Peerbits AI Scribe captures audio directly from your video call platform, generates a complete clinical note in real time, and delivers it to your EHR — without adding a single step to the virtual visit workflow.
DR. PATEL · PHYSICIAN
PATIENT · HOME
THE TELEHEALTH DOCUMENTATION GAP
Telehealth scaled faster than any healthcare technology in history. But the documentation infrastructure underneath it didn't change — physicians still type notes after calls, still lose visit context, still carry charting home.
38×
Virtual visit volume exploded — but documentation workflows remained manual. Every additional virtual visit added documentation burden without any tooling to absorb it.
11min
Physicians spend 11 minutes per virtual visit documenting after the call — equivalent to an entire additional appointment that generates no revenue.
Consumer-grade speech recognition degrades on video call audio — compressed streams, background noise, and network drops make standard transcription unreliable for clinical use.
Most telehealth platforms sit outside the EHR. Documentation happens in one system, the chart lives in another. Every note requires a manual bridge — copy, paste, format, code, submit.
HOW IT WORKS
AI Scribe for Telehealth integrates at the audio layer — capturing the physician-patient conversation directly from the video platform, without the patient seeing any additional interface or being aware of the documentation process.
AI Scribe connects directly to your telehealth platform's audio stream — Zoom Health, Doxy.me, Microsoft Teams Health, or your custom platform via SDK. No screen-share required. No third-party hardware. The physician sees a small AI indicator; the patient sees nothing different.
Video call audio is compressed — VoIP codecs degrade the signal that consumer speech recognition relies on. Our ASR model is trained specifically on compressed telehealth audio streams, maintaining 97% clinical accuracy even on low-bandwidth connections.
Speaker diarization identifies physician and patient speech independently — even accounting for the slight audio path differences in video calls — and routes each to the correct SOAP section without manual labelling.
As the call progresses, the note builds in the background — S, O, A, and P populating from the conversation in real time. By the time the call ends, the note is 90% complete and ready for a 30-second review.
The signed note pushes directly to the patient's EHR record via FHIR R4 — regardless of whether the EHR is the same platform as the telehealth tool or a completely separate system. No copy-paste. No portal switching.
Zoom Health (Zoom for Healthcare)
Native SDK Integration · HIPAA Business Associate eligible
Doxy.me
Browser-based · No download · HIPAA compliant
Microsoft Teams (Health)
Teams Health API · EHR-embedded virtual visits
Teladoc Health Platform
Enterprise API integration · Multi-specialty support
Custom / White-Label Platform
WebRTC SDK for any custom telehealth build
Amwell · MDLive · DrFirst
REST API integration for major telehealth networks
Clinical accuracy on compressed VoIP audio
Note delivered after video call ends
Post-call charting time eliminated per visit
Telehealth platforms with native integration
Workflow changes required for physician
BUILT SPECIFICALLY FOR VIRTUAL CARE
A documentation tool built for in-person care and adapted for telehealth is still fundamentally wrong for the job. Every capability in AI Scribe for Telehealth was designed from the ground up for virtual visits.
Clinical ASR trained on compressed video call audio — Opus, G.711, and G.722 codecs — maintaining accuracy at bandwidth levels that degrade standard speech recognition models by 30–40%.
When the connection drops mid-sentence, the AI fills the gap gracefully — using clinical context and surrounding conversation to reconstruct what was likely said, and flagging reconstructed segments for physician review.
Audio is captured at the physician's endpoint, encrypted immediately, processed in isolated infrastructure, and discarded within seconds of transcription. No patient audio is ever stored. Compliant by architecture, not policy.
After the call, AI Scribe simultaneously generates a plain-language visit summary for the patient — delivered automatically to your patient portal or sent via secure message. No additional physician action required.
Prescriptions mentioned during the virtual visit are captured, structured with correct drug name, dosage, frequency, and indication — and formatted for your e-prescribing workflow, ready to transmit without re-entry.
For asynchronous telehealth — where the patient submits a form or video message and the physician responds later — AI Scribe reads the patient submission and pre-populates the note's Subjective section before the physician even reviews it.
Telehealth E&M coding is different from in-person — the 2023 AMA guidelines, place of service codes, and modifier requirements for virtual visits are built into the coding engine. No manual modifier adjustment needed.
For telehealth platforms serving linguistically diverse populations, AI Scribe transcribes bilingual consultations and generates the clinical note in the physician's configured language — regardless of what language the patient speaks during the visit.
TELEHEALTH USE CASES
Virtual care isn't one thing — it's synchronous video, asynchronous messaging, remote monitoring follow-ups, and specialty consults. AI Scribe handles every format.
Synchronous Video · Primary Core
The core telehealth visit — physician and patient on video in real time. AI Scribe captures the full call, generates the SOAP note, and has it ready before the physician moves to the next patient.
WITHOUT AI SCRIBE
11 minutes of post-call charting after every video visit. 20 visits = 220 minutes of documentation that generates no additional revenue.
WITH AI SCRIBE
Note generated during the call. 30-second review at call end. Next patient within 2 minutes. No charting backlog at day's end.
