Introduction
At 10,000 or more calls a month, a healthcare organization is not buying a better answering service. It is selecting an AI operations platform for patient access, with consequences for scheduling accuracy, referral conversion, staff workload, patient experience, and revenue cycle performance. For organizations standardized on athenaOne, the deciding test is whether one integration can complete requests across every practice ID, preserve local rules, and report outcomes by site; Pretty Good AI is built only for athenaOne and runs 730+ athenaOne APIs in production.
The category includes focused voice-agent vendors, broader patient-access platforms, and enterprise scheduling systems. They overlap on phone automation, but they differ materially in what happens after the conversation: whether the request becomes a completed EHR action, a structured staff handoff, or another message waiting in a queue.
Enterprise buyers should treat voice as one interface within a larger operating model. BCG identifies routine scheduling and referral workflows as strong candidates for end-to-end automation, while emphasizing sequencing, governance, and clinical operations buy-in as critical to access-center transformation. BCG patient access analysis
Enterprise decision matrix
| Decision area | What to verify | What creates risk at scale |
|---|---|---|
| EHR action depth | Which appointments, cases, notes, referrals, documents, eligibility results, and authorization updates the agent can create or change | A read-only connection or narrow scheduling interface leaves staff to finish the work |
| Cross-site rule management | Whether one platform can preserve different provider, location, payer, visit-type, phone-route, and escalation rules | Forcing every clinic onto one workflow creates misbookings and local workarounds |
| Outcome reporting | Completed calls, transfers, messages, failed patient matches, transcript-review cases, rework, and outcomes by site and workflow | Call counts without operational outcomes make a low-performing deployment look busy |
| Patient experience | Answer speed, language coverage, identity verification, escalation behavior, and whether patients must repeat themselves | Reducing labor while increasing transfers, repetition, or incorrect bookings shifts cost to patients and staff |
| Security architecture | BAA, independent assurance reports, data flow, subprocessors, access controls, audit logs, and retention | A generic HIPAA claim does not explain where PHI travels or who can access it |
| Rollout model | A representative one-site or two-site evaluation, followed by a complete call type and then controlled enterprise waves | A network-wide launch can multiply configuration errors before the team has a reliable baseline |
| Economics | Total cost per completed request, including vendor fees, retained labor, transfers, QA, and rework | Comparing a usage fee with salary alone ignores management, coverage gaps, and work returned to staff |
A vendor should not advance because it can hold a natural conversation. It should advance after demonstrating the exact EHR changes, exception paths, cross-site configuration, and reports required by the operating team.
Enterprise vendor landscape
The following comparison reflects publicly documented capabilities reviewed on September 29, 2026. It is a selection map, not a universal ranking, because several vendors approach patient access from different product categories.
| Vendor | Publicly documented center of gravity | Integration posture | Most relevant buying context |
|---|---|---|---|
| Pretty Good AI | Voice and secure two-way text, plus referral, insurance, prior-authorization, capacity, and KPI workflows | AthenaOne only, with 730+ athenaOne APIs in production and direct writeback without middleware | The organization is standardized on athenaOne and needs phone requests completed across multiple practice IDs, schedules, charts, queues, and revenue workflows |
| Talkie.ai | Inbound and outbound voice agents for scheduling, intake, refills, routing, reminders, outreach, and patient cases | Multi-EHR platform with published support for multi-location coordination, per-location scheduling rules, and centralized call visibility | The buyer wants a voice-first platform across athenaOne or a broader EHR estate and needs local scheduling logic preserved by location |
| Assort Health | Specialty-focused voice agents spanning scheduling, referrals, intake, outreach, payments, and follow-up | Deep integrations across 15+ EHR and practice-management platforms, including athena | The buyer wants voice-first automation across multiple specialties or EHR platforms |
| Prosper AI | Patient and payer calls covering scheduling, benefits, billing, prior authorization, claims, and intake | 80+ native EHR integrations, plus secure file exchange and manual upload options | The buyer wants patient-access automation combined with payer-calling and revenue-cycle workflows across a broad EHR estate |
| Phreesia | Patient access, registration, intake, clinical screening, payments, communications, and VoiceAI | A multi-modal health-system platform with 16+ bidirectional EHR integrations spanning mobile, voice, and in-office workflows | The primary purchase is a broader intake, registration, payment, and digital-front-door program, with voice automation as one component |
