Pretty Good AI vs Spruce Health

A better inbox is not the same purchase as the work

Spruce and Pretty Good AI both show up in the same evaluation, usually under a heading like patient communication. They are not the same kind of product, and the difference is easy to miss on a feature grid because both lists contain the word phone and the word text.

Spruce sells a communication platform. Phone, secure messaging, two-way SMS, video, e-fax, team messaging and notes, in one shared inbox, priced per user. The pricing is public.

We sell an operation. The calls, the texts, the referrals, the schedule, the staff queues, the authorizations and the balances, run as one thing on one athenaOne integration, by us.

Put plainly: one of these makes your staff faster at the work. The other one does the work. They are not interchangeable, and practices get into trouble when they budget for the first and expect the second.

Per seat is a pricing model with an opinion in it

Spruce publishes $24 per month per user on Basic and $49 per month per user on Communicator. The price is not the point.

The point is what the unit is. When the unit is a seat, your capacity to answer patients is a function of how many people you employ. Volume goes up 30% in January and you have two options: add seats and the people to sit in them, or let the overflow go to voicemail. The license fee was never the constraint. The person in front of it always was.

We meter voice per call actually handled, on top of a platform fee per location and a per-provider fee for the modules you switch on. No setup fees, first 30 days live free from go-live, month to month after. When January arrives, the volume line moves and the staffing plan does not have to.

This is the single largest structural difference between the two products, and it is the one least likely to come up in a demo.

What a shared inbox does with the hard call

Here is a case that looks trivial and is not.

A mother calls a multi-site GI group about her daughter. Both are patients. The phone number on file belongs to the mother, so every message from that number threads into the mother’s chart. Staff notice a week later, when the daughter’s referral has not moved, and move the messages by hand.

A shared inbox handles this exactly as designed. The thread is correct, unified and in one place, and it is attached to the wrong patient. The tool did nothing wrong. The work is still wrong.

We treat this as a defect class rather than an edge case, because parent and child and spouse and spouse are common enough in any real panel to be routine. The agent asks whether the call is about the caller or someone else before anything gets written, and it detects the mid-call handoff when one person passes the phone to another. When identity cannot be resolved with confidence, it does not guess. It routes to a person with what it knows.

That is the general shape of the difference. A communication tool gives a human a better place to find the problem. An operation is accountable for the problem not happening.

Where the two products stop overlapping

Two things Spruce publishes that we do not sell, and it is worth being straight about both.

Internal team messaging with notes on a patient thread is one. Video visits are the other. Neither is something we sell, and neither conflicts with us.

What we would push back on is the assumption that the patient-facing phone line is covered because those two are. Those are collaboration and care-delivery tools. The inbound line is an operations problem, and it is the one with the revenue attached.

Depth is why the work finishes in the chart

Plenty of tools connect to athenaOne. Far fewer finish work inside it.

Sending a text or writing a note touches a handful of endpoints. Booking a new patient against the right appointment type and template, resolving a failed eligibility check, prepping an authorization by payer, or completing a referral that has sat since July touches far more of the system, and every gap becomes a task for a human.

We hold production read and write access to 820+ athenaOne APIs, with no middleware in between, and our app runs inside athenaOne with no extra login. That is the difference between a workflow that ends in the chart and a workflow that ends in somebody’s queue.

It is also why we only do athenaOne. Building to the full capability of one system and building to the common ground of many are different products, and we picked the first.

What this looks like in a real practice

Picture a four-site primary care group, about thirty providers, and six people on the phones across all of it.

They run a per-seat communication platform. The inbox is tidy, faxes arrive digitally, voicemails come in transcribed, and the after-hours tree routes correctly. Nobody is asking to rip it out.

Then pull the numbers. Roughly one in five inbound calls in the 8am to 10am block never got picked up. Voicemails from that block get worked by lunch on a good day and by the next morning on a normal one. Refill requests sit in the shared inbox behind scheduling calls because scheduling calls feel more urgent. The recall list has not been worked in six weeks because nobody has an hour.

None of that is a tool failure. Every one of those messages is sitting in a well-organized inbox. It is a capacity failure, and the only lever a per-seat tool offers against it is another seat.

We start with the 8am block. Calls get answered on the first ring, the booking happens in athenaOne against the right appointment type, refills get taken and routed, and the ones that need a person arrive as a case with the name, date of birth and reason for calling already in the chart. The recall list runs overnight because running it costs no hours. The group keeps its team messaging and its video visits, because those were never the problem.

Three questions that settle it

Is your problem organization or capacity? If your team can handle the volume and the mess is where things live, you are shopping for a better inbox. If the volume is beating six people every morning, a tidier queue does not add an hour to the day.

How much of the work do you want finished in athenaOne? There is depth available to an athenaOne practice that a broadly-sold platform cannot spend its roadmap on, and that gap is the whole comparison.

Who owns it after go-live? A per-seat tool is owned by whoever administers it, which is usually an office manager with another job. An operation spanning phone, referrals, scheduling and revenue is not something to hand to that person, which is why we run it. When athenaOne changes or a rule changes, fixing it is our job.

Before your next leadership meeting, count two things: calls that went to voicemail last week in your busiest hour, and how many of those patients called back. Put a dollar figure on the ones who did not. That number usually decides which of these two products you are actually shopping for.