Front offices are carrying more than they can hold. Call volume keeps climbing while the people who answer those calls turn over faster than almost any other role in the practice, and the gap between the two widens every quarter. Patients feel it as hold music and voicemail. Staff feel it as a queue that never empties.
When leaders evaluate voice AI to close that gap, the pitch usually centers on how natural the agent sounds. That is the wrong question. The one that matters is narrower and harder: what can it finish? A voice agent that answers a question and hands the work back to your staff has removed nothing from the queue. A capability is a task completed end-to-end, a booked appointment, a verified insurance record, a prescription renewal filed in the EHR, rather than a well-phrased reply.
This reference is organized around that distinction. It walks through the concrete workflows Artera’s agents handle today, from scheduling and intake through referral routing, prescription renewal, care gap outreach, and after-hours support, and for each one explains what the capability does, where it sits in the patient journey, and how to evaluate whether it is real.
Voice is treated throughout as one channel inside a full patient engagement platform, rather than a standalone phone bot bolted onto the front desk.
What AI voice agent capabilities for healthcare actually mean
Three different technologies get called voice AI, and confusing them is the fastest way to buy the wrong thing.
An IVR menu routes: Press 1 for scheduling, press 2 for billing. It follows a fixed decision tree and cannot handle anything the tree did not anticipate. An ambient scribe listens and documents. It turns a clinical conversation into a note, and it does not act on the patient’s behalf.
A voice agent differs in kind: it holds a natural conversation and completes the task the conversation is about. It does more than understand that a patient wants to move an appointment. It reschedules the appointment, checks the specialty’s rules, and writes the change back to the system of record.
Capability is the unit that matters, not conversation quality. The industry’s first wave of voice tools measured whether the agent answered. The standard now is whether it handled the task, start to finish, without a staff member picking up the remainder.
Read that way, capabilities become a maturity question, and this reference answers it plainly. Some workflows below are proven in production across many organizations. Others are emerging. A few are early pilots running at a single site. Where a capability sits on that curve appears in each section, because a capability being real and a capability being proven at scale are two separate claims, and buyers deserve to know which one they are getting.
Scheduling capabilities
Scheduling is where voice AI earns its keep, because scheduling is where the phone still carries the load. After years of patient-portal investment, 73% of medical groups report that a quarter or fewer of their patients use digital tools to book, and only 11% say a majority self-schedule (MGMA, Putting the Power of Scheduling into Patients’ Hands, 2026). The self-service future has not arrived. The calls have.
A capable scheduling program covers the full appointment lifecycle, well beyond net-new booking. Artera splits it across two proven agents. The scheduling agent handles inbound voice requests and outbound triggers, booking net-new appointments directly into the EHR, and runs in production across specialty and FQHC organizations on Athena, eCW, and NextGen. The appointment-management agent verifies, reschedules, and cancels existing appointments across inbound and outbound calls combined, proven across ten organizations from specialty clinics to FQHCs.
The difference between a demo and a deployment shows up in the write-back. If the agent books a visit and a staff member re-enters it in the EHR, the queue did not get shorter. These are the most mature capabilities in the platform, which is why they anchor the reference.
Specialty-aware scheduling rules
Generic scheduling breaks in specialty care because specialty care does not schedule generically. A gastroenterology visit that involves a procedure carries prep requirements and sequencing rules a primary-care slot never touches. Orthopedics routes by injury type and provider sub-specialty. A colonoscopy cannot be booked the way a follow-up is.
An agent that treats every appointment as an interchangeable 20-minute block will book the wrong thing with confidence, which is worse than not booking it, because now a staff member has to catch and unwind the error. Specialty-aware scheduling encodes the rules of this specialty in the booking logic: visit types, provider constraints, prep dependencies, and the questions that have to be answered before a slot is offered. That is the line between a scheduling agent that reduces work and one that quietly creates it.
Intake and insurance verification capabilities
Intake is the workflow text was supposed to solve and never fully did. Portal forms and pre-visit texts work for the patients who open them. The ones who do not are the same patients whose intake is incomplete when they arrive, and the front desk absorbs the difference. Voice fills that gap, because a patient who will not tap through six form screens will often answer the same questions on a call.
A capable intake flow handles pre-registration ahead of the visit, demographic capture and confirmation, form auto-population that turns spoken answers into structured fields, and insurance verification inside the scheduling conversation. The evaluation test matches scheduling: does the captured information land in the record as structured data, or as a transcript someone has to process? Auto-population is the point. If intake answers arrive as a block of text a staff member re-keys, the agent moved the work instead of removing it.
