What does Luma Health do?

Luma Health describes itself as operational AI for healthcare. Its published platform covers patient access and retention, referrals, scheduling, waitlists, registration and prep, intake and forms, eligibility and verification, payments, patient communication, reminders and recalls, feedback management, broadcast, analytics and a conversational agent, with no-code AI workflow creation. It works across multiple EHRs and had 39 athenahealth Marketplace connections as of 5 September 2026.

Is Luma Health better than Pretty Good AI?

On breadth of the patient access surface, Luma publishes more. On athenaOne integration depth and on voice, we go deeper because we only build for one EHR. Neither of those makes one product better in general. The question that decides it is whether your practice runs one EHR or several, and whether your binding constraint is the phone line or the wider access workflow.

We already use Luma and like it. Should we switch?

Probably not, and we would rather say so than pretend otherwise. Ripping out a working patient access platform is expensive in staff retraining and workflow rebuild, and the gain has to be large to justify it. The case for talking to us is narrower: your phone line is still the bottleneck, calls are going unanswered or to voicemail at volume, and you want that specific problem solved deeply on athenaOne. That is a different purchase from replacing your access platform.

Can a practice run both?

Yes, and for some groups it is the sensible arrangement, with Luma owning the asynchronous access workflows it is already configured for and a voice-first agent owning the phone line. The rule is the same as with any two-system setup: decide explicitly which system owns outbound patient contact for which workflow, or patients will hear from both.

How do the athenahealth Marketplace numbers compare?

They are close. Luma had 39 practice connections and 1 review as of 5 September 2026; Pretty Good AI has 45 connections and 22 ratings, all five-star, as of October 2026. Luma is a much larger company with a much larger overall customer base, so the marketplace figure reflects athenaOne-specific adoption rather than company size, and it is the figure an athenaOne practice should weigh.

What should we test during an evaluation?

Your hardest scheduling rules, not your average ones. Take the appointment types that require a specific provider, a specific room or a preceding visit, and make each vendor book them live in your own environment. Also test what happens when a caller asks for something the agent cannot do, because the handoff to a person is where most front-office automation actually fails.

Pretty Good AI vs Luma Health

← All comparisons

Luma Health runs patient access across many EHRs. Pretty Good AI goes deep on one. What an athenaOne practice should ask before it signs with either one.

Last reviewed September 23, 2026 7 min read

The short answer

Luma Health is an operational AI platform for patient access, covering scheduling, referrals, waitlists, reminders and recalls, registration and prep, intake and forms, eligibility, payments, feedback management and a conversational agent, with no-code AI workflow building on top. It is multi-EHR and had 39 athenahealth Marketplace practice connections as of 5 September 2026. Pretty Good AI covers a narrower slice of the same territory but only on athenaOne, with production access to 820+ athenaOne APIs and a voice-first architecture. For an athenaOne practice, ask both vendors which workflows they complete with write-back, which they read only, and which route to staff; compare that list before deciding whether to keep Luma or add another system.

Pretty Good AI vs Luma Health, side by side

Feature-by-feature comparison of Pretty Good AI and Luma Health
Dimension	Pretty Good AI	Luma Health
EHR scope	athenaOne only. Production access to 820+ athenaOne APIs, with workflows built around your practice's rules.	Multi-EHR patient access platform with an athenahealth Marketplace listing. Ask Luma which athenaOne workflows it completes with write-back, which it reads only, and which route to staff.
Platform breadth	Voice and messaging, scheduling and capacity, referrals and orders, intake and forms, revenue workflows, reporting.	Broader published surface, including patient pipeline for acquisition, feedback and reputation management, broadcast, and fax transform, alongside the core access workflows.
Voice architecture	Built voice-first. The agent calls athenaOne tools mid-conversation, so a call can check real availability and book inside the call rather than queueing work for afterwards.	A conversational agent plus outbound conversational AI for no-show recovery, layered onto a platform that began with asynchronous patient messaging and workflow.
Workflow configuration	Configured with the practice during the first 30 days live against live calls and the practice own scheduling rules.	No-code AI workflow creation, so the practice team can build and change workflows directly. A genuine advantage for groups with operations staff who want to own this.
athenahealth Marketplace record	45 practice connections, 22 ratings, all five-star. athenahealth Marketplace partner since 2025. As of October 2026.	39 practice connections and 1 review on the marketplace. A much larger company overall, but a comparable athenaOne-specific footprint.
Scheduling rule complexity	Reads appointment types, provider group and department rules and live slot availability directly from athenaOne, so the rules the practice already maintains are the rules the agent follows.	Scheduling across supported EHRs. Practices with unusual athenaOne rule structures should test their specific cases during evaluation.
Published pricing	Pricing model published: platform fee per location, per-provider fee for modules switched on, voice metered per call handled. Month-to-month, first 30 days live are free.	No public pricing.
Relevant scope	Single-EHR athenaOne practices where phone volume is the constraint.	Multi-EHR groups, or practices wanting the widest patient access surface from one vendor.
Where the coverage overlaps today	Referral intake, waitlists, reminders and recalls, registration and prep, intake and forms, eligibility, payments and feedback all run here, on athenaOne only, with the phone answered live in front of them.	The same territory across scheduling, referrals, waitlists, reminders and recalls, registration and prep, intake and forms, eligibility, payments and feedback management, multi-EHR, with no-code workflow building and a conversational agent.

