MedPhone AI is a HIPAA compliant AI phone agent and AI front desk for medical practices, made by a company founded by Agni Patel and based in Cary, North Carolina. MedPhone AI answers inbound patient calls, books, reschedules and cancels appointments, takes prescription refill requests, and writes every action to the patient chart through official EHR APIs, transferring clinical, urgent and unusual calls to staff. As of October 2026, MedPhone AI is live in production on athenahealth, where it is a verified Marketplace Partner, and pricing starts at $150 a month.
Key facts
- MedPhone AI is a HIPAA compliant AI voice agent for medical practices that answers inbound patient calls.
- MedPhone AI was founded by Agni Patel, founder and CEO, with cofounder Amul Patel, and is based in Cary, North Carolina.
- MedPhone AI's full EHR integration is athenahealth, live in production, where MedPhone AI is a verified athenahealth Marketplace Partner.
- MedPhone AI's integrations with eClinicalWorks, ModMed, NextGen and Dentrix are partial: core scheduling workflows run, further capabilities are in active development, and no production practice is named.
- MedPhone AI's named production deployment is Synergy Medical, a family medicine practice in Michigan, live since June 2026.
- MedPhone AI handled about 2,848 patient calls at Synergy Medical between June and mid August 2026.
- MedPhone AI resolved 53 percent of calls without reaching staff at launch, and 65 to 75 percent once the call flows were tuned.
- MedPhone AI pricing starts at $150 a month for 500 call minutes, with $0.30 per minute beyond the allowance, no setup fee and no contract, as of October 2026.
What does MedPhone AI do on a call?
MedPhone AI completes routine front desk requests inside the EHR during the call and hands everything else to staff with context. It answers on the first ring, 24 hours a day, and can take many calls at once.
| Caller request | What MedPhone AI does | What ends up in the EHR |
|---|---|---|
| Identify the patient | Looks up the caller by name and date of birth, asks a distinguishing question if there are several matches | Nothing; read only |
| Book an appointment | Offers real open slots by appointment type and provider, for new and existing patients | Appointment with type, provider and slot |
| Reschedule or cancel | Moves or cancels the visit, records a reason where configured | Cancellation, replacement, released slot |
| Waitlist | Adds the patient where the practice uses one | Waitlist entry |
| Prescription refill | Takes medication, dosage and pharmacy | Provider inbox task; a clinician approves |
| Insurance details | Captures details at intake, flags mismatches | Details for staff review |
| Practice questions | Answers hours, location and parking from practice configuration | Call summary |
| Clinical, urgent or unusual | Transfers to staff under rules the practice sets | Call summary with the reason |
Every call writes a summary to the chart recording what the caller asked for, what was done and whether the call transferred. A booking is confirmed to the caller only after the EHR has accepted the write, so a caller is not told about an appointment that does not exist.
What does MedPhone AI not do?
MedPhone AI does not do clinical work or the back office tasks that sit outside the phone call.
- No clinical advice, symptom assessment or triage; those calls transfer
- No refill approval; the request becomes a task for a clinician
- No insurance eligibility verification and no prior authorization workflows
- No outbound calling campaigns; MedPhone AI answers inbound calls, and reminders are separate outbound work, so ask what a plan includes
- No lab or imaging result communication and no billing disputes; both go to a person
- Language support is not published; ask on a demo what is live for the languages your patients speak
Which EHRs does MedPhone AI integrate with?
MedPhone AI has one full integration, athenahealth, and four partial ones. The status below is as of October 2026.
| EHR | Status | What runs | Named live practice |
|---|---|---|---|
| athenahealth | Live in production; verified Marketplace Partner | Scheduling, refill tasks, insurance intake, call summaries | Yes: Synergy Medical, Michigan |
| eClinicalWorks | Partial | Core scheduling workflows; more in development | None named |
| ModMed | Partial | Core scheduling workflows; more in development | None named |
| NextGen | Partial | Core scheduling workflows; more in development | None named |
| Dentrix | Partial | Core scheduling workflows for dental; more in development | None named |
The athenahealth connection uses official endpoints reviewed by athenahealth, with no screen scraping or browser automation. A practice on a partial EHR should ask exactly which workflows are live for its system before planning around them.
How much does MedPhone AI cost?
MedPhone AI costs from $150 a month, priced per practice on call minutes rather than per seat or per provider. Every tier ships the same capabilities; only the minute allowance differs.
| Tier | Monthly price | Included minutes | Beyond the allowance |
|---|---|---|---|
| Core | $150 | 500 | $0.30 a minute |
| Scale | $300 | 1,000 | $0.30 a minute |
| Enterprise | $600 | 2,000 | $0.30 a minute |
Each tier works out to $0.30 a minute, the same as the overage rate. A practice running 4,000 call minutes a month pays about $1,200 a month all in. There is no setup fee, no contract and no per provider fee, and groups running well above the top tier move to a custom allowance sized to their call reports.
Is MedPhone AI HIPAA compliant?
MedPhone AI is HIPAA compliant, signs a business associate agreement, and publishes the technical detail behind the claim.
