Product feature
EMR Integration: Connecting AI Reception to Your Practice’s System of Record
Before any other question, buyers ask the same thing: does this actually work with our EMR? This page answers that. We’ve built a tiered integration architecture across 23 supported systems, so your AI receptionist reads live availability, writes confirmed appointments, and pushes patient demographics back to the same chart your clinicians open every morning. No swivel chair. No spreadsheet exports. No staff retyping anything.
EMRs Supported
23
Production integrations live across three tiers
Tier 1 Deep API
14
Full bidirectional sync EMRs (athena, AdvancedMD, Tebra and more)
Write Success Rate
99.7%
Appointments written successfully on first attempt
Avg Onboarding
4 weeks
From contract to first AI booked appointment
Integration architecture
Three Tiers, Set by What Your EMR Vendor Actually Exposes
Not every EMR ships a modern REST API. Some publish robust FHIR endpoints. Some only allow vendor partners. A handful are still Citrix only. We probe each vendor for the deepest available path, then place the integration in one of three tiers. You always know what you’re getting before you sign.
Tier 1: Deep API
14 EMRs, full bidirectional sync
- • Real time schedule reads via vendor REST or FHIR
- • Direct appointment writes with confirmation IDs
- • Patient demographic create and update
- • Insurance and copay capture on the booking record
- • eRx triage routing into the correct provider inbox
- • Webhook subscriptions for cancellations and rebooks
Includes: AdvancedMD, athenahealth, Charm, Jane App, ModMed family, Tebra, NextGen, eClinicalWorks, DrChrono, Kareo, PracticeFusion, OncoEMR, Office Practicum, Greenway.
Tier 2: Hybrid
9 EMRs, API reads plus automation writes
- • Read appointment availability via published API
- • Write appointments via headed browser automation
- • Demographic writes through validated UI flows
- • Session keepers maintain authenticated state on EC2
- • Auto relogin with 24 hour cooldown on credential failure
- • Vision fallback for screens that lack stable selectors
Includes: Office Ally, Accuro, Med Access, ChiroHD, Weave, Practice Mate, PMD, ModMed Ortho, ModMed Podiatry.
Tier 3: Automation Only
Legacy systems, browser based end to end
- • Both reads and writes through automated UI flows
- • Suited to EMRs with no public API at all
- • Same monitoring and alerting as Tier 1 and 2
- • Centralized alert manager covers broken sessions
- • Citrix HTML5 supported via vision plus xdotool
- • Custom built when a client’s EMR isn’t on our list
Used for niche or self hosted legacy systems. Onboarding adds two to four weeks of selector mapping.
Capability matrix
What the AI Can Actually Do, Tier by Tier
Capabilities aren’t binary. A Tier 2 integration may handle 90 percent of what Tier 1 does, just by a slower path. Here’s the honest breakdown of what ships on day one for each tier.
| Capability | Tier 1 Deep API | Tier 2 Hybrid | Tier 3 Automation |
|---|---|---|---|
| Read live schedule availability | Yes, sub second | Yes, sub second | Yes, 2 to 5 seconds |
| Write confirmed appointments | Yes, with confirmation ID | Yes, with post save verify | Yes, with post save verify |
| Write patient demographics | Yes | Yes | Yes, slower |
| Push intake form data | Yes | Yes | Partial, varies by EMR |
| eRx triage routing | Yes | Read only, route in dashboard | Dashboard only |
| Billing and copay capture | Yes, on appointment record | Yes, on appointment record | Captured separately |
| Webhook driven cancellation sync | Yes | Polling, 5 minute | Polling, 15 minute |
Buyer FAQ
The Five Questions Every Practice Owner Asks
We’ve fielded these in hundreds of evaluation calls. Here are the direct answers we give before the demo so you can decide if we’re even worth your forty five minutes.
Does the AI need EMR credentials?
For Tier 1 integrations, no. We use vendor sanctioned API keys or OAuth scopes. For Tier 2 and Tier 3 we use a dedicated service user that you create inside your EMR, with the minimum permissions required to read your schedule and write appointments. Credentials live encrypted in AWS Secrets Manager, never in the AI prompt, never on a laptop.
What fields actually get written?
Patient first and last name, date of birth, phone, email, primary insurance, reason for visit, preferred provider, appointment date and time, and a note that the booking originated from the AI receptionist with a call recording link. You can add or remove fields per location during onboarding. We never write clinical findings or diagnoses, only scheduling and demographic data.
Can we test in a sandbox first?
Yes. Every Tier 1 integration runs in your vendor’s sandbox during the first week of onboarding. For Tier 2 and Tier 3 we mirror your live EMR into a test tenant when the vendor supports it, otherwise we create test patients and test appointment types so production data is never touched until you sign off. We require sandbox sign off before flipping the AI to production.
What happens if the EMR API goes down?
The AI keeps answering. New appointments queue in our durable write buffer and replay as soon as the EMR responds. Patients receive a real confirmation either immediately or via SMS once the write completes. You get a Slack alert from our centralized alert manager the moment writes start failing, with the EMR vendor status checked automatically before any human is paged.
What’s the PHI scope?
Minimum necessary. We process the patient’s identifying details and reason for visit only for the duration of the call, then store a redacted call summary and the structured booking payload. A signed BAA is required before any production data flows. PHI is encrypted at rest and in transit, and your call recordings live in an S3 bucket you can audit. Full details are in our compliance overview.
Onboarding flow
From Signed Contract to AI Booked Appointment in Four Weeks
Week 1: Connection Wizard
You enter EMR vendor and tenant ID. We auto detect the right tier, request sandbox access, and provision the service user. Most clients finish this in under an hour.
Week 2: Sandbox Test
The AI receptionist runs in sandbox with your real scheduling rules. We book one hundred test appointments end to end and you review them before approving the writes.
Week 3: Production Cutover
We forward your phone number, flip the AI to production, and keep your front desk in the loop on a shared dashboard. Real bookings flow into your live EMR.
Week 4: Monitoring Live
Sentinel alarms watch every Lambda. The centralized alert manager pings Slack on any broken EMR session. You see write success rate, booking volume, and recovered revenue in your portal.
Supported EMRs
23 Production Integrations, Each With Its Own Hub Page
Each EMR hub documents the exact API endpoints we use, the field mapping, the sandbox URL, and any known vendor quirks. If your EMR isn’t listed, ask and we’ll scope a custom integration.
For technical buyers
Want the Engineering Detail?
Our EHR integration technical landing covers the deeper architecture: REST probe order, vision fallback, session keeper EC2 design, and the centralized alert manager state machine. Built for buyers whose IT lead is in the call.
EHR Integration Architecture →Ready to evaluate
See It Live on Your EMR in a 30 Minute Demo
We’ll log into our sandbox of your EMR and book a real appointment, with you watching. No slideware, no canned video, just the live integration handling whatever scheduling rules you throw at it.
Book a Live Demo →Where to go next