A 2025 study in QJM tracked how NHS resident doctors actually spend their time. For every hour they spent with a patient, they spent four hours on administration. Most of that administration was documentation.
That ratio is not an anomaly specific to junior doctors either. The RCGP’s own tracking data found that GPs spend, on average, a third of their working time on what they themselves consider unnecessary workload and bureaucracy. Ambient AI Scribing is the process that exists to change that ratio.
What is ambient AI scribing?
Ambient AI scribing is the process of using an AI-powered scribe to automatically generate structured clinical notes from a live consultation, without the clinician needing to dictate, type, or interrupt the appointment to capture anything.
The word “ambient” matters here. Unlike dictation tools, which require you to speak directly to the software after the patient has left, an ambient AI scribe runs passively in the background. It listens to the natural dialogue between you and your patient. The consultation happens. The note appears. You review it, edit if needed, and approve it.
That is the whole workflow. Not a transcription tool that produces a wall of text for you to turn into a note. A structured clinical draft, formatted for your clinical system, ready for your review in under a minute.
How is this different from dictation?
The distinction is real and worth spelling out.
Traditional dictation requires you to finish seeing the patient and then perform the documentation task: mentally reconstruct the consultation, organise it into a clinical structure, and narrate it. The tool removes the typing, but the cognitive work of writing the note still belongs to you.
An ambient medical scribe removes that step entirely. You do not reconstruct the consultation because the scribe was present for it. You are reviewing a draft, not composing a document from memory. For clinicians working through a full clinic list, that difference accumulates quickly.
What does it actually look like in practice?
HealthOrbit Ambient Scribe works directly within NHS clinical environments, with integrations for both SystmOne and EMIS. The workflow is straightforward from the first session.
Before your first patient, you open the application and start a session. It runs in the background while you consult. When the patient leaves, you end the session.
Within moments, a structured clinical note appears. In primary care, this typically follows a SOAP format covering the presenting complaint, history, relevant examination findings, your clinical impression, and the management plan, including safety-netting. In secondary care, the output adapts to what you need: an outpatient letter, a clinic note, a discharge summary.
You read the note, make any edits, and approve it. It then pushes directly into SystmOne or EMIS, without any copy-pasting or screen-switching.
The Ambient Scribe has been piloted across NHS primary and community care settings, including Haxby Group Practice and Navigo Care NHS Trust, where clinicians reported meaningful reductions in post-consultation admin. Most find their footing with the review stage within a session or two.
What does the evidence say?
The research picture on ambient AI scribing has developed significantly over the past two years, and the findings are worth knowing.
A quality improvement study in JAMA Network Open followed 263 physicians across six health systems. After 30 days with an ambient AI scribe, burnout among ambulatory clinicians fell from 51.9% to 38.8%, with significant improvements in after-hours documentation time, cognitive load, and patient attention.
A randomised clinical trial in NEJM AI, led by UCLA Health, examined 238 physicians across 14 specialties and 72,000 patient encounters. Physicians using the ambient AI tool reduced documentation time by nearly 10% compared to usual care, with measurable improvements in burnout and work-related stress. Fewer than 10% of patients declined its use, which matters for any clinician uncertain about how patients will respond.
The NHS Long Term Workforce Plan estimates that AI can automate up to 44% of general practice administrative workload, and that speech recognition alone can save one minute per consultation. One minute across a full clinical day adds up faster than it sounds.
Why this matters beyond saving time
The documentation burden carries consequences that go beyond a long working day. According to the RCGP, 76% of GPs believe patient safety is being compromised by their workloads, and 60% say they do not have enough time to adequately assess and treat patients during appointments.
42% say they are likely to leave the profession within five years, with administrative load consistently cited as a primary reason.
Ambient AI scribing addresses one concrete slice of that pressure. When the documentation is handled passively during the consultation, the work that was previously waiting for you at 6pm is either done or reduced to a brief review. That changes the shape of the working day.
There is also something less easily quantified but consistently reported by clinicians who have adopted an ambient AI scribe: being present in the consultation feels different when you are not simultaneously trying to document it. Patients notice when you are looking at them rather than a screen. The quality of the clinical interaction improves when your attention is undivided.
What about accuracy? Should you be concerned?
Yes, and you should stay that way.
