Four NHS trusts in south west London have signed one of the largest artificial intelligence deals in UK healthcare history, with plans to deploy AI-powered clinical scribing technology to up to 20,000 clinicians across the region over the next four years.
The contract — between Lyrebird Health and the South West London Acute Provider Collaborative — marks a significant step in the NHS’s push to use technology to reduce the administrative burden on clinical staff and return more time to direct patient care. For Londoners using NHS services across south west London, the practical outcome should be doctors and nurses who spend less time typing notes and more time focused on the patient in front of them.
Which London Hospitals Are Involved?
The deployment covers four NHS trusts that together serve a large population across south and west London:
- St George’s University Hospitals NHS Foundation Trust — one of the UK’s largest teaching hospitals, based in Tooting
- Epsom and St Helier University Hospitals NHS Trust — serving communities across Surrey and south London
- Croydon Health Services NHS Trust — providing acute and community services across the London Borough of Croydon
- Kingston and Richmond NHS Foundation Trust — serving Kingston, Richmond and surrounding areas in south west London
Together these trusts represent a broad cross-section of NHS services — from major acute hospitals to community health settings — making this one of the most geographically and clinically diverse AI deployments the NHS has undertaken to date.
The programme aims to onboard 10,000 clinicians in the first year, with the rollout expanding to the full 20,000 over a four-year period.
What Is AI Scribing and How Does It Work?
AI scribing — also known as ambient voice technology (AVT) — is a category of software that listens to conversations between clinicians and patients during consultations and automatically generates clinical notes, summaries, letters and other documentation in real time.
The technology works in the background. A clinician begins a consultation as normal — the AI listens, transcribes and structures the conversation into formatted clinical documentation. The clinician then reviews, edits and approves the notes before they are saved to the patient’s record. Crucially, the AI does not make any clinical decisions — it handles documentation only. All clinical judgement remains entirely with the human clinician.
The Lyrebird Health platform used in this deployment is integrated directly with the Oracle Cerner Millennium electronic patient record (EPR) system shared by all four south west London trusts. This means clinical notes generated by the AI can be entered directly into patient records without manual re-entry. The system also supports AI-powered form population and gives clinicians automated access to patient demographics, medication history and medical background during consultations.
Why Does This Matter? The Documentation Burden on NHS Clinicians
To understand why this technology is attracting such significant investment, it helps to understand the scale of the documentation problem in the NHS.
Clinical documentation — the writing of consultation notes, referral letters, discharge summaries, coding entries and other administrative records — consumes a significant proportion of every clinician’s working day. Studies consistently show that many NHS doctors and nurses spend as much time on administrative tasks as they do on direct patient care.
The consequences are significant: clinicians face higher levels of burnout, documentation quality can suffer when staff are rushed, and time that could be spent with patients is diverted into typing.
Matthew Laundy, chief clinical information officer at South West London ICS, described the current situation plainly: documentation is currently “a huge admin burden on clinicians, which leads to some very poor quality documentation, which can impact on patient safety.”
NHS England estimates that AI scribing technology could save clinicians an average of two to three minutes per consultation — a figure that, multiplied across thousands of clinicians and millions of consultations, translates into an enormous amount of recovered clinical time.
A major NHS England-sponsored study led by Great Ormond Street Hospital’s innovation unit, conducted across nine NHS sites in London and more than 17,000 patient encounters, found that AI-scribing technology increased direct patient interaction time by 23.5% during appointments. If scaled nationally across A&E alone, researchers estimated the technology could generate more than 9,000 additional A&E consultations per day and save £176 million in documentation time annually.
What the Oxford Pilot Found
Before the south west London deployment was confirmed, Lyrebird Health’s technology was part of a pilot programme at Oxford University Hospitals NHS Foundation Trust, run by the trust’s digital innovation team between July and November 2025.
The results were encouraging. 87% of users — including doctors, nurses, allied health professionals and operational staff — reported a reduction in time spent on documentation. The most common benefit reported was a saving of between one and ten minutes per patient interaction, which across a full clinical day represents a substantial amount of recovered time.
The Oxford pilot included several AI scribing platforms — Lyrebird, Accurx Scribe, Heidi and TORTUS — and its findings directly informed NHS England’s decision to launch a national self-certified registry of approved ambient voice technology suppliers in January 2026. Lyrebird Health is one of 19 suppliers currently listed on that registry, all of which have demonstrated compliance with NHS standards on clinical safety, technology assurance and data protection.
