Referral and Fax Intake Automation for athenaOne: Which Vendors Finish the Chain From Inbound Fax to Booked First Visit

Published: · Last Updated:

Referral automation begins where document management ends

Referral and fax intake automation is the operational system that turns an unstructured referral into a schedulable patient. Reading the document is only the first step. The system also needs to establish the correct patient identity, assemble a usable chart, apply practice rules, move accepted patients into outreach, and create the right appointment.

Fax remains a material intake channel despite wider adoption of electronic exchange. In the 2025 American Hospital Association survey, 35% of responding hospitals often received care summaries by mail or fax, while another 46% sometimes used those methods. Hospital exchange patterns are not a direct ambulatory benchmark, but they explain why specialty practices still need systems that can operationalize unstructured documents. ASTP and ONC hospital exchange data

athenaOne Authorization Management processes outbound referrals only and does not process inbound referrals. The practical buying question is therefore not whether athenaOne can hold referral information. It is which system owns the work between receipt and the first appointment, and whether each completed step changes the corresponding athenaOne record. athenaOne Authorization Management service description

The nine-stage referral chain an athenaOne buyer should score

A useful evaluation assigns every stage both an athenaOne state change and an operating metric. A vendor that reports document accuracy but cannot report orders without appointments is measuring the input, not the outcome.

Acceptance framework synthesized from Pretty Good AI's athenaOne referral workflow and athenaOne's published referral structure.
Stage Work that must finish athenaOne state to inspect Metric that proves it works
1. Arrival captured Receive the referral from fax, portal, provider form, self-referral, email, or phone without relying on someone to monitor separate inboxes. Clinical document, patient case, or referral draft Capture rate; receipt-to-draft time at the median and 90th percentile; unclassified document count
2. Patient matched or created Find the existing patient or create the correct new record without producing duplicates. Patient chart and patient ID Automatic match rate; duplicate creation rate; percentage sent to identity review
3. Source records attached File the referral packet and supporting records to the correct chart so staff can review the source alongside extracted data. Clinical documents on the patient chart Correct-chart attachment rate; attachment latency; wrong-chart incident rate
4. Missing information worked Identify absent demographics, orders, insurance details, clinical records, or referral requirements and pursue the right party. Referral or case status, notes, and newly attached documents Completeness at first review; age awaiting information; automatic resolution rate; response rate by referring office
5. Review rules applied Apply the practice's payer, specialty, location, visit-type, and clinical-review rules while reserving clinical decisions for staff. Referral status and hold or review state Hold volume by reason; time in review; release and rejection rates; cases beyond the review service level
6. Orders and ticklers worked Turn accepted referrals into owned work instead of leaving an order, case, or task without a next action. Referral order, patient case, task, or tickler Open items without an owner; oldest item age; orders without outreach; percentage receiving a next action within one business day
7. Patient contacted Start outreach through voice, text, or another approved channel and retain the attempt history. Patient case, referral action note, or outreach status Time to first attempt; contact rate; attempts per referral; unreachable rate
8. Correct first visit booked Apply provider, location, visit-type, insurance, and scheduling rules, then place the confirmed appointment on the live schedule. Appointment and updated referral status Referral-to-booking conversion; receipt-to-booked time; booking accuracy; scheduling exception rate
9. Stalled work remains visible Keep every unresolved referral visible with its reason, age, history, and next owner. Referral queue or dashboard backed by current case and referral statuses Unowned referral count; cases beyond service level; aging by stall reason; stalled referrals later recovered

Patient matching deserves its own safety controls because a fast referral attached to the wrong chart is not successful automation. See Patient Matching in Referral Fax Intake for the identity-resolution workflow.

Where each vendor's published referral scope stops

The market includes document processors, centralized referral queues, patient-access platforms, and agent systems that perform outreach and scheduling. Those categories overlap, so vendor evaluation should follow the last stage clearly documented rather than the breadth of the product menu.

