What is ambient clinical documentation?

For most physicians, the patient encounter does not end when the patient leaves the room. It continues at a keyboard, filling in SOAP notes, updating EHR records, and catching up on documentation that could not be completed during the visit. The American Medical Association’s 2024 data shows physicians work an average 57.8-hour week, spending 13 hours on indirect patient care tasks including documentation. Primary care physicians spend approximately three hours per day on clinical documentation alone . That administrative weight is the primary driver of a burnout rate that, while improving, still affects 43.2% of the profession .

Ambient clinical documentation addresses this directly, not by reducing the quality of records, but by removing the manual effort required to produce them.

How clinical documentation works today #

Today, a physician documents a patient encounter in one of three ways: typing notes in real time during the visit, dictating into the EHR immediately after, or completing charts after hours. Each carries a cost.

Typing during the visit divides attention between the screen and the patient. After-hours documentation, referred to in clinical settings as “pajama time”, consumes personal time and is consistently linked to burnout. Dictation reduces the gap, but still requires the physician to narrate observations rather than simply practise medicine.

Some practices address this with human scribes: a trained person who documents in real time from inside the consultation room. The American Academy of Family Physicians (2023) estimates the cost at $2,500–$4,500 per month per clinician. That covers wages but adds scheduling complexity and introduces an additional person into every patient encounter.

What is ambient clinical documentation? #

Ambient clinical documentation (also referred to as ambient scribing) is an AI-powered system that passively listens to the patient-clinician conversation, processes it using natural language processing and large language models, and produces a structured clinical note, ready for the clinician to review and sign off before it enters the EHR.

The clinician changes nothing about how they speak, how they move through a consultation, or how they structure their thinking. The system works in the background.

The output is typically a SOAP note: a four-part format covering the patient’s Subjective report, the clinician’s Objective findings, Assessment, and Plan. This is the standard structure in EHR systems for clinical documentation. The ambient system generates a draft of this note from the encounter in real time.

This distinguishes ambient documentation from two related but different approaches. Voice dictation still requires the physician to narrate observations; it transcribes speech but does not interpret clinical context. Human scribing produces a similar output but requires a trained person in the room. Automated clinical note-taking through ambient AI requires neither.

Current platforms operating in this space include Nuance DAX (Microsoft), AWS HealthScribe, Abridge, and DeepScribe.

What the evidence shows #

The case for ambient clinical documentation is increasingly supported by peer-reviewed evidence rather than vendor claims alone.

A study published in JAMA Network Open (2025) , covering 46 clinicians across 17 specialties, found that ambient scribe use reduced documentation time by a median of 2.6 minutes per appointment and cut after-hours EHR work by 29.3%. A separate study on Nuance DAX, published in Future Healthcare Journal (PMC, 2025) , found a reduction of 2.5 hours per week in off-hours documentation. The University of Iowa Health Care reported that following its systemwide rollout in September 2024, physicians recorded an average of 2.6 hours per week saved on after-hours documentation, alongside a greater than 30% reduction in burnout scores within 90 days .

On clinician satisfaction, a study at Emory Healthcare, published in the Journal of General Internal Medicine (2024) and replicated in a larger joint study with Mass General Brigham covering 1,430 clinicians ( JAMA Network Open, 2025 ), found that the proportion of clinicians who said their documentation process met their needs rose from 41.9% to 71% within 60 days.

Patient experience has also shown measurable improvement. Sachin Shah, MD, Chief Medical Information Officer at UChicago Medicine, summarised feedback received after deploying ambient documentation across 550 clinicians: “We heard really meaningful feedback from patients, who said things like: ‘My doctor is more present in our conversations.’”

What to watch out for #

The efficiency case is real, but implementation requires caution.

AI medical documentation automation systems can hallucinate, generating content that sounds clinically plausible but was not part of the encounter. Controlled LLM-based studies report hallucination rates of approximately 1–3% . In specific clinical contexts, including oncology nursing deployments, rates between 3% and 28% have been observed, reflecting variation by setting, specialty, and model. In a clinical record, even a low error rate carries meaningful patient safety implications.

Beyond hallucinations, ambient systems face structural limitations. They cannot capture nonverbal information: body language, visible signs of distress, or observations a clinician makes without verbalising them. Critical details discussed during the encounter can be omitted if not clearly articulated in the conversation.

Privacy and informed consent are active regulatory concerns. The FDA is currently reviewing ambient clinical documentation tools under medical device frameworks . Patients should be informed when this technology is in use; consent processes are not yet standardised across health systems.

The output is a draft. Clinician review before sign-off is not optional and should be treated as such operationally, not as a formality.

Ambient documentation and agentic AI in healthcare #

Ambient documentation solves one problem: it removes the manual effort of creating a clinical note from the encounter. What it does not do is connect that note to the broader workflow: pre-appointment data gathering, post-visit follow-up, care coordination across teams, or the patient communication that surrounds a clinical interaction.

This is where agentic AI in healthcare extends the model. Rather than handling a single documentation task, agentic systems coordinate across the full patient journey, gathering information before the appointment, updating records after it, and surfacing relevant history when the physician needs it.

At Vstorm, we built a multi-channel, pre-appointment AI agent for a US healthcare provider serving more than 100,000 members. The system integrates with clinic management and physician platforms, personalises communication based on patient history, and updates records automatically after each encounter. Each physician using the system now saves more than five hours per week. Read the full case study at vstorm.co/case-study/multi-channel-ai-agent-in-healthcare/ .

For healthcare organisations evaluating where ambient clinical documentation fits within a wider automation strategy, our Agentic AI in Healthcare page covers the implementation considerations in detail.

Ambient clinical documentation returns a straightforward thing to the physician: the ability to be present with the patient. The technology works in the background, the note is ready to review when the encounter ends, and the charts do not follow the physician home. The practical value is not in the sophistication of the model. It is in the hours it returns, and what a physician chooses to do with them.