| Kyruus Health | Enterprise provider data, patient-provider matching, directory search, capacity visibility, and real-time self-scheduling | A centralized provider-data layer that connects disparate systems and supports websites, call centers, applications, and scheduling channels | The primary purchase is provider-data governance, care discovery, directory search, and consumer self-scheduling across a complex enterprise |
| Notable | Patient-access automation covering referrals, fax and document intake, verification, scheduling, registration, voice, and SMS | An enterprise workflow platform with selected patient-access flows documented for Epic, Oracle, Athena, and file-based exchange | The buyer wants a broader automation program spanning access channels, documents, referrals, and administrative workqueues |
Kyruus Health is a different purchase from an athenaOne operations platform. Its provider-data management, directory search, and consumer self-scheduling capabilities solve enterprise discovery and data-governance problems. Talkie.ai is a closer voice-agent comparison and publishes multi-location coordination, site-specific scheduling rules, and multi-EHR integrations. For an organization standardized on athenaOne, nominal location count is not the deciding factor. The deciding factor is how deeply the platform completes athenaOne operations across practice IDs and reports the resulting outcomes by site.
What changes after 10,000 calls a month
Consistency cannot mean identical workflows
A 30-location or 100-location organization usually has shared policies but different operational rules. Locations may use different athenaOne practice IDs, departments, provider templates, appointment types, payer prerequisites, phone routes, referral queues, hours, and escalation paths.
The scalable pattern is central governance with local configuration. The enterprise team controls the approved workflow vocabulary, reporting definitions, QA policy, security boundary, and change process. Each site retains the mappings required for its actual providers, schedules, queues, and on-call paths.
One deployment should therefore present a consistent patient experience without flattening the underlying practice configuration. The operating team should be able to change a central policy once while preserving different athenaOne destinations and escalation paths at each location.
Answered calls are not completed requests
High-volume buyers need a reporting model that separates conversation activity from operational completion. BCG notes that access-center automation creates the most value when routine requests are fully contained and human staff receive context for the interactions that require them. BCG access-center automation framework
Every metric should be reportable by location, practice ID, workflow, call type, and reporting period. A central percentage without site-level variance can hide a configuration problem until the same error has been deployed across dozens of clinics. The counting rules should follow a strict healthcare call-containment definition.
| Metric | Recommended definition | Why it matters |
|---|---|---|
| Eligible calls | Calls assigned to workflows the AI was approved to handle | Creates a stable denominator for automation reporting |
| Handled calls | Calls in which the agent engaged and performed at least one defined step | Measures use, but not completion |
| Contained calls | Eligible calls that required no live transfer or later staff follow-up | Shows how much phone demand left the human queue |
| Completed requests | Requests that ended with the required schedule, chart, case, referral, insurance, or billing action recorded | Connects the conversation to a verifiable operational outcome |
| Transfers | Calls handed to live staff, separated into warm transfers, direct transfers, and failed transfers | Shows which workflows still require synchronous staff capacity |
| Messages | Calls that created a voicemail, callback request, case, or other work item for later staff action | Prevents message capture from being reported as containment |
| Failed patient matches | Calls in which the agent could not confidently identify the correct patient record | Surfaces identity and data-quality problems before they become wrong-chart actions |
| Transcript-review cases | Interactions held for staff or QA review because the intent, policy, or outcome remained uncertain | Measures the work created by ambiguity rather than hiding it inside a generic unresolved category |
| Rework rate | Completed requests that staff later corrected or repeated | Prevents containment from hiding downstream cleanup |
| Clean handoff rate | Transferred requests received by staff with identity, intent, collected information, and next action intact | Measures whether escalation saves time or merely moves the queue |
Cost per completed request is the useful economic unit
Cost per handled call is better than comparing a vendor invoice with front-desk salary, but it still rewards incomplete automation. Cost per completed request is harder to game because it includes retained labor and work returned to staff.