Behavioral Health · Telepsychiatry
Telepsychiatry requires complete presence and therapeutic rapport — any visible documentation activity breaks the therapeutic frame. AI Scribe runs completely in the background, invisible to the patient.
WITHOUT AI SCRIBE
Psychiatrists type DAP notes after every session — often 20–30 minutes each. Patient sees a distracted clinician who types during emotional disclosures.
WITH AI SCRIBE
Full therapeutic presence. DAP note generated invisibly. Risk assessment, mood, and intervention captured from natural session dialogue.
Asynchronous · Store-and-Forward
For dermatology, optometry, and follow-up care — where patients submit photos or video asynchronously. AI Scribe reads the patient submission and pre-populates the Subjective section before the physician review begins.
WITHOUT AI SCRIBE
Physician reads patient submission, then manually writes the Subjective section — duplicating information that's already been documented by the patient.
WITH AI SCRIBE
Patient submission auto-ingested. Subjective pre-populated. Physician adds assessment and plan only — total documentation time under 3 minutes.
Remote Monitoring · Chronic Care
Remote patient monitoring generates data that needs clinical interpretation and documentation. AI Scribe ingests the RPM device readings alongside the video call and incorporates them into the Objective section automatically.
WITHOUT AI SCRIBE
Physician manually transcribes blood pressure, glucose, and weight readings from the RPM dashboard into the clinical note — duplicating existing data.
WITH AI SCRIBE
RPM data auto-incorporated into Objective. Trend language generated automatically: "BP improved from 162/98 to 138/82 over the past 14 days."
Specialty Consult · Second Opinion
Specialist consults conducted by video — cardiology, neurology, endocrinology — with the consulting physician seeing the patient remotely. AI Scribe generates the consult note and routes it back to the requesting PCP automatically.
WITHOUT AI SCRIBE
Specialist dictates consult note after the call, which gets typed, returned, and reaches the PCP hours or days later — delaying clinical decisions.
WITH AI SCRIBE
Consult note generated within 60 seconds of call end. Automatically routed to requesting PCP. Clinical decisions made same day.
On-Call · After-Hours
On-call physicians taking after-hours virtual visits are already fatigued. AI Scribe ensures that every after-hours encounter is documented completely — so nothing is missed and no documentation carries to the morning.
WITHOUT AI SCRIBE
On-call physicians often defer documentation on after-hours calls until the next morning — creating overnight documentation gaps and recall errors.
WITH AI SCRIBE
Every call documented in real time. On-call physician reviews and signs before the next call. No documentation deferred to morning.
THE TELEHEALTH DOCUMENTATION WORKFLOW
No new hardware. No separate app. No patient-facing changes. AI Scribe slots into your existing telehealth workflow at every point where documentation currently creates friction.
STEP 01
Open your telehealth platform as normal. AI Scribe activates automatically when the call begins — no extra button, no mode to select.
STEP 02
Talk to your patient exactly as you always do. Maintain eye contact. Ask the questions you need. The AI listens to everything — both sides of the call.
STEP 03
When the call ends, a complete SOAP note — with ICD-10 codes and E&M level — is ready within 60 seconds. One pass to verify, then sign.
STEP 04
Signed note flows directly to the patient's EHR record via FHIR. Patient summary sent to the portal. Next call in minutes, not after finishing notes.
TELEHEALTH PLATFORM COVERAGE
We don't ask you to switch telehealth platforms to access AI documentation. AI Scribe integrates at the audio layer of whatever video infrastructure your practice or platform already runs on.
Native SDK Integration
Embedded directly via Zoom Healthcare SDK. HIPAA Business Associate eligible. Works with existing Zoom licenses.
✓ NativeBrowser Extension + API
Works via Doxy.me API and a lightweight browser integration. No software install required for physician or patient.
✓ NativeTeams Health API
For Epic-embedded Teams virtual visits and standalone Teams Health deployments. Integrates with the existing Teams meeting infrastructure.
✓ NativeEnterprise API
Integration via Teladoc's enterprise API for large telehealth network deployments. Supports multi-specialty routing.
✓ APIREST API Integration
Connects to Amwell's platform API for enterprise telehealth networks and health system virtual care programs.
✓ APIWebRTC SDK
For telehealth companies running proprietary video infrastructure — our WebRTC SDK integrates at the media layer of any custom build.
✓ SDKTELEHEALTH COMPLIANCE & SECURITY
Virtual care documentation has compliance dimensions beyond the in-person setting — cross-state licensing, place-of-service coding, and audio data handling under HIPAA. Every requirement is addressed at the architecture level.
Audio captured at the physician endpoint. Encrypted immediately (AES-256). Processed in isolated infrastructure. Discarded within 60 seconds of transcription. Zero patient audio storage. BAA executed as standard.
Telehealth E&M visits require correct Place of Service codes (POS 02/10) and modifier 95 or 93 on applicable CPT codes. These are applied automatically based on the visit type — no manual modifier selection required.
