Skip to main content
Workflow · Inbox and task automation

The queue is the backlog. AI works the administrative half of it.

Most of what sits in an athenaOne queue is not a clinical decision waiting on a clinician. It is a phone call nobody has made, a date nobody has set, a form nobody has chased. Pretty Good AI works those items to completion inside athenaOne and leaves everything that needs clinical judgment in the queue for the person who owns it.

Written for: Practice managers, COOs and clinical operations leaders at athenaOne practices where staff and providers are working down queues that keep refilling.

The problem

A growing queue is usually read as a staffing problem. It is more often a sorting problem. Administrative items that need a call, a booking or a document sit in the same list as items that genuinely need a clinician, so the clinical work waits behind the clerical work and the practice sees one undifferentiated backlog. Hiring clears it for a quarter, then it refills.

What starts it

A case lands in an athenaOne queue: a voicemail written to the chart as a case, a tickler coming due, an order opened without an appointment, a result released with a callback to arrange, a refill request arriving from a patient or pharmacy, or a referral held for review.

What the AI does, step by step

  1. Read the queue and sort it

    Items are classified against the rules your practice agreed: what is administrative and can be worked, what is clinical and stays, and what is ambiguous and therefore stays. Sorting is the whole job. A queue that has been split into work-this and leave-this is already easier for staff to face than one that has not.

  2. Work the administrative items to completion

    The agent makes the call, offers the slots your templates allow, books the visit, sends the form, chases the missing insurance detail or requests the document. This is the same rule set the voice agent follows on an inbound call, so a patient gets one answer whichever direction the contact runs.

  3. Write the result back to the record

    The appointment lands on the athenaOne schedule, the case is documented and closed or reassigned, the tickler that prompted the outreach is cancelled once the patient is booked, and the document returns to the chart. Nothing finishes in a separate system that staff then have to reconcile.

  4. Return the rest, better prepared

    Items that need a clinician go back to the person who owns them with the administrative groundwork already done: the patient reached, the callback scheduled, the outside records requested, the coverage confirmed. The decision is untouched. The clerical work in front of it is not.

What changes in athenaOne

  • Cases documented and closed, or reassigned to the queue your practice designated, with the outreach recorded.
  • Appointments booked on the athenaOne schedule against the provider, location and visit type the order or tickler requires.
  • Ticklers cancelled once the patient they prompted is scheduled, so the same reminder does not generate a second call.
  • Requested forms, consents and insurance updates returned to the patient record.
  • Outreach attempts and outcomes logged against the item, so an unreached patient is visible rather than silently dropped.

Completion milestones

Worked is not the same as resolved. An item is worked when the agent has done everything the workflow allows and recorded the outcome. It is resolved when the underlying task is actually finished, which for an unreachable patient may take several attempts and for a clinical item is never the agent’s call to make. Agree which athenaOne signal marks each of those, separately, before launch.

What your practice controls

  • Which queues are in scope, and which are never touched.
  • The classification rules that decide what counts as administrative at your practice, reviewed by your clinical leadership rather than assumed.
  • How many outreach attempts an item gets, over what window, and at what times of day.
  • Which items are always handed to a person regardless of how they classify.
  • Where a worked item goes when it closes, and who owns the exception queue.

Where a person steps in

  • Clinical decisions stay with your care team. The agent does not assess symptoms, rank patients by clinical urgency, interpret a result or decide what care a patient needs.
  • Prescription requests are administrative handling only. The agent can gather what the request is missing, match it to the right chart and route it to the queue your practice designates. It does not create, approve, sign or transmit a prescription, and it does not make refill decisions.
  • Results are scheduled, never conveyed. The agent can arrange the callback and confirm the patient will be there. What the result means is said by your clinical staff.
  • An item the rules cannot classify confidently is left in place and flagged, not guessed at. Ambiguity resolves toward the human every time.
  • Patients who ask a clinical question mid-call are routed to your care team with the administrative part of the task already recorded.

Your queue conventions, not ours

Queue conventions are the least standardized thing in any practice. Two athenaOne groups in the same specialty will use different case types, route results differently and disagree about what the front desk is allowed to close. A product with one opinion about how a queue should work forces one of them to change. We build to the convention you already run.

