How Clearway Rolled Out AI Across athenaOne

Clearway Pain Solutions rolled out AI on its athenaOne patient access team in stages: one-hour “power hours” live in the queue, a command center watching calls as they came in, the simpler workflows first, and expansion only after the organization trusted it. Jeannine Spagna, vice president of patient access and experience, and Sami of the patient access team walked through the whole rollout on a recorded webinar. This page is the written version, with what they said and where their staff still step in.

Practice context

Clearway is a pain management organization with more than 100 locations on athenaOne. One centralized patient access team answers for every clinic. Patients see one organization; behind the phone line are many providers, several specialties, and scheduling workflows that differ by office, state and region.

The organization was growing quickly, and call volume grew with it. Spagna’s description of the problem was blunt: hiring a new class of agents every month or two was not going to be a sustainable answer, and she did not want growth to come at the expense of the patient experience. Her framing on the webinar was that AI was never the strategy. Patient access was the strategy, and AI was one way to get there.

That pressure is common. In a March 2026 MGMA Stat poll, practice leaders named eligibility and prior authorization (45%) and scheduling (31%) as the most time-intensive tasks for staff on the phones.

Why they picked a vendor that could act in the chart

The team’s test for every vendor was simple: show us. Most could not. Several products they saw only transcribed calls, which is not what a centralized access team needs.

What decided it was range inside athenaOne. Creating patient cases, scheduling, checking details and validating notes in the chart were all one system rather than a patchwork of point vendors. As the Clearway team put it on the webinar: “It’s not just like a layer on top… they’re in there.”

The complication: the same patient, different providers, different queues

Pain management charts are messy in a specific way. A patient may see one provider for medication management and a different provider for procedures, and which provider that is can change from month to month as treatment moves. When a patient calls and asks “where’s my appointment?”, the agent has to work out whether they mean the last visit or the next one, and with which provider and department.

The routing is just as specific. Take a routine call: a patient’s pharmacy cannot fill a prescription and they want it sent to a different location. The AI confirms the patient’s identity, which medication, and which pharmacy, often from a street name or cross streets rather than an address. It then creates the case in athenaOne and puts it in the right bucket for that patient’s office and team. The AI does not move or approve the prescription. The clinic team processes the request and follows up with the patient. As Sami pointed out, the conversation is the visible part; the real work is getting the case to the right queue so it is not delayed.

Outbound works the same way. When an order is on file for a procedure, the AI calls the patient, confirms the order, offers open procedure slots and books one. Questions about the procedure itself go to the clinic.

How the rollout worked

1. Discovery. Clearway already had written SOPs and detailed work logs for its processes, which the team used to build the agent’s workflows and routing rules.
2. Power hours. The AI went live in the queue for one hour at a time. Spagna sent meeting invitations to providers, office staff and even HR, and asked them to call in and give it the hardest request they could think of.
3. Command center. During each live window the two teams monitored calls together as they came in and made adjustments in real time.
4. Longer windows. One hour became half a day, then a full day, then 24/7 coverage.
5. Easy workflows first. Clearway deliberately started with simpler, high-volume call types to show the organization what the AI could do before moving to the more complex, revenue-driving workflows.

Spagna said the pace was set by Clearway, not by the technology. The slower parts were stakeholder communication and buy-in, which she treated as the actual work of the rollout.

What the practice controls

• The wording. Spagna still tunes the agent’s scripts herself, including the outbound scripts, to keep the Clearway tone of voice. She describes it as coaching the agent the way she would coach a new team member.
• What the AI does not handle. Some requests always go to a person, by design. In her words, just because the AI could do something “doesn’t mean she should.”
• Choice for the patient. Any caller who asks for a person is transferred. Clearway has done no promotion of the AI to patients yet.
• The business case. Spagna presented the project to leadership as budget neutral, funded from what the organization already spends on access staff, rather than as a new line item.

Where staff still step in

• Every caller who asks for a person.
• Call types Clearway has chosen to keep with staff.
• Prescription requests, which the AI captures and routes to the clinic team rather than acting on.
• Clinical questions, including questions about a procedure, which go to the clinic.

The staff side of the change was handled before launch. Clearway held a leadership offsite first, then positioned the AI as a tool that takes simple, repetitive calls so agents can spend their time on the calls that need a person. Staff tested it themselves during power hours and fed back what they heard.

Results

This story is qualitative. Clearway discussed early results on the webinar, but no Clearway figures are published here, and none should be attached to Clearway from any other customer’s numbers. Spagna’s own summary of what mattered was not a metric: “it’s one thing to book the appointment, right? It’s another to get them in the door.” Her answer to the cost question was the question her leadership asked themselves: could they afford not to do it, given their growth?

For a step-by-step version of the go-live checks, see the athenaOne workflow acceptance checklist. For related pain management workflows, see procedure outreach for pain management and prior auth expiry windows.

Sources: the recorded Clearway webinar linked above; MGMA, Phones are still a bottleneck costing medical practices time they can’t afford.