Pretty Good AI vs CallMyDoc

Start with the thing that is true

CallMyDoc has been on the athenahealth Marketplace since 2017 and has 300 practice connections and 113 reviews at a 4.8 average. We have 45 practice connections, 22 ratings, all five-star, as of October 2026. On length of track record alone the numbers favor them.

What we would argue is that track record answers a different question than the one most practices are actually asking in 2026.

Message routing versus work completion

The generation of tools CallMyDoc comes from was built to make sure nothing gets dropped. A call comes in after hours, the system captures it, categorizes it, and gets it in front of the right person by morning. That was a large improvement on an answering service and it is still valuable.

The question now is different: how much of that work can end without a person at all? Not “was the message routed correctly”, but “was the appointment booked against the right appointment type and provider group, was the refill put in the right queue, was the referral attached to the order, was the tickler closed”.

That is the problem Pretty Good AI is built for, with production access to 820+ athenaOne APIs. We build workflows around athenaOne’s records and the practice’s rules, connecting the steps needed to move the work forward and making exceptions visible to the team.

The overlap is real

Both products answer calls, both do self-scheduling, both send reminders, both write back to the chart. A practice that buys either one will stop missing calls. If that is the whole brief, this is a close comparison. Most practices we talk to find the brief gets bigger the moment somebody counts what the front office is still doing by hand.

Scope to confirm for your practice

If the board conversation is “we are paying an answering service too much and patients complain about it”, and nobody is talking about referral leakage, denials or schedule density, you are evaluating a narrower product than the one we sell. Worth asking why those other numbers are not on the agenda.

If you run practice management systems other than athenaOne, they support that and we do not.

Where we would argue for us

The practice that should be talking to us is the one where the phone is one symptom of a front office that is behind. Referrals sitting in a fax queue. Ticklers nobody works. Eligibility checked twice or not at all. Denials worked late. Those are separate purchases with most vendors. With us they are modules on an integration you already have, and you turn them on when you are ready.

Pretty Good AI also serves smaller practices, but we have a stronger focus on call centers and central patient-access and support teams. We are especially suited to complex specialties, multiple locations, and high or complex appointment and phone volume, where the work needs to finish inside athenaOne rather than stop at message capture.

What a 2017 architecture assumes

CallMyDoc has been on the athenahealth Marketplace since 2017, and the product is built around routing a message to the right person. That was the right design for 2017: the software captured and routed, and a staff member did the work.

Products designed after generative models became reliable start somewhere else. The agent is expected to complete the administrative request in the chart, not to deliver it to a queue. A vendor with eight years of customers on the routing model carries that model in its data structures and its contracts, which is why the newer capability tends to show up alongside the old one rather than replacing it.

How to decide in thirty days

Ask both vendors for their first 30 days live on calls in your own environment, then compare two numbers: what percentage of calls ended with no human action required, and what percentage ended with a task in someone’s queue. Everything else in the evaluation is commentary on those two numbers.