  • We map the queues as your staff actually use them, including the local habits that never made it into a policy document, and automate that. Renaming your case types to suit a vendor is not implementation work worth doing.
  • Depth of integration is what makes an unusual workflow buildable at all. Most vendors on the marketplace pick a narrower wedge and build deeply inside it; we carry significant depth across athenaOne, which is why a queue rule specific to your practice is a scoping question rather than a roadmap request.
  • New workflow scope is measured in days and weeks, not quarters. A queue behavior identified during onboarding ships inside the same onboarding window.
  • The classification boundary is yours and is reviewed by your clinical leadership before anything runs. We do not ship a default opinion about which of your queues is administrative.

Illustrative walkthrough

An order sitting open for three weeks

An order exists with no appointment against it. The agent calls the patient, translates the order into the name of the test rather than the internal code, offers the slots that visit type allows, books it in athenaOne and documents the outreach on the case. If the patient asks why the test was ordered, that question goes to the clinic. If the patient cannot be reached after the agreed attempts, the item returns to staff showing what was tried and when, instead of aging quietly in the queue.

Synthetic example, not a customer result or a live product test.

How it gets implemented

  1. Agree the first queue and the boundary

    We pick one queue, usually ticklers or open orders, and write down what is administrative in it. Your clinical leadership signs off on that boundary. This conversation is the implementation; the automation is the easy part.

  2. Build to your athenaOne configuration

    Production access to 820+ athenaOne APIs, with workflows built around your practice's rules. Your case types, queue routing, appointment types and provider templates stay as they are.

  3. Review and go live

    You watch the agent work a sample of real items against your own rules before it runs unattended, and you keep the exception queue in view afterwards. Typical time from kickoff to the first workflow live is 3 to 6 weeks. We agree on the first workflow, how success will be measured, and the pricing for continued use before launch. There are no setup fees. Your first 30 days live are free. The clock starts when your first agreed workflow goes live in your practice, whether that is voice, web scheduling, referrals or another workflow. When the free 30 days end, service continues automatically, month to month, at the pricing in your order form. New workflows are priced when activated.

Evidence and its limits

Commonwealth Pain & Spine: deployment

Commonwealth Pain & Spine runs Pretty Good AI across 35 locations on its existing athenaOne configuration. That establishes deployment at scale across a multi-site group, not a measured reduction in queue volume. Ask for the workflow, the period and the denominator behind any queue figure, ours included.

See where it is deployed →

Deployment evidence is not a measured conversion improvement. Read any outcome figure with its stated workflow, period and denominator. Customer-reported results. Your numbers will vary by workflow, staffing, seasonality and call mix.

Questions buyers ask

How do you decide what is administrative and what is clinical?

Your clinical leadership decides, and it is written down before anything runs. We bring a starting boundary from other athenaOne practices, you change it, and ambiguous items resolve toward the human by default. A vendor that claims to know where that line sits at your practice has not asked.

Does this touch refill requests?

Only the administrative handling. The agent can gather missing information, match the request to the right chart and route it to the queue you designate. It does not create, approve, sign or transmit a prescription, and it does not decide whether a refill is appropriate. That boundary is regulatory, not a configuration setting.

Will it tell patients their results?

No. It arranges the callback and confirms the patient will be reachable. Your clinical staff say what the result means.

What happens to an item it cannot finish?

It returns to staff showing what was attempted and when. An unreachable patient after the agreed number of attempts is a visible exception rather than a closed item, which is the opposite of what happens when the same task quietly ages in a queue.

Is this a separate inbox our staff have to learn?

No. The work happens in athenaOne against the cases, ticklers and orders your staff already use. Reporting on what the agent worked is available in the Command Center.

How is it priced?

We agree on the first workflow, how success will be measured, and the pricing for continued use before launch. There are no setup fees. Your first 30 days live are free. The clock starts when your first agreed workflow goes live in your practice, whether that is voice, web scheduling, referrals or another workflow. When the free 30 days end, service continues automatically, month to month, at the pricing in your order form. New workflows are priced when activated.

Tell us how your front office runs today.

Book a scoping session and we will walk one of your real queues, using a synthetic example, and mark which items are administrative at your practice. Do not send patient information through public website or booking forms.

Book a scoping session →

Opens our scheduling calendar. Choosing a time books the session.

Book a strategy call