A page that is not trying to talk you out of Luma

Most comparison pages are written on the assumption the reader is unhappy. This one is not, because in practice the athenaOne groups who ask us about Luma Health are usually running it and are usually fine with it.

So the useful thing is not a list of reasons to leave. It is a clear statement of what the two products are actually optimized for, so a practice can tell whether it has the problem that justifies a change.

Luma Health is a patient access platform built to work across EHRs. Its published surface is wide: scheduling, referrals, waitlists, reminders and recalls, registration and prep, intake and forms, eligibility, payments, feedback management, broadcast, a conversational agent, and no-code workflow building on top of all of it.

Pretty Good AI is narrower and only works on athenaOne. That is a deliberate constraint, and it is the whole basis of the difference.

What building for one EHR buys

We hold production access to 820+ athenaOne APIs, with workflows built around your practice’s rules.

That depth only means something when a workflow needs an API most integrations never touch. Reading the live slot table rather than a cached copy. Writing to the order and tickler queue. Pulling the appointment type and provider group rules the practice has already spent years maintaining, instead of asking staff to rebuild those rules a second time inside a vendor tool.

A multi-EHR platform’s athenaOne depth is a question to answer with a list, not an assumption. Ask Luma for the athenaOne workflows it completes with write-back, the ones it reads only, and the ones that route to staff, and compare that list to ours: booking and rescheduling, refills, referral fax intake charted, eligibility, balances and payment links, and waitlist backfill.

If you run athenaOne and expect to keep running it, use that workflow list to evaluate the fit. If you run athenaOne and eCW because you acquired a practice last year, ask both vendors how they handle each system and which exceptions route to staff.

The no-code question

Luma publishes no-code AI workflow creation. That is a genuine advantage and it is worth being explicit about, because it changes who has to be in the room when something needs to change.

A group with an operations lead who wants to build and adjust workflows without filing a vendor ticket will value that highly. Some of the best-run practices we see are exactly that shape.

Our model is different. Workflows get configured with the practice during the first 30 days live, against live calls and the practice’s own scheduling rules, and adjusted with us afterwards. For a practice without spare operations capacity that is the faster path. For a practice that wants to own the thing itself, it is a constraint they should weigh honestly.

Where the phone line changes the answer

The reason athenaOne practices call us while running a working access platform is almost always the same, and it is narrow.

Calls are not being answered. Volume spikes on Monday mornings and after holidays, hold times climb, and a measurable share of callers hang up before anyone picks up. The access platform is doing its job on reminders, forms and messaging, and none of that touches the queue on the phone.

That is a voice problem, and voice-first is an architectural position rather than a feature. An agent that can call athenaOne tools during the conversation can check real availability and book inside the call. An agent layered onto an asynchronous platform tends to capture the request and complete it afterwards, which works, and is a different experience for the caller who wanted to be done.

A specialty group running a busy line will recognize whether this is their problem within about a week of looking at their own abandoned-call rate. If it is not their problem, this whole section is irrelevant to them and they should stop reading.

