Transcribe specialist referral calls and auto-populate referral forms
This solution automatically listens to specialist referral calls and instantly fills out the required paperwork in the practice's system. It gives MSPs a high-value offering to help clinics eliminate costly data entry errors and stop patient leakage.
The problem today
30 mins
wasted on manual data entry per referral call
$10K+
lost per month in wasted clinical staff time
Maria Castillo is the office manager for a seven-provider orthopedic group outside Columbus, Ohio. Her biggest daily frustration is watching her two best front desk employees spend their afternoons doing data entry instead of talking to patients — and knowing that at least one referral per week is getting rejected because someone misheard a number on a call.
01The Problem
Two front desk employees lose entire afternoons to data entry while patients at the desk go unattended.
A single transposed digit returns the referral days later, forcing staff to rebuild from scratch while the patient waits.
Slow submission hands volume to whichever competitor processes faster — no marketing budget required on their end.
Reconstructing calls from memory and scribbled notes guarantees missing clinical detail and compounds downstream rejection risk.
Staff hired for care coordination burn expensive hours on keyboard work that requires no clinical judgment and generates no revenue.
No record ties what was said on the call to what landed on the form, so disputed referrals cannot be defended.
02The Solution
Solution Brief
Fictional portrayal · illustrative
- Maria's two front desk staff field 11+ referral calls before noon
- Insurance IDs half-heard, diagnosis codes guessed, sticky notes lost
- One rejection per week minimum; one patient already booked elsewhere
- Each rejection costs a full rebuild plus resubmission cycle
- One rejection/week compounds to several full workdays of rework per month
- Clinical staff consumed by keyboard tasks that build no patient relationships
- No call-to-form audit trail when a referral goes wrong
- Call ends; system transcribes, extracts, and pre-populates EHR referral form
- 30-minute manual process collapses to 90-second confirmation
- Referrals submitted same day; rejection rate drops
- Touches phone system, EHR, and daily workflow — replacement friction is high
- $150–$200/user/month against ~$70 vendor cost; single practice funds itself inside year one
“I used to find my staff at 4:30 still typing from calls they took at 10 in the morning. Now the form is basically done by the time they hang up. I didn't realize how much of our day we were just handing to a keyboard.”
— Maria Castillo is the office manager for a seven-provider orthopedic group outside Columbus, Ohio
03What the AI Actually Does
Ambient Call Transcription
Captures every referral call in real time through a HIPAA-compliant VoIP system, producing a full verbatim transcript with each speaker's words clearly separated — no manual recording, no missed details, no relying on staff memory.
Referral Data Extraction Engine
Reads the call transcript and automatically identifies and structures the key referral fields — patient demographics, insurance information, referring diagnosis, urgency level, and clinical notes — ready to be written into the EHR without anyone retyping a word.
EHR Auto-Population Agent
Pushes the extracted referral data directly into the practice's existing EHR or dental practice management system via standard healthcare data interfaces, so the form arrives pre-filled and staff only need to review and confirm before sending.
Compliance & Consent Manager
Handles call recording consent announcements automatically, enforces role-based access to call transcripts and PHI, and maintains a complete audit trail linking every submitted referral form back to the original call — critical protection if a referral is ever disputed.
04Technology Stack
Dialpad AI (HIPAA-Compliant Business Plan)
$25/user/month (HIPAA plan with BAA); estimate 6 users = $150/month
HIPAA-compliant cloud VoIP phone system with built-in AI transcription and call recording. Serves as the primary call capture layer: records all refer…
HealOS AI Scribe + Referrals Agent
$49/provider/month (Premium tier with Referrals Agent); estimate 5 providers = $245/month. Free tier available for initial proof-of-concept (20 sessions/month).
Core AI platform for the referral workflow. The Scribe module transcribes referral call audio with speaker diarization. The Referrals Agent extracts s…
OraCore Scribe (Dental Practices Only)
$149/month for unlimited providers per practice location
Dental-specific ambient AI scribe that integrates with virtually all major dental PMS systems including Dentrix, Open Dental, and Eaglesoft. Replaces …
Microsoft Azure OpenAI Service (Whisper + GPT-5.4)
Whisper transcription: $0.006/minute; GPT-5.4 extraction: ~$0.01–$0.03 per referral call processing. Estimated $30–$80/month for a 5-provider practice handling 200–500 referral calls/month.
