Pretty Good AI vs Transform9

Two vendors pointed at the same buyer

Transform9 describes itself as AI voice agents built for athena practices. We describe ourselves as an AI operations platform built exclusively for athenaOne practices.

Those sentences overlap almost entirely on the phone and diverge completely everywhere else. Working out which difference matters to you takes about an hour with data you already have.

Start by finding out whether you have a phone problem

This sounds obvious and is skipped constantly.

Pull your abandoned call rate, your average hold time, and the share of calls arriving after hours. Then pull three other numbers: how many referrals are sitting unworked, how many items are in your order and tickler queue, and how many appointment slots went unfilled last month after a cancellation.

If the first set is bad and the second set is fine, you have a phone problem. Buy the best voice agent you can and stop reading comparison pages. Transform9 is a serious option, particularly if you are orthopedic.

If both sets are bad, a voice agent fixes the part your patients complain about and leaves the part that costs you money. Unworked referrals and empty slots are revenue that was already yours.

Most specialty practices we talk to have both, and are surprised by the second set because nobody reports on it.

Where Transform9 is genuinely strong

Their athenaOne install base is concentrated: ten of eighteen practice connections are orthopedic surgery, and four more are ambulatory surgery centers.

Concentration like that is worth something real. Orthopedic scheduling is one of the harder rule sets in ambulatory care. Procedures and office visits book differently. Imaging has to precede certain visits. Post-operative follow-ups sit in fixed windows relative to a surgery date. Surgical block time constrains everything around it.

A vendor that has configured those rules across a dozen orthopedic groups starts ahead of one seeing them for the first time. If you are an orthopedic or sports medicine group, ask them for a reference in your subspecialty at your size, and take it seriously.

They also publish a scheduling rules engine as a headline capability, which is the right thing to build a voice product around. Rules are where these deployments succeed or fail.

Where the architectures differ

We take a different route to the same problem. Rather than building a rules engine the practice configures, we read the rules already configured in athenaOne: appointment types, scheduling templates and slots, provider groups, departments.

The reason is maintenance rather than elegance. Rules configured in two places drift. A practice that adds a new visit type in athenaOne and forgets to add it to the vendor’s engine gets an agent confidently booking something wrong, and nobody finds out until a patient arrives on the wrong day for the wrong thing.

That is also the argument for connecting 500+ of the roughly 800 endpoints athenahealth exposes. Reading the live configuration rather than a copy of it requires reaching most of the system.

The complication that separates a voice agent from a platform

A multi-site orthopedic group we work with had a capacity problem hiding behind a phone problem.

Their phone metrics were poor and got fixed quickly. Calls answered, holds gone, bookings made in the call. The expected result.

Their visit volume barely moved.

The reason was on the other side of the schedule. Cancellations inside 48 hours left slots that nobody backfilled, because backfilling meant somebody working a waitlist by hand against rules about which patients could take which slot. A post-operative follow-up could not take a new-patient slot. An imaging-dependent visit could not take a slot before the imaging. The waitlist existed and was almost never worked.

What changed it was reading the cancellation, filtering the waitlist against the same appointment type and provider group rules already in athenaOne, and calling the patients who actually fit, in order, until the slot filled.

Two things still go to a person. If the only candidates need a clinical decision about whether an earlier visit is appropriate, that goes to the practice, because it is a care judgment and not the automation’s to make. And if a slot is inside the post-operative window for a surgeon who is out, a human reassigns it.

The phone work and the backfill work are the same integration. They are not the same product, and a voice agent will not do the second one.

Both of us will be at Thrive

Transform9 is a sponsor of athenahealth Thrive 2026 and we are a Platinum sponsor, so both companies will be in front of the same athenaOne buyers in the same week.

Bring your hardest appointment type to both booths and ask for it to be booked live in an athenaOne environment, not a sandbox. For a surgical practice that means the procedure requiring a specific provider, a specific location and a preceding visit.

Then ask each vendor what happens when the agent cannot finish, and ask to see what the staff member on the other end of the handoff actually receives. Ask how many athenahealth API endpoints each connects.

And ask both companies for the same three evaluation numbers measured in your environment rather than in a case study: calls finished without a human, calls handed off, and what happened to the handoffs.

The honest summary

If your problem is the phone and you are orthopedic, Transform9 has real depth where you need it and a shorter product to run.

If the phone is one of several places work is piling up, and referrals, recalls, empty slots and denials are the rest, the phone agent is a third of the answer. That is the case an athenaOne-only platform is built for. Pricing is published, month-to-month, and the first 30 days live are free.