Beyond Note-Taking: Automated Coding and Referral Tracking
The south west London deployment goes further than simply automating clinical notes. Two additional capabilities are being introduced as part of the programme:
Automated clinical coding: Clinical codes — the standardised identifiers used to record diagnoses, treatments and procedures in NHS systems — will be generated automatically from consultation content. Clinical coding is currently a labour-intensive process that requires specialist staff to review documentation and assign codes manually. Automating this step reduces delays, improves accuracy and frees coding teams for more complex work.
Referral to treatment (RTT) automation: RTT pathway data — which tracks the time between a patient’s referral and their treatment, a key NHS performance metric — will be captured automatically from consultation records without manual data entry. This directly supports the NHS’s ambition to reduce waiting times, including the target for 92% of patients to be seen within 18 weeks of referral by March 2029.
Both capabilities are significant. Together with the documentation automation, they represent a broader shift from manual administrative processes toward AI-assisted workflows across the clinical pathway.
Data Protection and Patient Safety: What Are the Safeguards?
Given that the technology involves recording patient consultations, questions of data protection and patient consent are central to any responsible deployment. NHS England has published guidance specifically covering the use of AI-enabled ambient scribing products in health and care settings, developed in consultation with the Information Commissioner’s Office (ICO) and the National Data Guardian.
Key safeguards built into compliant AI scribing deployments include the following. Audio recordings are processed in real time and are not stored long-term. All generated documentation is reviewed and approved by the clinician before being added to any patient record. The AI does not make clinical decisions — its role is limited entirely to documentation. Patients must be informed that the technology is in use during their consultation. Data processing arrangements must be clearly documented between the NHS trust and the technology supplier.
The South West London ICS has indicated that the deployment has been designed with safety and clinical governance at its core. Martin Ellis, chief digital information officer at the ICS, confirmed that the programme was structured to be “safe, deeply-integrated, and genuinely helpful to clinicians” — with a clear emphasis on IT and clinical governance supporting the technology rollout rather than running behind it.
Where South West London Fits in the National Picture
The south west London deployment is the largest single ambient voice technology rollout in the NHS to date, but it sits within a rapidly growing national movement toward AI-assisted clinical documentation.
Accurx Scribe — another platform on the NHS’s approved registry — is already in use across more than 40 NHS trusts and is available in 97% of GP practices in England. Hertfordshire Community NHS Trust recently announced its own AVT rollout following a successful pilot in community and mental health services.
Nationally, NHS England’s chief clinical information officer has described ambient voice technology not as “just another digital tool” but as “an enabler to truly reimagine healthcare” — signalling that AI scribing is now firmly established as a priority technology within the NHS’s digital transformation agenda.
The 10 Year NHS Health Plan, published earlier in 2026, specifically encourages closer collaboration between NHS trusts and health technology companies to embed AVT at scale — giving the south west London deployment a clear alignment with the direction of national NHS policy.
What This Means for Patients in South West London
For patients attending appointments at St George’s, Epsom and St Helier, Croydon Health Services or Kingston and Richmond NHS trusts, the transition to AI scribing should produce several tangible improvements over time.
Clinicians who spend less time on documentation during and after appointments can give more focused attention to the patient in front of them. The quality and completeness of clinical records should improve as AI captures information that might otherwise be abbreviated when clinicians are pressed for time. And automated RTT tracking should support the trusts in meeting their waiting time targets more consistently, benefiting patients who are waiting for treatment on NHS pathways.
This AI investment forms part of a broader package of improvements coming to London’s NHS services in 2026. Alongside the technology rollout, London patients are also set to benefit from major new investment in diagnostic and emergency care infrastructure — including upgraded MRI scanning capacity at Wembley, new equipment at Eltham Community Diagnostic Centre, and expanded urgent treatment facilities at North Middlesex and West Middlesex hospitals.
Key Facts: South West London AI Scribing Deployment
| Detail | Information |
|---|---|
| Technology | Lyrebird Health ambient voice technology (AVT) |
| Trusts covered | St George’s, Epsom & St Helier, Croydon Health, Kingston & Richmond |
| Total clinicians | Up to 20,000 over four years |
| Year one target | 10,000 clinicians onboarded |
| EPR integration | Oracle Cerner Millennium |
| Additional features | Automated clinical coding, RTT pathway automation |
| NHS registry status | Lyrebird is one of 19 approved AVT suppliers |
| Oxford pilot result | 87% of users saved time on admin tasks |
| GOSH study finding | 23.5% increase in direct patient interaction time |
All information in this article is based on confirmed announcements from the South West London Acute Provider Collaborative, NHS England, and Lyrebird Health, April 2026, supplemented by published NHS research and official guidance. For the latest information on NHS digital programmes in London, visit england.nhs.uk/london.