Documentation supports multi-state telehealth compliance by capturing the patient's physical location state during the call — essential for licensing documentation and CMS interstate telehealth billing requirements.
CMS telehealth documentation requirements — including patient consent capture, technology type documentation, and connection quality notation — are built into every telehealth note template.
Annual third-party security audits specifically covering telehealth audio data handling. Available as part of our security documentation package for enterprise telehealth platform reviews.
Every telehealth note includes a call metadata log — call duration, platform used, connection quality, patient consent timestamp, and physician sign-off time — creating a complete audit trail without retaining audio.
SIDE-BY-SIDE COMPARISON
| Capability | Manual Post-Call Charting | Generic AI Transcription | In-Person AI Scribe (Adapted) | Peerbits AI Scribe for Telehealth |
|---|---|---|---|---|
| Works inside video call audio | ✗ | ~ Raw text only | ~ Limited accuracy | ✓ Native integration, 97% accuracy |
| VoIP-optimised speech recognition | N/A | ✗ Consumer model | ✗ In-person trained | ✓ Compressed audio trained |
| Structured SOAP output (not raw transcript) | ✗ Physician writes it | ✗ Unstructured text | ~ Partial | ✓ Complete SOAP / H&P / DAP |
| Telehealth-specific coding (POS, Modifier 95) | ✗ Manual | ✗ | ✗ | ✓ Auto-applied |
| Patient-facing visit summary to portal | ✗ | ✗ | ✗ | ✓ Auto-generated post-call |
| RPM data integration | ✗ Manual transcription | ✗ | ✗ | ✓ Auto-ingested to Objective |
| HIPAA audio compliance — zero retention | N/A | ✗ Often stores audio | ~ Varies | ✓ Zero retention · BAA standard |
"Telehealth gave medicine access it never had before. AI documentation gives telehealth the infrastructure it always lacked."
— Product philosophy, Peerbits AI Scribe for Telehealth
See how virtual care platforms are using Peerbits AI Scribe to eliminate post-call charting and improve documentation quality.
Healthtech ,
Remote patient monitoring app helps to bridge the gap between patients and healthcare providers. It tracks the vitals of the patients and sends it to the doctors.
Healthtech , AWS / Cloud ,
A healthcare startup struggled with increasing loads of data and manual infrastructure management as its business expanded. Peerbits successfully built cloud infrastructure using AWS for their system possessing auto-scaling, automated and more.
Healthtech ,
This is a native iOS app that helps to bridge the gap between the patients and healthcare providers. Patients can monitor their health on a regular basis and share the data with the doctors and healthcare professionals.
This is a consent question, not just a technical one. Yes — patients should be informed that an AI documentation tool is active during the visit. We provide compliant consent language that can be incorporated into your existing telehealth consent flow, typically presented before the call begins. The phrasing is straightforward: the AI assists the physician with documentation and the audio is not stored. Most patients find this unremarkable — it's analogous to informing them that a nurse might be in the room taking notes.
Our engine is specifically designed for the interrupted-audio reality of telehealth. When a connection drop creates a gap in the transcript, the AI uses clinical context — what was discussed before the interruption, what the physician said after — to reconstruct the most likely content, flagging the reconstructed segment for physician review with a yellow indicator. For drops longer than 15 seconds, the corresponding section remains blank with a prompt for the physician to complete it. Audio is never fabricated without disclosure.
Every telehealth note captures the patient's physical location at the time of the call — not just their billing address. This is essential for licensing documentation in states with cross-state telehealth practice requirements. The patient's state is auto-populated in the note metadata and available for your compliance reporting. For states with specific documentation requirements around interstate telehealth, those fields are configurable in the template builder.
Yes — this is explicitly supported via our WebRTC SDK. If you've built a proprietary telehealth platform, our SDK integrates at the media layer, capturing the physician-side audio stream before it reaches the video codec. Documentation: full API and SDK reference, sandbox environment, and dedicated integration engineering support are included in the enterprise engagement. Implementation typically takes 2–4 weeks for a clean WebRTC integration.
AI Scribe applies the correct telehealth-specific coding automatically — including Place of Service code 02 (telehealth other than patient home) or 10 (patient home), and modifier 95 or 93 as appropriate for the CPT code and payer. The 2023 AMA E&M guidelines apply equally to telehealth visits, and MDM-based E&M levels are calculated from documented complexity, not time. The coding engine is updated with CMS telehealth waivers and policy changes, which have been frequent in recent years.
Yes. For asynchronous visits where the patient submits information (photos, video, questionnaire) and the physician responds later, AI Scribe reads the patient submission and pre-populates the Subjective section of the note. The physician then adds Assessment and Plan during their review. For dermatology, optometry, and follow-up management — where async is the primary care model — this reduces per-visit documentation time to under 3 minutes.
Have more questions?
Ask our expertsSEE IT LIVE ON A VIDEO CALL
Our demo is a live video session. You'll see the AI annotation panel building in real time alongside a simulated telehealth visit — and we'll show you the note, the codes, and the portal summary it produces.
Expert insights on AI documentation for telehealth, virtual visit workflows, and clinical note automation.