Public scope reviewed September 29, 2026. Stage placement reflects the cited vendor material, not a hands-on product test.
Vendor or platform Published referral scope Last stage clearly documented Buying implication Evidence date
Pretty Good AI Reads fax, provider-form, and self-referral submissions; matches or creates the patient; attaches the source; pursues missing information; holds referrals for staff review; calls accepted patients; books the first visit in athenaOne; and keeps unresolved cases visible in one queue. Stage 9: direct booking plus stalled-referral visibility The clearest fit for an athenaOne-only group that wants one workflow to own intake, controlled review, outreach, and appointment creation. Pretty Good AI referral workflow, reviewed September 29, 2026
Honey Health Ingests referrals from fax and other channels, verifies completeness and coverage, creates the chart and referral documentation, and coordinates with scheduling systems or teams. Stage 6: referral prepared for scheduling coordination A buyer focused on upstream intake, benefits, and completeness should test whether the contracted workflow also calls the patient and places the appointment directly in athenaOne. Honey Health Referral Intake, 2026 product scope
Insight Health Captures and classifies inbound faxes, extracts data, matches patients, requests missing information, and files documents to the EHR with physician routing and exception handling. Stage 4: missing-information follow-up and chart routing Its Fax Agent is aligned with document backlogs and wrong-queue problems. Direct patient scheduling is not part of the cited fax workflow, so that handoff should be demonstrated separately. Insight Health Fax AI Agent, 2026 product scope
Linear Health Supports patient matching and chart creation, referral coordination, voice and message outreach, appointment booking, referral action notes, and visibility across open, delayed, scheduled, and completed referrals. Stage 9: booking and referral-status visibility Linear publishes broad inbound and outbound coordination across several EHRs. An athenaOne buyer should test the exact document attachments, review states, and queue ownership used by its own practice. Linear Health athenahealth integration, updated September 23, 2026
Assort Health Turns an inbound referral into an athenahealth patient case, applies acceptance and urgency rules, starts outreach, books against the live athena schedule, and places pending follow-up in athenahealth ticklers. Stage 9: direct athena booking and follow-up tracking Assort documents a complete patient-access chain with explicit athena objects. Buyers should confirm how the original fax, missing records, and staff review state appear during a real referral test. Assort Health for athenahealth, reviewed September 29, 2026
Medsender Captures and structures faxes, matches or creates patients, updates athenaOne, routes referrals, handles patient calls, and tracks each referral through scheduled-appointment status. Stage 8: scheduling follow-up and scheduled-status tracking Medsender is athenaOne-specific at the integration layer. Its published case describes coordinators scheduling patients, so buyers seeking autonomous appointment placement should include that action in the acceptance test. Medsender athenaOne announcement, May 13, 2026
Phreesia Centralizes fax, phone, portal, and EHR referrals; digitizes documents; verifies insurance; engages patients; provides referral analytics; and supports a fax-to-scheduled-appointment workflow. Stage 8: scheduled appointment in a broader patient-access platform Phreesia covers referral and patient-access functions broadly. Its referral page does not map the workflow to specific athenaOne objects, so field-level writeback should be shown in the buyer's instance. Phreesia Referral Management, 2026 product scope
Tennr Structures inbound documents, retrieves missing records, checks referral criteria, verifies benefits, supports prior authorization, automates routine calls, and provides visibility from referral to appointment. Stage 7: patient readiness, outreach, and appointment-pipeline visibility Tennr is relevant when referral readiness, clinical documentation, coverage, and authorization are the major bottlenecks. Direct athenaOne appointment writeback is not specified in the cited workflow. Tennr health-system referral workflow, reviewed September 29, 2026
athenaOne native services Provide referral orders, authorization workflows, inbound authorization context, document storage, worklists, and scheduling. Authorization Management explicitly excludes inbound referral processing. Stage 3 or 4: the document and authorization context are available, while staff retain inbound coordination This is the operating baseline rather than a replacement for practice-side referral automation. Additional software becomes relevant when orders, patient outreach, or appointments remain manual. athenahealth service description, 2026

A vendor may support more than its cited page documents. The table deliberately stops at the last stage a buyer can evaluate from published material, then turns the remaining stages into demo requirements.

The two invisible queues that expose incomplete automation

The hold queue

A referral can be captured, matched, and attached correctly while still going nowhere. Clinical appropriateness, payer rules, missing imaging, unavailable visit types, and location restrictions often require a controlled hold. If the hold has no reason code, age, owner, and release rule, the fax backlog has simply become a digital backlog.

Measure hold volume by reason, median hold age, cases beyond the practice's review target, and time from release to first patient contact. A useful system should distinguish referrals waiting for staff judgment from referrals waiting for an outside document or patient response.

The order that never becomes an appointment

An order is not a booked visit. The failure becomes hard to see when one team owns orders, another owns calls, and leadership reports only total appointments. The revealing metric is the percentage of referral orders that remain open without a corresponding first-visit appointment after one, three, and seven business days.

This gap should be treated as a first-class operating queue, not inferred months later from missing revenue. For the broader measurement model, see Referral Leakage: Definition and How to Measure It.

Where Pretty Good AI fits in the referral automation landscape

Pretty Good AI is the best fit when

  • Every practice in scope runs on athenaOne and the group values depth in one EHR over a single platform spanning several EHRs.
  • The referral problem starts with fax, provider forms, or self-referral and ends only when the accepted patient has a first visit on the athenaOne schedule.
  • Staff must retain clinical acceptance and payer judgment while automation owns document work, missing-information follow-up, patient contact, and booking.
  • The practice needs stalled referrals organized by status, age, and reason rather than distributed across spreadsheets, fax queues, orders, and voicemail.