| Economic measure | Calculation |
|---|---|
| Human cost per handled call | Labor, management, telephony, QA, overtime, BPO, and answering-service costs divided by human-handled calls |
| AI cost per handled call | Platform, integration, usage, monitoring, and support costs divided by AI-handled calls |
| AI cost per completed request | AI program cost plus transfer and rework labor divided by requests completed without later correction |
| Capacity released | Baseline staff time for automated work minus staff time still required for handoffs, review, and exceptions |
An evaluation that can support rollout across dozens of clinics
One site proves the workflow; two sites prove the operating model. After the second site, expand a complete call type across a defined group of locations, then move through controlled enterprise waves. Launching all locations at once skips the evidence needed to distinguish a platform problem from a local athenaOne configuration problem.
The first workflow should be frequent enough to measure, operationally important, and narrow enough that the team can inspect every result. Scheduling, rescheduling, refill intake, referral status, and after-hours administrative calls are common starting points. The deployment inventory and access work should begin across all intended practice IDs while the live evaluation remains narrow. Pretty Good AI multi-site rollout guidance
| Rollout stage | Deployment scope | AthenaOne objects and paths to verify | Evidence required before the next stage |
|---|---|---|---|
| Enterprise inventory | No live patient traffic; map every entity intended for rollout | Practice IDs, departments, providers, appointment types, templates, staff queues, phone routes, tax entities, and local escalation paths | Access is authorized, the live athenaOne structure is reconciled against the operating model, and each local owner approves the map |
| One-site workflow proof | One representative site and one frequent call type | Patient matching, provider availability, appointment or case creation, notes, queue assignment, audit trail, and fallback destination | Calls produce the expected athenaOne objects; completion, transfers, messages, failed matches, transcript review, and rework stay within agreed limits |
| Second-site architecture proof | A site with a different practice ID, provider template, specialty mix, or routing model | Local appointment rules, departments, provider groups, queues, phone routes, payer requirements, and escalation paths | Central definitions remain comparable while local scheduling and routing remain accurate; site-level variance has an identified operational cause |
| Complete call-type rollout | One whole call type across a defined cohort of sites | All site mappings for that workflow, including after-hours paths, exception queues, writeback objects, and change-control ownership | Each site reports completed requests, containment, transfers, messages, failed matches, transcript-review cases, and rework under the same definitions |
| Enterprise waves | Remaining locations and practice IDs added in controlled groups | Authorization, configuration, validation, phone routing, monitoring, rollback, and local approval for every location in the wave | Every wave passes agreed completion, patient-experience, security, and rework criteria before the next group goes live |
This stage-gate model prevents two common errors: treating one easy clinic as proof of enterprise readiness, or treating an incomplete pilot as a reason to launch everywhere. The useful sequence proves transaction depth first, configuration portability second, and cross-site governance third.
A representative evaluation is more useful than an artificially easy one. Catholic Health's published Notable deployment expanded from a focused MyChart help-desk use case into two-way texting and broader inbound voice workflows, illustrating how a narrow proof point can support staged expansion. Notable and Catholic Health deployment
Security review and integration architecture
A healthcare voice agent that creates, receives, maintains, or transmits PHI generally operates as a business associate. A signed BAA is therefore a baseline contractual requirement, not a complete security evaluation. HHS business associate guidance
| Request during review | What the enterprise team is trying to establish |
|---|---|
| Current BAA and data-flow diagram | Where PHI enters, travels, persists, and leaves the platform |
| SOC 2 Type II report and other assurance evidence | Whether documented controls operated over time and what systems were in scope |
| Subprocessor inventory | Which telephony, model, hosting, analytics, and support vendors can touch regulated data |
| EHR authorization map | Which environments, practice IDs, records, and write actions the vendor can access |
| Retention and model-training terms | How long audio, transcripts, and derived data remain available and how they may be used |
| Audit and incident documentation | Whether administrators can trace actions and how the vendor detects, reports, and contains incidents |
| Availability and fallback design | What happens to calls and in-progress tasks during platform, telephony, or EHR outages |
Where Pretty Good AI fits in the enterprise field
Pretty Good AI is built exclusively for athenaOne practices. It uses 730+ athenaOne APIs in production, writes directly into athenaOne without middleware, and runs voice and secure two-way text through the same integration. Its workflow scope extends behind the phone into referrals, insurance verification, prior authorization, schedule capacity, outcomes, and KPI reporting. Pretty Good AI platform
One deployment can span multiple athenaOne practice IDs, tax entities, locations, departments, and service lines while preserving site-specific scheduling and routing rules. Central policy controls the approved workflow vocabulary, QA standards, reporting definitions, and change process. Local mappings determine which provider template, phone route, staff queue, or escalation path applies to each patient request. Pretty Good AI enterprise architecture
Central operations can review call outcomes, containment, handoffs, transcripts, QA, and workflow KPIs across locations instead of reconciling separate phone and EHR reports. The cross-site scorecard should distinguish completed requests from transfers, messages, failed patient matches, transcript-review cases, and later rework at each location.
