Solutions
Ambient AI
Scribe
An ambient AI scribe listens to the visit and drafts the note, so you don't dictate or type. Scribeberry records the conversation once the patient agrees, labels who is speaking, and turns it into a structured, review-ready note you can edit and send to your EMR.
Key capabilities
Hands-free capture
Press record at the start of the visit. The AI captures the full conversation, labels speakers, and structures the note. No dictation, no typing.
Review-ready drafts in seconds
Generation runs 5-15 seconds end-to-end after the visit. The draft uses SOAP, consult, or your custom template — edit inline and push to EMR.
Works in every clinical setting
In-person, virtual, telephone, and even noisy multi-room clinics. 40+ languages. Apple Silicon supports offline transcription for true privacy.
About Ambient AI Scribe
How ambient listening works. A laptop, phone, or tablet microphone captures the conversation after the patient agrees. Speech recognition produces a transcript, speaker labels separate clinician and patient, and a language model drafts the note in your template, whether SOAP, consult letter, or your own format. You review, edit, and sign, then send it to your EMR. Scribeberry doesn't keep the audio; only the transcript and the final note are retained.
What the evidence shows. At The Permanente Medical Group, 7,260 physicians used ambient AI scribes in more than 2.5 million encounters over 15 months; 84% reported a positive effect on communication with patients and 82% on work satisfaction (NEJM Catalyst, 2025). In the first randomized trial, 238 UCLA Health physicians across 14 specialties (NEJM AI, 2025), one scribe cut time per note by 9.5% and the other showed no significant change, while both showed signs of lower burnout. Physicians still saw occasional inaccuracies, so every note needs review.
Consent and when to switch modes. Canadian guidance from the CPSO and CMPA calls for telling patients about the AI scribe and getting their consent before recording. Scribeberry prompts that conversation when you start a scribe. In the UCLA trial, physicians estimated 6 to 7% of patients declined; for them, or in loud settings like the ED, switch to dictation and the same template still applies.
At a glance
Ambient AI scribe vs dictation vs human scribe
| Ambient AI scribe | Dictation software | Human scribe | |
|---|---|---|---|
| How the note gets written | AI drafts it from the visit conversation | You narrate; software types it | A person writes while listening |
| Your effort in the visit | Consent, then talk to the patient | Narrate during pauses or after | None beyond briefing the scribe |
| Note structure | Your template, filled in | As dictated unless AI formats it | Depends on the scribe |
| Typical cost | Free tier; Scribeberry Pro $99/month | Free built-in tools to enterprise quotes | $31,000–$42,000 a year full-time (PayScale 2024) |
| Who signs | You review and sign every note | You proofread and sign | You review and sign |
FAQ
What is an ambient AI scribe?
An ambient AI scribe is software that captures audio from a clinical encounter (in-person or virtual) and auto-generates a structured clinical note for review. The clinician doesn't dictate or type — the AI listens and drafts.
How accurate is ambient scribing?
Transcription is usually accurate for routine visits, but note quality varies with the visit and the specialty. In the UCLA randomized trial (NEJM AI, 2025), physicians reported occasional inaccuracies with both scribes tested, mostly omissions and formatting. Review every note before you sign it; complex multi-system visits need the most edits.
Does ambient AI work in noisy environments?
Modern speech-to-text models handle moderate background noise well — clinic hallways, exam-room chatter, and HVAC are typically fine. ICU and ED settings are harder; for those, dictation mode or a directional mic is more reliable than ambient capture.
Is patient consent required for ambient AI scribing?
Yes. Most jurisdictions (CPSO, CMPA, US state medical boards) require explicit patient consent before audio recording. Scribeberry includes a consent-first workflow that prompts the consent conversation at scribe creation.
What is ambient listening in healthcare?
Ambient listening is the technology that powers modern AI medical scribes. A microphone captures the natural patient-provider conversation during a clinical encounter, and AI converts the audio into a structured clinical note that the clinician reviews and signs.
Does ambient listening capture other conversations in the room?
Speaker labeling separates provider and patient voices. Background conversations (hallway noise, intercom announcements) are typically not transcribed if they're below a noise threshold. For shared exam rooms, position the mic close to the patient-provider conversation.
What happens to the audio after the encounter?
Scribeberry never stores audio permanently. The audio stream ends when the recording stops. Only the transcript and final note are retained — and they're encrypted, audit-logged, and deletable on demand.
Compare more