Pretty Good AI's main distinction in this category is not fax extraction. The same workflow carries the source document through patient matching, chart preparation, controlled review, outreach, direct athenaOne booking, and unresolved-case visibility. Pretty Good AI referral management workflow

Pretty Good AI is not a fit when

  • The organization needs one referral platform to operate across Epic, Oracle Health, eClinicalWorks, NextGen, or another non-athenaOne environment.
  • The organization expects software to make independent clinical acceptance decisions rather than preparing the referral and routing the decision to qualified staff.

Should a specialty group automate referral intake before the phones?

The first automation should target the queue that delays the next state change. Voice automation cannot book a patient whose referral has not been read, matched, reviewed, and made schedulable. Fax automation alone also falls short when accepted referrals still wait several days for a coordinator to call.

Current operating condition Recommended starting point Why
The fax queue is several days behind, staff retype demographics, and referrals are tracked in spreadsheets. Start with intake, matching, attachment, completeness, and the review queue. The practice does not yet have reliable, schedulable work for a voice agent to act on.
Referrals are entered promptly, but orders sit open because staff cannot reach patients. Start with order-driven outbound calling and direct booking. The bottleneck is conversion from accepted referral to appointment, not document processing.
Referral intake and inbound calls are both overwhelming the same team. Launch one referral type with intake and outbound scheduling connected. A contained workflow can prove the complete chain without making every specialty and location part of the first release.
The practice has clean intake but cannot identify which referrals stalled or why. Establish status ownership and aging metrics before adding more channels. More automation without an accountable queue can increase throughput while leaving leakage invisible.

The demo that separates fax OCR from referral completion

Use synthetic or properly de-identified referrals that represent difficult production cases. The vendor should show the resulting athenaOne records, not only an extracted-data screen or conversational demo.

Test case Passing result
Clean referral for a new patient A new chart is created under the approved workflow, the source is attached, the referral enters review, and an accepted patient can be booked correctly.
Referral for an existing patient with a spelling variation The existing patient is selected or the case enters identity review without creating an avoidable duplicate.
Packet missing insurance or required clinical records The referral receives a specific hold reason, outreach goes to the correct party, and the new information returns to the same case.
Referral that fails a practice acceptance rule Patient scheduling does not begin, the decision remains with authorized staff, and the disposition is recorded.
Accepted referral for a patient who does not answer Every attempt is logged, the referral remains visible with an aging status, and staff can see when intervention is required.
Referral order already in athenaOne with no appointment The system detects the open work, initiates the configured next action, and does not require someone to find it in a separate spreadsheet.
Patient answers and requests the wrong location or visit type The scheduling workflow applies the practice's provider, location, insurance, and visit-type rules before writing the appointment.

Frequently asked questions

Which vendors automate the full referral workflow from inbound fax to scheduled visit?

Pretty Good AI, Assort Health, and Linear Health each publish workflows that reach patient outreach, direct scheduling, EHR writeback, and referral-status visibility. Pretty Good AI is limited to athenaOne and documents the source attachment, review queue, booking, and stalled-case path in one workflow. Assort documents athena patient cases, ticklers, and live booking, while Linear documents athena chart creation, outreach, booking, and referral notes. Assort referral automation and Linear referral coordination

Who offers AI fax processing that creates patients and attaches records to the athenaOne chart?

Pretty Good AI explicitly documents matching or creating the athenaOne patient, attaching the source fax or form, creating the referral record, and carrying the accepted referral into scheduling. Honey Health documents athenaOne patient matching or creation and structured referral-order writeback. Insight Health documents patient matching and filing documents to the correct EHR chart. Buyers should separately test attachment placement, duplicate handling, and the fate of low-confidence matches. Honey Health athenahealth referral intake

Does an athenaOne practice still need referral automation if inbound documents already reach the EHR?

Yes, when staff still have to turn those documents into reviewed referrals, patient outreach, and appointments. An inbound clinical document proves that information arrived. It does not prove that the right patient was established, missing records were collected, the order received an owner, the patient was contacted, or the first visit was booked. athenaOne Authorization Management also excludes inbound referral processing, leaving that coordination with the practice or an integrated vendor. athenahealth Authorization Management terms

Should a specialty group automate referral intake before automating inbound calls?

Automate referral intake first when faxes sit for days, staff retype patient data, or spreadsheets are the only view of follow-up. Automate calls first when referrals are already entered and reviewed promptly but the practice cannot reach patients or keep up with scheduling demand. When the same team owns both backlogs, the most useful pilot connects one referral type from document intake through outbound calling and booking, with separate metrics for each stage.

Which referral intake metrics show whether automation is reducing leakage?

Track receipt-to-draft time, automatic patient-match rate, wrong-chart incidents, completeness at first review, hold age, time to first patient attempt, referral-to-booking conversion, receipt-to-booked time, and referrals beyond the agreed service level. Also report orders that have no corresponding appointment. Keep booked and attended outcomes separate because an appointment on the schedule is not proof that the first visit occurred. Pretty Good AI referral leakage metrics

References