Insurance verification and form auto-population
Insurance verification is one of the highest-friction moments in intake and one of the most error-prone when the front desk rushes it. An agent that verifies eligibility as part of the scheduling conversation catches coverage problems before the visit instead of at check-in, when the fix is a phone call rather than a turned-away patient.
Pre-appointment prep capabilities
Some visits fail before the patient arrives, because the prep did not happen. An MRI gets canceled over an unscreened contraindication. A colonoscopy is wasted because the patient did not follow the prep. A form that should have been done in advance backs up the waiting room. Pre-appointment prep is outbound by nature: the agent reaches the patient ahead of the visit and walks them through what has to be true before they show up.
A great example of these capabilities at work: Atlantic Health, a large health system based serving patients across New Jersey, New York and Pennsylvania, used Artera’s AI agents specifically for colonoscopy outreach – guiding them through the prep journey and closing the screening gap.
Prescription renewal capabilities
Refill requests are high-volume, repetitive, and clinically bounded, a strong fit for a voice agent and a real relief for staff who field the same call dozens of times a day. The capability goes past answering the renewal call. It files the request into the EHR so the clinical team can act on it without transcribing anything.
Referral routing capabilities
Referrals leak. A referral that comes in by phone or fax and waits in a pile is a patient who may never get scheduled, which is a care gap and lost volume at once in specialty care. Referral routing tracks the inbound referral and moves it to the right place instead of letting it sit.
Care gap outreach capabilities
Care gap outreach is proactive rather than reactive. Instead of waiting for the patient to call, the agent reaches out to patients overdue for preventive care, closing gaps and running screenings, including social-determinant screenings, at population scale. It is the clearest example of voice AI doing work the practice would otherwise never staff, because no front office has the hours to call every overdue patient.
Call routing and telephone encounter capabilities
Not every call is a task the agent can finish itself, and a capable system knows the difference. Call routing triages inbound calls, sends them to the right provider or team, and creates the telephone encounter in the EHR so nothing falls through, turning the switchboard from a bottleneck into a routing layer that documents itself.
Warm handoff capabilities
When an agent does need to transfer to a person, the handoff itself is a capability. A cold transfer that drops the patient into another queue undoes the goodwill the agent earned. Warm handoff detects a live human pickup during a transfer, telling a real person apart from a hold queue or voicemail, so the patient connects only when someone is there.
Medication adherence capabilities
Adherence outreach is the clinical cousin of prescription renewal: outbound calls that check whether patients are staying on their medications, which shapes outcomes long after the visit ends. It extends the patient relationship into the space between appointments, where most of care happens.
Conversational FAQ and knowledge base capabilities
Not every patient question needs a person, and every patient question needs an accurate answer. A conversational FAQ agent answers patient questions from a maintained knowledge base, across voice and messaging, so the front desk stops answering the same questions all day and patients stop waiting on hold to hear something the practice already knows.
HR benefits FAQ capabilities
The same agentic approach that serves patients can serve the workforce behind them. An HR benefits agent answers employee questions, the repetitive and well-documented queries that tie up HR teams, freeing people for the work that needs judgment.
SDOH and patient-reported outcome capabilities
Payers and value-based arrangements increasingly require structured data the practice has to collect from patients, including social-determinant screenings and patient-reported outcomes. A conversational agent gathers that data at scale and returns it as structured fields, instead of asking staff to run screenings one call at a time.
Artera’s SDOH questionnaire agent is an early-stage capability collecting patient-reported outcomes for payer requirements. It is a reminder that patient interaction includes the data obligations behind the visit, not only the visit itself.
After-hours support capabilities
Care needs do not stop at closing time, and the practice’s ability to respond should not either. An after-hours agent handles triage calls when the office is closed, routing urgent needs to the right destination, an affiliated healthcare organization or an on-call line, instead of leaving patients with voicemail.
How to read these capabilities
Read as a set, these workflows make one point: voice AI for healthcare is not a single product with a single maturity. Scheduling and appointment management are proven across many organizations. Prep and prescription renewal are emerging. Most of the rest are early pilots, real and running and being measured, and not yet proven at scale. A reference that blurred those distinctions would be easier to write and worse to trust.
The through-line is that these are capabilities, not conversations. Each one earns its place the same way: it completes the task, writes back to the system of record, and respects the rules of the specialty it serves. And each one is one channel inside a full patient engagement platform that sees the whole journey, which is what separates an agent that finishes the work from a phone bot that answers.
