone uses them)

Examples at a glance: each feature has a concrete example from a real job plus a short Try this you can run this week.

1. Recall and care-gap outreach agents (primary)

What it is: Luma sends recall messages when a patient is due — for example a year after a Medicare Annual Wellness Visit or annual eye exam — on each patient’s own cadence, with rules such as age ranges, by SMS/RCS or an outbound AI call. Luma agents then follow up and help patients book; if a patient says they do not need or want the visit, Luma can opt them out. Luma says at one customer, SJ Health, 28% of contacted parents booked a pediatric checkup (Sep 2026 post) — a customer figure.

A recall that names the wrong visit type books the wrong appointment length, or tells a patient they are due for something they are not.

Concrete example (practice manager at a family medicine practice) A practice manager reviews one week of automated recalls — 140 messages. 131 name the right visit. Five offer an annual wellness visit to patients under 65 whose plans call it a preventive visit. Three tell patients they are due for a diabetic eye exam that an outside clinic already did. One goes to a patient who left the practice. She fixes the visit-type mapping, adds outside results to the check, and expands to the next cohort a week later.

Try this

Primary workflow:

Review one week of Luma automated recalls. Have the practice manager confirm every message names the right visit type before expanding cohorts.

2. Waitlist offers and no-show rebooking

What it is: Luma’s scheduling suite sends waitlist offers that “respect the original appointment,” and agents follow up after a no-show or late cancel until the patient rebooks, offering “new times with the same provider and appointment type.” Luma reports a customer whose waitlist offers got rheumatology patients in 33 days sooner on average, and another (CLS Health) where agentic outreach rescheduled 2.3 times as many cancellations as SMS reminders — customer figures.

Rebooking fast is good until the rebooked visit is the wrong length or provider. Sample what the agent booked.

Concrete example (scheduling lead at a multi-site dermatology group) The scheduling lead checks twenty appointments the agent rebooked after no-shows. Seventeen match the original provider and visit type. Two book a procedure visit into a short follow-up opening. One rebooks with a provider who has since left the group. She tightens the visit-length rule and removes the departed provider from the rebook list.

Try this

Sample rebooks:

Have the scheduling lead check twenty agent-rebooked appointments for provider, visit type, and length before turning rebooking on for every site.

3. Navigator AI agent for calls, chat, and text

What it is: Navigator is Luma’s AI agent for voice, chat, and text. It helps patients schedule, cancel, reschedule, and ask questions such as “what do you have for weekday mornings?”, handles more than one need in a call, works in English and Spanish, and draws on the EHR and past interactions. Luma says a Banner Health chat use got imaging patients answers 6 hours sooner — a customer figure — and that clinical questions and requests for a person are escalated (care gaps page).

A voice agent that books is acting for your front desk. Listen to the calls it finished, not just the ones it escalated.

Concrete example (practice manager at a pediatric practice) The practice manager listens to twenty-five Navigator calls. Twenty-one are handled well. Two book a sick visit for a well-child request. One gives a parent the wrong office’s hours. One correctly escalates a fever question to the nurse line. She fixes the visit mapping and office hours and keeps a weekly listen.

Try this

Listen weekly:

Have the practice manager listen to twenty-five Navigator calls each week, including handled ones, and fix every wrong booking before widening hours.

4. Fax Transform for referrals and documents

What it is: Luma says it parses referral faxes, extracts and classifies referral data, matches the referral to the correct patient in the EHR, and kicks off the referral workflow, and that its fax parsing has a 98.7% accuracy rate (Our AI page) — Luma’s figure. Fax Transform also documents in the EHR when a care gap was completed elsewhere (care gaps page).

A referral filed to the wrong patient is a privacy problem and a missed referral at once. Check matches before trusting them.

Concrete example (referral coordinator at a neurology practice) A referral coordinator checks thirty faxes Fax Transform processed. Twenty-seven are matched and classified correctly. Two are matched to a parent instead of a child with the same name. One cover sheet is read as a referral. She keeps manual review for same-name matches and reports the cover-sheet pattern to Luma.

Try this

Check matches:

Have the referral coordinator check thirty Fax Transform matches against the fax before referrals route without review.

5. EHR-connected messaging, broadcasts, and languages

What it is: Luma’s communication tools send messages by RCS/SMS, phone, or email from your number in 35+ languages, with broadcasts customized per patient and a conversation hub staff can monitor. Messages can start from a chart update, a provider’s day off, or a patient action, and Luma says patients can schedule, confirm, complete intake, or pay in the same thread. Luma also runs eligibility checks during scheduling (scheduling page).

A broadcast reaches everyone at once. One wrong field in a template becomes hundreds of wrong messages.

Concrete example (practice manager at an OB/GYN practice) Before a provider’s day off, the practice manager previews a rescheduling broadcast in English and Spanish for six patients. Five read well. One Spanish message shows the provider’s name in the wrong field. She fixes the template and keeps previewing each new broadcast with a few real patients before sending.