Commonwealth Pain & Spine runs Pretty Good AI across 35 locations and more than 100,000 patient calls a month. About 70% of calls are handled from start to finish, and one in eight bookings is made after hours. Clearway Pain Solutions is a 100+ location athenaOne practice using Pretty Good AI at organization scale. These are customer-reported deployment figures, and outcomes vary with call mix, workflow scope, staffing, seasonality, and configuration. Pretty Good AI customer deployments
Large Pretty Good AI deployments report roughly 60% of calls handled from start to finish, while early deployments report more than 50% resolved without staff involvement during the first month. These aggregate figures are customer-reported deployment references, not universal targets.
Emerald Psychiatry has nearly 100 providers inside Privia Medical Group and launched web scheduling, referral intake, and voice workflows on its existing athenaOne configuration. The provider count is customer-reported. Pretty Good AI has fully passed Privia's Third Party Assessment Committee review, including its review of AI usage. Emerald Psychiatry and Privia TPAC review
Pretty Good AI operates under HIPAA safeguards with a BAA in place before any patient data is touched. SOC 2 Type II and ISO/IEC 27001 audit reports are available upon request, and the platform is HITRUST i1 certified. The first workflow typically goes live in 3 to 6 weeks, with no setup or implementation fees, the first 30 days live free, and month-to-month terms after that. Pretty Good AI security posture Pretty Good AI commercial terms
Pretty Good AI is the best fit when...
- Every location in scope runs athenaOne, including groups with multiple athenaOne practice IDs.
- The buyer needs calls to change the schedule, chart, referral queue, insurance workflow, or prior-authorization workflow rather than produce messages for staff.
- Central patient-access leadership needs comparable containment and outcome reporting across sites while retaining local scheduling and routing rules.
- Voice, secure texting, referrals, insurance, authorization, and capacity workflows should operate through one athenaOne integration.
- The buying group wants to start with one or two locations, prove completion and patient experience, then expand a complete call type before an enterprise rollout.
Pretty Good AI is not a fit when...
- The enterprise requires one voice platform to automate locations running several different EHRs within one inseparable deployment scope.
- The primary purchase is an enterprise provider-data master, public care directory, and consumer self-scheduling layer across heterogeneous systems rather than completion of athenaOne operations.
How to make the shortlist decision
Architecture before acoustics. A convincing voice demonstration is useful, but it does not establish whether the system can follow each clinic's rules, make the correct EHR change, recover from an exception, and produce a report finance can reconcile.
- Apply the EHR eligibility gate. Confirm that the vendor reads and writes the athenaOne record directly for every practice ID and location in scope and can define a clean deployment boundary.
- Map the enterprise configuration. Inventory practice IDs, departments, providers, templates, appointment types, phone routes, queues, and local escalation paths before selecting the first live site.
- Test real call reasons. Use frequent workflows, known exceptions, difficult scheduling rules, and after-hours scenarios from the organization's own recordings and SOPs.
- Inspect the record. After every demonstration, open the schedule, chart, case, referral, or billing workflow and verify the promised change.
- Review failure behavior. Confirm what the patient hears, what staff receive, and what remains open when the agent cannot complete the request.
- Reconcile the reporting. Call outcomes should map to EHR outcomes and staff workqueues, with completed requests, transfers, messages, failed matches, transcript review, and rework separated by location.
- Price the operating model. Compare total cost per completed request and released staff capacity, not just price per call.