HIPAA-compliant (BAA available through Microsoft Enterprise Agreement) AI backend for custom NLP extraction pipeline. Used when HealOS Referrals Agent…
Microsoft 365 Business Premium (or existing license)
$22/user/month (if not already licensed)
Provides Azure AD / Entra ID for SSO and MFA across all platforms (Dialpad, HealOS, EHR), Teams for internal communication, and SharePoint/OneDrive fo…
Compliancy Group (The Guard)
$299–$499/month depending on practice size
HIPAA compliance management platform for tracking BAAs, conducting risk assessments, managing consent documentation, audit logging, and breach notific…
05Alternative Approaches
Microsoft Dragon Copilot (DAX)
$600/provider/month
Enterprise-grade ambient AI platform with deep Epic EHR integration and comprehensive clinical documentation capabilities beyond referrals, including progress notes, H&Ps, and discharge summaries.
Strengths
- Deepest EHR integration, especially with Epic
- Comprehensive clinical documentation beyond referrals
- Dedicated implementation support teams
- Single ambient AI platform for all documentation needs
Tradeoffs
- 4–12x higher per-provider monthly cost ($600 vs. $49–$75)
- Requires significant IT infrastructure
- 3–6 month deployment timeline
- Referral capture requires custom workflow configuration
Best for: Practices with 10+ providers already on Epic or athenahealth who want a single ambient AI platform for all documentation needs, not just referrals
Abridge
$400–$600/provider/month
Enterprise-grade ambient AI scribe with broad clinical documentation capabilities and dedicated implementation support. Handles referral capture through general ambient recording with custom workflow configuration.
Strengths
- Comprehensive clinical documentation capabilities
- Dedicated implementation support
- Broad EHR integration
Tradeoffs
- 4–12x higher per-provider monthly cost ($400–$600 vs. $49–$75)
- Referral capture requires custom workflow configuration
- Not recommended for practices that only need referral workflow automation
Best for: Practices wanting a single ambient AI platform for all documentation needs including referrals, progress notes, H&Ps, and discharge summaries
Suki
$399/provider/month
Enterprise-grade ambient AI scribe platform offering comprehensive clinical documentation capabilities and dedicated implementation support. Handles referral capture through general ambient recording.
Strengths
- Comprehensive clinical documentation beyond referrals
- Dedicated implementation support
- Competitive enterprise pricing relative to DAX
Tradeoffs
- 4–12x higher per-provider monthly cost ($399 vs. $49–$75)
- Referral capture requires custom workflow configuration for referral-specific data extraction
- Not recommended for practices that only need referral workflow automation
Best for: Practices with 10+ providers wanting a single ambient AI platform for all documentation needs, not just referrals
Freed AI
$39–$99/provider/month
Simplest, fastest-to-deploy ambient AI scribe option for practices without IT departments. Provides call transcription with export capability but lacks a dedicated referral agent for auto-populating referral forms.
Strengths
- Easier to deploy than HealOS — no EHR integration required for transcription
- Similar price point to HealOS ($39–$99 vs. $49)
- Well-suited for solo or 2-provider practices focused on transcription
Tradeoffs
- No dedicated referral agent — core value proposition of auto-populating referral forms is lost
- Referral form auto-population requires manual copy-paste or custom Azure OpenAI middleware
- Not recommended if the primary goal is eliminating manual referral form data entry
Best for: Solo practitioners or 2-provider practices that primarily want transcription of referral calls for documentation purposes, with manual form population acceptable
HealOS AI Scribe + Referrals Agent
$49/provider/month
Core AI platform with dedicated Referrals Agent for structured data extraction and auto-population of referral forms in the EHR. Supports API/HL7/FHIR integrations and payer referral requirement verification.
Strengths
- Dedicated referral agent for auto-populating referral forms
- EHR write-back via API/HL7/FHIR
- Payer referral requirement and network rule verification
- Free tier available for proof-of-concept
Tradeoffs
- Requires EHR integration setup
- Slightly more complex deployment than Freed
Best for: Practices where the primary goal is eliminating manual referral form data entry and automating the full referral workflow
Custom Azure OpenAI Pipeline (Build-Your-Own)
$0.01–$0.05 per referral call ($30–$80/month Azure consumption for 200–500 calls/month)
Skip HealOS and Dialpad's built-in AI entirely. Build a custom pipeline using Azure OpenAI Whisper for transcription and GPT-5.4 for entity extraction, with custom Azure Functions middleware connecting the practice's existing phone system to the EHR via FHIR APIs.
