Dear colleagues,

The NHS is increasingly using ambient voice technology (AVT) to improve the quality and experience of patient care, reduce administrative burden, improve productivity and support frontline staff.

We are writing to let you know that the Medicines and Healthcare products Regulatory Agency (MHRA) has today published guidance on the regulatory status of AVT products.

The guidance has been developed with input from NHS clinicians and patient safety advocates and provides greater clarity on how AVT products are regulated. It reflects the National Commission into the Regulation of AI in Healthcare’s recommendation for a clear, risk-proportionate approach.

The 10 Year Health Plan and Life Science Sector Plan set ambitions for a new regulatory framework that supports the NHS to become the most AI-enabled health system in the world.

The new guidance confirms that some AVT products solely intended to transcribe and summarise clinical conversations, draft correspondence or suggest clinical codes for review by a clinician are not to be regulated as medical devices under the current regulatory framework (MDR 2002).

The guidance also confirms that AVT products intended to support diagnosis, treatment and automated clinical actions are regulated as medical devices and remain subject to the relevant regulatory requirements.

The MHRA will continue to work with NHS England, the Department of Health and Social Care and other partners to develop further guidance as the technology evolves and the needs of the health system change.

What this means for NHS organisations

NHS organisations should pay close attention to the capabilities of their chosen AVT product and ensure it has the appropriate medical device certification, where applicable.

Further information is available in NHS England’s guidance: Adopting ambient scribing products in health and care settings.

NHS organisations remain responsible for ensuring AVT products are deployed safely and appropriately.

The Ambient Voice Technology Self-Certified Supplier Registry remains available to support organisations with their AVT adoption.

Integration requirements

Regardless of their regulatory status, NHS organisations should procure AVT products that can be interfaced into the organisation’s electronic patient record (either natively or through full integration) to support more advanced electronic patient record workflow functionality.

As AVT capabilities develop, they are likely to support more advanced actions, such as creating discharge summaries and ordering medicines or diagnostic tests.

National support for the scaled adoption of AI

NHS England is establishing national support to help NHS organisations adopt AVT.

This support will begin with outpatient settings, where AVT can reduce administrative burden, make better use of clinical time, and help create capacity for additional appointments. This can help reduce waiting times and ease pressure on overstretched clinical teams.

The national support will include implementation and change management support, commercial guidance and a clear benefits framework.

Support will then expand to other care settings: capturing what works, sharing practical lessons and supporting continuous improvement as adoption grows.

NHS England will work with exemplar trusts that have successfully trialled or deployed AVT to establish an AVT Centre of Excellence. This will build on the national AVT resources already available. The aim is to capture learning, improve the evidence base, and spread what works across the NHS quickly.

Expectations for boards and executive teams

Regulatory clarity is only one part of adopting AI-enabled technologies safely. Effective governance, workforce capability, professional standards and leadership are all crucial factors.

NHS organisations will therefore play a critical role in ensuring that AVT delivers benefits for patients and staff while maintaining public and staff trust and confidence. When deploying AVT products, boards and executive teams should assure themselves the following arrangements are in place:

1. Oversight and accountability 

• Users remain responsible for reviewing, validating and approving any information generated by AVT before it is relied upon for patient care.
• Appropriate professional judgement is maintained when using AI-generated transcripts, summaries and other outputs.
• Clear lines of accountability exist for decisions informed by AVT outputs.

2. Local governance

• AVT deployment is supported by appropriate clinical safety, information governance and digital assurance processes.
• Organisations understand the intended purpose and capabilities of products being used, particularly where they are enabling more advanced functionalities beyond transcription and summarisation.
• Processes are in place to review any change in functionality that may alter the regulatory status of a product.

3. Staff training and awareness

• Staff receive appropriate training on the capabilities and limitations of AVT.
• Users understand that AVT outputs may contain inaccuracies and require review to ensure accuracy.
• Safe use of AVT is embedded within existing clinical workflows and governance arrangements.

4. Risk management

• Organisations consider risks relating to privacy, consent, information quality and automation bias.
• Incident reporting arrangements are clear and understood.
• Learning from implementation and use is shared through existing governance processes.
• There is clear accountability for the review and use of AVT outputs.
• There is ongoing monitoring of risks, benefits and implementation outcomes.

5. Commercial and procurement

• Procurement processes include appropriate assessment of clinical safety, information governance, cyber security, interoperability and regulatory compliance requirements.
• Procurement decisions take account of long-term interoperability, scalability and the organisation’s ability to safely adopt future functionality.
• Organisations undertake appropriate due diligence to understand product functionality, limitations and deployment arrangements.
• Contractual arrangements provide appropriate rights to monitor and address performance, audit compliance, manage changes in functionality and address emerging risks throughout the contract term, including at AI model update milestones.

We would like to thank all colleagues involved in this work for their ongoing commitment to digitising the NHS and supporting the transition from analogue to digital ways of working, improving experiences for patients and staff alike.

Yours sincerely,

Publication reference: PRN02607_i