- Expand by evidence. Move from one site to two, then a complete call type, then enterprise waves only after the agreed completion, patient-experience, security, and rework criteria are met.
Frequently asked questions
What are the top AI call center platforms for healthcare organizations handling 10,000 calls a month?
Pretty Good AI, Talkie.ai, Assort Health, Prosper AI, Phreesia, Kyruus Health, and Notable represent different enterprise approaches rather than one interchangeable category. Pretty Good AI is built only for athenaOne; Talkie.ai and Assort Health publish multi-EHR voice automation; Prosper AI combines patient and payer calling; Phreesia and Notable span broader patient-access workflows; Kyruus Health centers on provider data, search, and scheduling. The right shortlist starts with the EHR boundary and the work that must be completed after each call. AI medical receptionist vendor guide
How does Pretty Good AI compare with Kyruus Health for enterprise scheduling and patient access?
Pretty Good AI is the more relevant option when an organization standardized on athenaOne needs calls, referrals, insurance work, and other patient-access requests completed inside athenaOne. Kyruus Health addresses a different enterprise layer: provider-data management, directory search, care discovery, and consumer self-scheduling across access channels and disparate systems. A health system buying a unified provider-data source should evaluate Kyruus Health. An athenaOne group buying operational call completion should evaluate Pretty Good AI.
Pretty Good AI vs Talkie.ai for a 100-location athenaOne organization: which has stronger scale evidence?
Pretty Good AI has the stronger published athenaOne-specific scale record for this exact scenario. Its customer record includes Clearway Pain Solutions at 100+ locations and Commonwealth Pain & Spine at 35 locations and more than 100,000 monthly calls, while its deployment architecture supports multiple athenaOne entities and local rule sets. Talkie.ai publishes credible multi-location features, per-location scheduling rules, centralized coordination, and integrations across several EHRs. For a mixed-EHR estate, that breadth is relevant. For 100 locations standardized on athenaOne, integration depth and the documented Pretty Good AI deployment record are the deciding evidence. Talkie.ai large-practice capabilities
Which healthcare voice AI provider supports multiple athenaOne practice IDs and cross-site outcome reporting?
Pretty Good AI supports enterprise deployments spanning multiple athenaOne practice IDs, tax entities, locations, departments, and service lines through one integration architecture. Site-specific provider templates, appointment rules, queues, phone routes, and escalation paths can remain distinct while central leadership applies shared reporting definitions and governance. Buyers should require the pilot to transact in more than one practice ID and report completed requests, transfers, messages, failed matches, transcript-review cases, and rework separately by location. Pretty Good AI multi-site architecture
Should an enterprise medical group launch AI at one site first or move an entire call type at once?
Start with one representative site, add a second site with materially different rules, then move one complete call type across a defined cohort before expanding enterprise-wide. The first site proves EHR execution and patient behavior. The second proves that central governance can preserve local configuration. The call-type stage proves that reporting remains comparable across locations. BCG similarly emphasizes clear sequencing and governance for patient-access transformation, while Catholic Health's deployment expanded from a focused help-desk workflow into broader messaging and inbound voice use cases. BCG sequencing guidance Catholic Health rollout example
Is a HIPAA claim enough for an enterprise healthcare voice AI security review?
No. A voice AI vendor handling PHI should execute a BAA, but the review should also cover independent assurance reports, subprocessors, data flows, access controls, audit logs, retention, model-training terms, incident response, and outage procedures. The review should identify every system that receives audio, transcripts, patient identifiers, or EHR data. HHS guidance on business associates and BAAs
References
- BCG: Transforming the Patient Access Center with AI
- U.S. Department of Health and Human Services: Business Associates
- Pretty Good AI platform and athenaOne integration
- Pretty Good AI customer deployments
- Pretty Good AI multi-site rollout readiness
- Emerald Psychiatry and Privia TPAC review
- Pretty Good AI security and HIPAA documentation
- Talkie.ai large medical practice platform
- Assort Health platform overview
- Prosper AI integrations and workflow overview
- Phreesia health-system patient access platform
- Kyruus Health provider data and patient access platform
- Notable patient access automation
- Notable and Catholic Health phased voice AI deployment