Ambient AI scribe for healthcare used in NHS settings are trained on clinical language and handle the way real consultations unfold: interruptions, tangents, patients who mention their blood pressure at the end of the appointment, and the clinical shorthand that general-purpose transcription tools miss entirely.
But no system is infallible. The UCLA randomised trial specifically noted that AI scribes occasionally generate clinically significant inaccuracies, and that physicians must remain vigilant in reviewing AI-generated documentation. The clinician review step exists precisely for this reason. The tool drafts. You verify. The record only reflects what you approve.
With HealthOrbit AI’s Ambient Scribe, which is 95% accurate, that review step is still not an optional feature. It is the point. Nothing enters the clinical record without your explicit sign-off, and the interface is structured to make that review fast without making it easy to skip.
Treating the note as a working draft rather than a finished product is the right posture, at least until you have developed enough familiarity with how the scribe handles your clinical style and specialty.
Does it work in secondary care?
Yes, though the documentation requirements differ from primary care and the scribe needs to be configured accordingly.
Outpatient letters, clinic notes, and discharge summaries each have their own structure, clinical vocabulary, and downstream recipients. A respiratory clinic note does not follow the same template as a GP SOAP note, and a well-configured ambient AI scribe handles that distinction.
HealthOrbit AI supports both primary and secondary care documentation, with output templates that adapt to clinical context. For secondary care teams managing high outpatient volumes, that means the scribe fits the work rather than the other way around.
The secondary care documentation backlog is a separate but related problem. Outpatient letters that are delayed or incomplete create friction for patients, for the GP receiving the letter, and for the hospital managing its own referral and follow-up workflows. When clinicians can review and approve notes within the same clinic session, that backlog shrinks.
What about patient data and NHS compliance?
Any ambient AI scribe operating in an NHS environment needs to meet DTAC requirements, maintain DSPT compliance, and sit within the MHRA’s framework for Software as a Medical Device. Where audio is processed, how long it is retained, and whether it leaves UK infrastructure are live questions for any procurement decision.
HealthOrbit Ambient Scribe is NHS-approved and DTAC-compliant, with all data processed within UK infrastructure. It is also a member of the NVIDIA Inception program. Before any ambient AI scribe goes live in your practice or trust, those compliance questions should already have been answered by the vendor. If they have not been, that in itself is the answer.
Conclusion
The case for ambient AI scribing as a process is not built on efficiency metrics alone, though the metrics are real and well-documented. It is built on the fact that a significant portion of what currently fills a clinician’s working day, and spills into their evenings, is documentation work that an AI scribe can now handle reliably.
With patient-to-GP ratios in England now at 2,294 per full-time equivalent, an increase of 7.2% since 2019, the pressures on clinical time are only going in one direction. An ambient scribe technology does not resolve that structural problem. But it does return hours to the working week that most clinicians would rather spend elsewhere.
HealthOrbit’s Ambient Scribe is one of the few solutions built specifically for NHS workflows, with live deployments across primary and community care and direct integrations into the clinical systems you already use.
The consultation stays yours. The paperwork does not have to.
See HealthOrbit Ambient Scribe in action. Book a demo today.
FAQs
Do I need to change how I consult for the ambient AI scribe to work?
No. The ambient AI scribe listens to your consultation as it naturally happens. You do not need to speak differently, pause for the system, or follow a specific format. The process is designed to work around you!
How much time does it take to revise and accept the generated note?
In general, reviewing and accepting a note for a routine consultation takes less than one minute. This is because the system delivers the note to you pre-formatted and organized; all that remains is for you to check and edit it.
What happens if the note contains an error?
You catch it before it enters the record. The clinician review step is mandatory, and nothing is submitted to SystmOne or EMIS without your explicit approval. The scribe produces a draft. You remain the author of the final note.
Can the ambient AI be used for more complex and sensitive consultations?
Yes. The ambient AI scribe can capture notes from consultations that involve mental health care and chronic disease management. Indeed, clinicians working within community mental health care clinics, such as the Navigo Care NHS Trust, have used this software successfully in just such circumstances.
How is patient consent handled?
Patients are informed that an ambient AI scribe is in use before the consultation begins. In clinical trials, fewer than 10% of patients declined, and most found the explanation straightforward. Your practice or trust will have a consent process in place as part of the deployment.