Strengths
- Lowest per-call AI cost ($0.01–$0.05 per referral call vs. $49+/provider/month)
- Highest MSP margin potential — charge $150–$200/provider/month for a service costing $30–$80/month in Azure consumption
- Maximum control over pipeline, customization, and accuracy
- Works with EHR systems not supported by HealOS or other turnkey platforms
Tradeoffs
- Highest complexity — requires Python development, Azure DevOps, FHIR API expertise, and ongoing prompt engineering
- 8–16 week deployment timeline vs. 4–6 weeks for turnkey
- MSP owns entire pipeline including uptime, accuracy, and compliance
- No vendor support to fall back on
Best for: MSPs with in-house developers seeking maximum control and margin, or practices with EHR systems not supported by HealOS or other turnkey platforms
HealOS + Dialpad Turnkey Stack
$49/provider/month (HealOS) + $25/user/month (Dialpad)
Deploy HealOS Referrals Agent with Dialpad AI as the primary call capture layer. Turnkey deployment with vendor support, built-in HIPAA compliance, and EHR integrations.
Strengths
- 4–6 week deployment timeline
- Vendor support and uptime responsibility
- Built-in HIPAA compliance and BAA
- Pre-built EHR integrations via API/HL7/FHIR
Tradeoffs
- Higher per-provider monthly cost than build-your-own
- Less customization capability
- Lower MSP margin potential
- May not support all EHR systems
Best for: MSPs without in-house developers, or practices needing fast deployment with minimal technical risk
Bola AI
Contact vendor (estimated $100–$300/month)
Purpose-built for dental workflows with deep Dentrix and Open Dental integration. Offers dental-specific NLP that natively understands CDT codes, tooth numbering systems, and dental specialty referral workflows.
Strengths
- More dental-specific NLP than OraCore for complex referral needs
- Deep Dentrix and Open Dental integration
- Native understanding of CDT codes and tooth numbering systems
- Purpose-built for dental specialty referrals (implant, TMJ, sleep dentistry)
Tradeoffs
- Pricing unknown — requires direct vendor contact (estimated $100–$300/month)
- Newer vendor with less market validation and fewer reviews/case studies
- Not recommended if the practice also has medical providers
Best for: Dental practices with high-volume specialist referrals, especially oral surgery and periodontics, where dental-specific NLP accuracy is critical
OraCore Scribe
$149/month for unlimited providers per practice location
Dental-specific ambient AI scribe integrating with all major dental PMS systems including Dentrix, Open Dental, and Eaglesoft. Multi-speaker capture for hygienist, assistant, and provider.
Strengths
- Flat monthly rate with unlimited providers per location
- Integrates with virtually all major dental PMS systems
- No special microphone required
- More market validation than Bola AI
- Supports mixed medical/dental practices when combined with HealOS
Tradeoffs
- Less dental-specialty-specific NLP than Bola AI for complex referral workflows
- May not natively understand CDT codes and tooth numbering as deeply as Bola AI
Best for: Dental practices needing straightforward PMS integration across Dentrix, Open Dental, or Eaglesoft, or mixed medical/dental practices
RingCentral RingEX
$20–$35/user/month
Established HIPAA-compliant VoIP platform with AI transcription, call recording, BAA availability, and a broad app marketplace. Larger market share in healthcare than Dialpad.
Strengths
- Larger, more established vendor — lower risk for practices concerned about vendor stability
- Broader app marketplace
- Advanced call center features (IVR, call queuing, skills-based routing)
- Good fit for practices already using RingCentral
Tradeoffs
- Slightly higher cost ($20–$35/user/month vs. $15–$25/user/month for Dialpad)
- Dialpad's native AI transcription may be more tightly integrated than RingCentral's
Best for: Practices that already use RingCentral, prefer a larger vendor, or need advanced call center features for their referral lines
Dialpad AI (HIPAA-Compliant Business Plan)
$25/user/month (HIPAA plan with BAA)
HIPAA-compliant cloud VoIP with tightly integrated native AI transcription, real-time call recording, and API for downstream AI processing. Primary recommendation for new deployments.
Strengths
- Tightly integrated native AI transcription
- Real-time transcripts and API export to downstream AI processing
- Automatic call recording consent announcements
- Slightly lower cost than RingCentral
Tradeoffs
- Smaller vendor than RingCentral — higher vendor stability risk perception
- Smaller app marketplace than RingCentral
Best for: New deployments where native AI transcription integration and API access are the primary requirements
Ready to build this?