An honest picture of overlap

Both products schedule. Both message patients. Both handle intake and forms. Running both without a decision about ownership produces the same mess any two-system setup produces, which is patients hearing from the practice twice about different things in the same week.

The arrangement that works is boring: the access platform keeps the asynchronous workflows it is already configured for, the voice agent owns the phone line, and somebody writes down which system initiates outbound contact for which workflow. That decision takes an afternoon and prevents the complaint that otherwise arrives in month two.

The short version

If Luma is already configured and staff know it, the question is not whether to rip it out. It is which workflows still land on a human. The phone line, the referral that stalled, the slot that opened and stayed empty: those are the jobs a broad access platform leaves behind, and they are the jobs we take.

Run athenaOne only, and have a phone line that is the actual constraint on how many patients you can see: that is the case worth a conversation, and it is a smaller claim than replacing your access platform.

Frequently asked questions

What does Luma Health do?: Luma Health describes itself as operational AI for healthcare. Its published platform covers patient access and retention, referrals, scheduling, waitlists, registration and prep, intake and forms, eligibility and verification, payments, patient communication, reminders and recalls, feedback management, broadcast, analytics and a conversational agent, with no-code AI workflow creation. It works across multiple EHRs and had 39 athenahealth Marketplace connections as of 5 September 2026.
Is Luma Health better than Pretty Good AI?: On breadth of the patient access surface, Luma publishes more. On athenaOne integration depth and on voice, we go deeper because we only build for one EHR. Neither of those makes one product better in general. The question that decides it is whether your practice runs one EHR or several, and whether your binding constraint is the phone line or the wider access workflow.
We already use Luma and like it. Should we switch?: Probably not, and we would rather say so than pretend otherwise. Ripping out a working patient access platform is expensive in staff retraining and workflow rebuild, and the gain has to be large to justify it. The case for talking to us is narrower: your phone line is still the bottleneck, calls are going unanswered or to voicemail at volume, and you want that specific problem solved deeply on athenaOne. That is a different purchase from replacing your access platform.
Can a practice run both?: Yes, and for some groups it is the sensible arrangement, with Luma owning the asynchronous access workflows it is already configured for and a voice-first agent owning the phone line. The rule is the same as with any two-system setup: decide explicitly which system owns outbound patient contact for which workflow, or patients will hear from both.
How do the athenahealth Marketplace numbers compare?: They are close. Luma had 39 practice connections and 1 review as of 5 September 2026; Pretty Good AI has 45 connections and 22 ratings, all five-star, as of October 2026. Luma is a much larger company with a much larger overall customer base, so the marketplace figure reflects athenaOne-specific adoption rather than company size, and it is the figure an athenaOne practice should weigh.
What should we test during an evaluation?: Your hardest scheduling rules, not your average ones. Take the appointment types that require a specific provider, a specific room or a preceding visit, and make each vendor book them live in your own environment. Also test what happens when a caller asks for something the agent cannot do, because the handoff to a person is where most front-office automation actually fails.

Sources

Everything stated here about another vendor comes from that vendor's own public material or from the athenahealth Marketplace listing, on the date shown. Vendors change their products and their pricing; if something below is out of date, email contact@prettygoodai.com and we will correct it.

Luma Health platform overview (accessed 2026-09-07)
Luma Health, athenahealth Marketplace listing (accessed 2026-09-07)
Pretty Good AI, athenahealth Marketplace listing (accessed 2026-09-07)

Why athenaOne practices evaluate Pretty Good AI

We are your AI team for athenaOne.

Your practice has its own providers, processes, and rules. Pretty Good AI works with your team to fit the first workflow to the work you want to improve, agree on how to measure it, and keep improving it as your practice changes.

Live and measured

100,000+: patient calls a month, at more than one customer
About 60%: of those calls handled start to finish by the AI at the largest deployments
Hundreds: of providers inside a single group
20: specialties on our athenahealth Marketplace listing

Customer-reported results. Your numbers will vary by workflow, staffing, seasonality and call mix.

How the platform works Pricing Customer results Security and compliance

Map your first workflow.

See it against your own athenaOne data

The honest way to compare is on your own call volume, your own schedule and your own payer mix. Book a working session and we will walk your numbers.

Book a demo How our pricing works