Pretty Good AI vs OhMD

A texting product and an operations platform

Pretty Good AI started at voice, with an AI agent, against athenaOne, and grew outward from there into secure two-way text and the operations work behind the conversation. OhMD started as a texting platform and has added an AI agent on top. Both of us will be Platinum sponsors at athenahealth Thrive 2026, so an athenaOne practice can put the two side by side this year and ask the same questions of each.

OhMD covers the patient conversation. We cover the conversation and the operations behind it, on one athenaOne integration. That is the whole comparison, and the rest of this page is the evidence.

What “added AI” means in practice

OhMD’s agent is called Nia, and the company has been public about how it is built. Its own November 2025 announcement credits a third-party communications platform for the live voice, messaging, workflow and analytics layers, and quotes its CEO saying that platform is what lets them deliver at scale.

That is a reasonable way to ship an agent quickly on top of a messaging business. It also means the agent and the EHR integration are separate pieces of engineering, assembled, and voice is layered onto an architecture designed for asynchronous messaging. Ours are the same piece, built voice-first. When the athenaOne write path needs to change because a practice’s scheduling rules changed, that is one team, one codebase, one release.

Why the integration is the product

Voice works under a specific constraint: there is no draft. The patient is on the line, the agent gets one pass, and whatever it decides has to be right and has to land in athenaOne before the call ends.

That constraint is why we hold production access to 820+ athenaOne APIs, significant depth across the whole of athenaOne where most vendors pick a narrower wedge and build inside it, and it is the honest answer to the question practices actually ask, which is why some vendors’ calls finish and others’ calls turn into a task. It is not voice quality. Everyone’s voice is good now. Calls finish when the agent can read the live athenaOne record mid-call, the real scheduling template and slot type, the appointment types a given department actually uses, the provider group, the insurance on file, the order and tickler queues, and write back before it hangs up.

Here is what that looks like when it meets a real schedule. Practices do not build calendars out of tidy named appointment types. They build them out of generic slots, an “Any 15” or an “Any 30”, and the system will return a generic slot when you search for a specific visit. An “Any 15” may genuinely need to be an hour for a new patient. Some appointment types can only be booked when a companion type can be booked too. Resolving that means knowing which specific appointment types each generic slot is eligible for, per provider, per department, and the mapping is different at every practice. Get it wrong and the front desk finds out at check-in. There is no conversational cleverness that substitutes for reading the configuration.

Texting is one channel, not the product

Secure two-way text runs on the same athenaOne integration as voice, so the thread and the call share context and write to the same records, and staff pick up a conversation where the agent left it rather than starting over. If texting is what you mostly want, you get it from us, on one integration with the phone.

What you also get is everything a messaging platform does not reach: fax, portal and web referral intake, tickler and order outreach, waitlist backfill and no-show recovery, eligibility and prior-auth prep, denials and appeals, balance outreach under your billing team’s rules, post-visit outcome calls and surveys, forms signed inside athenaOne, and per-call QA with KPI reporting. One vendor, one integration, one set of records, from the phone and the thread through to the back office. Every figure on this page was checked against the vendor’s own pages and the athenahealth Marketplace in September 2026.

What we would tell you to check

Ask both vendors, at Thrive or on a call, to show a call and a text thread that end with athenaOne updated and no task created for staff. Then ask what happens on the ones that fail, who sees the exception, and who maintains the workflow when your rules change. Then ask what runs behind the conversation.

Then get the total. Ours is a platform fee per location, a per-provider fee for the modules you switch on, and voice per call, with the first 30 days live free and month-to-month terms after. Ask what a comparable month costs once calling usage and any tier minimums are added on. The failure path, the scope behind the conversation and the real monthly number are where these products differ, and none of them appears on a feature grid.