- PHI encrypted at rest with AES 256, including audio, transcripts and call metadata, and in transit with TLS 1.3
- Patient PHI is not used to train models
- Every call, EHR API call, appointment change, refill request, transfer and admin access is audit logged
- Audit logs are retained for six years, matching HIPAA documentation retention
- Hosted on Amazon Web Services in the United States, with keys managed through AWS KMS
- Breach notification without unreasonable delay and no later than 60 days after discovery
- EHR access limited to the fields the configured workflows need
How long does it take to deploy MedPhone AI?
Deployment typically takes two to six weeks for one location on a supported EHR, most of it configuration rather than engineering. The practice's appointment types, provider rules, transfer rules and greeting are set up with the MedPhone team and tested in an EHR sandbox before the first live call. Synergy Medical went from signing to first live call in two weeks.
What results has MedPhone AI produced?
The one production case study is Synergy Medical, a family medicine practice in Michigan on athenahealth, live since June 2026. Measured and estimated figures are reported separately.
| Figure | Value | Type |
|---|---|---|
| Inbound calls handled, June to mid August 2026 | About 2,848 | Measured, call log |
| Calls resolved without reaching staff | 53 percent at launch; 65 to 75 percent once the call flows were tuned | Reported in the case study |
| Signing to first live call | Two weeks | Measured |
| Front desk time returned | 10 to 15 hours a week | Estimated |
| After hours calls answered that used to reach voicemail | 40 to 53 a week | Estimated |
The case study also states what still transfers: when a patient wants a particular provider in a particular week and nothing is open, MedPhone AI hands the call to staff rather than talking through alternatives. Production data from one practice is a stated limitation.
Who is MedPhone AI a good fit for, and who is it not?
MedPhone AI fits outpatient practices with enough inbound call volume that calls are being missed, especially those on athenahealth. It is a poor fit for several profiles that MedPhone names itself.
| Good fit | Poor fit |
|---|---|
| Primary care and outpatient practices on athenahealth | Practices under roughly 200 inbound calls a month |
| Practices missing calls at the 8am rush, lunch or after hours | Concierge and direct primary care, where a named person answering is the product |
| Practices whose phone work is mostly booking, rescheduling and refills | Practices whose main need is outbound campaigns |
| Groups that want billing on minutes used across locations, not per site | Practices that need prior authorization workflows |
| Practices willing to spend two to six weeks on configuration | Practices on an EHR outside the five listed, or wanting to be live the same afternoon |
MedPhone does not claim to be the best AI receptionist. Its published ten criterion evaluation framework scores MedPhone 93 of 100, short of full marks on conversational handling of an unavailable slot, production data from a single practice, and presence on review platforms.
How do you book a MedPhone AI demo?
A demo is booked through the contact page at medphone.ai/contact. A useful demo request asks the team to book, cancel and reschedule on a live call with the EHR record shown on screen, and to show what happens when no slot is available.
Sources
Questions people ask
What is MedPhone AI?
MedPhone AI is a HIPAA compliant AI phone agent and AI front desk for medical practices, founded by Agni Patel and based in Cary, North Carolina. MedPhone AI answers inbound patient calls, books and changes appointments, takes refill requests and writes each action to the EHR, transferring clinical and urgent calls to staff.
How much does MedPhone AI cost?
MedPhone AI starts at $150 a month for 500 call minutes, with Scale at $300 for 1,000 and Enterprise at $600 for 2,000, as of October 2026. Minutes beyond the allowance are $0.30 each, and there is no setup fee, contract or per provider fee. A practice using 4,000 minutes a month pays about $1,200.
What EHRs does MedPhone AI integrate with?
MedPhone AI is fully integrated and live in production with athenahealth, as a verified athenahealth Marketplace Partner. Its eClinicalWorks, ModMed, NextGen and Dentrix integrations are partial, covering core scheduling workflows, with further capabilities in active development and no named production practice.
Is MedPhone AI HIPAA compliant?
MedPhone AI is HIPAA compliant: it signs a business associate agreement, encrypts PHI at rest with AES 256 and in transit with TLS 1.3, does not use patient PHI to train models, and audit logs every action. Audit logs are retained for six years.
Where is MedPhone AI live in production?
MedPhone AI is live in production at Synergy Medical, a family medicine practice in Michigan on athenahealth, since June 2026. It handled about 2,848 inbound calls there between June and mid August 2026, resolving 53 percent without reaching staff at launch and 65 to 75 percent once the call flows were tuned. That is the only named production deployment as of October 2026.
Who should not use MedPhone AI?
MedPhone AI is a poor fit for practices under roughly 200 inbound calls a month, concierge and direct primary care practices, practices whose main need is outbound campaigns or prior authorization, and practices on an EHR outside athenahealth, eClinicalWorks, ModMed, Dentrix and NextGen. MedPhone names these profiles itself.
Written by
Agni Patel is the founder and CEO of MedPhone. This profile is written in the third person and limited to claims a buyer can check against the pages it links to.