Try this

Preview first:

Have the practice manager preview every new broadcast template in each language with five real patient records before it goes out.

6. Agent limits, escalation rules, and the MCP Server

What it is: Luma says your team defines what each agent can act on — “which workflows it handles, which actions it can take in your EHR, what hours it runs” — with escalation rules at the platform and agent level, and that agents “don’t give clinical advice, interpret symptoms, or change medications” (AI trust and governance page). Luma says each release runs 500+ test cases, agents escalate 3% of engagements on average, and it uses models from OpenAI, Anthropic, and ElevenLabs. Its 15 September 2026 post adds an MCP Server in Early Access so assistants such as Claude, Microsoft Copilot, or ChatGPT can send outreach lists to Luma.

An agent that can book, cancel, and message patients needs written limits. Connecting outside assistants widens what can start a workflow.

Concrete example (IT lead at a community health center) Before widening Navigator, the IT lead writes down each agent’s allowed actions. Scheduling may book and reschedule but not cancel procedures. Outreach may text but not call after 7 pm. She finds the recall agent was allowed to cancel duplicate visits, which no one had approved, and removes it. She keeps the MCP Server off until the compliance officer reviews which assistants may send lists.

Try this

Write the limits:

Have the IT lead write each Luma agent’s allowed EHR actions, hours, and escalation rules, and get the practice manager’s sign-off before widening any agent.

Pricing context (October 2026)

Luma publishes no price list. When we checked on 19 Oct 2026, lumahealth.io/pricing returned a not-found page and its pages ask you to book a demo. Treat it as quote-only.

Ask in writing whether Navigator, outbound AI calls, Fax Transform, and the MCP Server are priced separately from the core platform; whether pricing is per provider, per location, or per message; and what implementation costs.

Uncertainty: We did not test Luma. Rebooking, booking-rate, faster-access, escalation, and fax-accuracy figures are Luma or customer claims. Agentic referral scheduling is marked coming soon and the MCP Server is Early Access. Confirm BAA, how opt-outs and Do-Not-Contact flags work, and which models handle which tasks.

When it’s useful (honest take)

Useful when you run a practice, community health center, or health system with high no-show, recall, and referral-fax volume, your EHR is one Luma integrates with, and the practice manager will approve every outreach template.

Less useful when your main job is front-desk intake and copays (Phreesia AI), patient billing conversations (Cedar AI), population-health care-gap programs on a shared data platform (Innovaccer AI), or prior auth and denials (Notable AI).

Related options (educational)

FAQ

Does Luma Health use AI or just rules-based reminders?

Both. Luma still runs rules-based reminders and recalls, and adds AI: Navigator handles calls, chats, and texts; agents follow up on care gaps, cancellations, and no-shows until the patient books; and Fax Transform reads referral and document faxes.

What is Luma Navigator?

Navigator is Luma’s AI agent for voice, chat, and text. It helps patients schedule, cancel, reschedule, and ask routine questions in English and Spanish, and escalates clinical questions and requests for a person to staff.

How much does Luma Health cost in October 2026?

Luma publishes no prices; its pricing URL returned a not-found page when we checked on 19 Oct 2026. Ask for a written quote that lists Navigator, outbound AI calls, and Fax Transform separately.

Can Luma agents give clinical advice?

Luma says its agents do not give clinical advice, interpret symptoms, or change medications, and that clinical questions are escalated to staff. Check this in your own call samples.

Who should check Luma AI output?

The practice manager for every outreach template, recall, and broadcast; the scheduling lead for rebooked appointments; the referral coordinator for fax matches; and the IT lead for each agent’s allowed actions and escalation rules.

Optional next step (if seats overlap)

If you pay for Luma, a separate intake vendor, and an outsourced call center that all contact the same patients, that overlap is worth mapping. Map the overlap with the $19 AI toolkit audit or the $9 AI toolkit spreadsheet. Email Audit $19 or Toolkit spreadsheet $9.

Checkout is live via Stripe. After you pay, we email delivery (spreadsheet) or intake next steps (audit).

Related: athenahealth AI · Epic Systems EHR AI · Waystar AI · Oracle Health AI · Abridge AI · Microsoft Dragon Copilot AI · Suki AI · athenahealth vs Epic Systems EHR · eClinicalWorks AI · Akasa AI · Phreesia AI · Oracle Health vs athenahealth · NextGen Healthcare AI · Elation Health AI · Innovaccer AI · Notable AI · Cedar AI · Applied Epic AI (insurance AMS — different product) · Microsoft 365 Copilot · ChatGPT for finance · Claude · Three checks before another AI seat · How-to: Run one Luma Health recall and care-gap outreach campaign, and set agent controls before it texts patients · ops hub