NHS AI procurement must be more flexible, says national CCIO

  • 28 October 2025
NHS AI procurement must be more flexible, says national CCIO
Digital Health AI and Productivity Gains webinar panel (Credit: Digital Health)
  • NHS procurement of AI must be more flexible, said Alec Price-Forbes, national chief clinical information officer at NHS England
  • In a Digital Health Webinar Price-Forbes highlighted the need for a "robust monitoring system" for the regulation of AI in healthcare
  • Concerns were raised that AI adoption is outpacing regulation because digital clinical safety standards need updating

NHS procurement cycles for AI must be more flexible to keep pace with the rapidly changing technology, said the national chief clinical information officer for England.

Speaking in a Digital Health webinar on ‘AI and Productivity Gains’, Alec Price-Forbes said: “Absolutely procurements can’t be eight or 10 years now.

“They’re going to be one plus one plus one because this [AI] is moving at pace and we need to learn as we go along”.

At the event, which featured a panel of NHS digital leaders, Price-Forbes said that procurement for ambient voice technologies (AVT) was “a bit of a challenge at the moment” because there are “a lot of suppliers approaching clinicians and giving trials for free both within primary and secondary care and we need to be robust in ensuring that what we are doing is safe, assured and is going to deliver benefit”.

He confirmed the imminent launch of the national ambient voice technologies (AVT) self-certified registry for suppliers to show evidence of compliance.

“What we’ve done is create a sort of model blueprint, we’re not calling it a commercial framework but more a strategic framework, which outlines the self-registry compliance requirements, both from a business capability perspective, as well as a compliance, adherence, platform stability, understanding whether a supplier has the ability or a roadmap to integrate into the electronic record systems and what other functionality they’ve got.

“This is moving fast, but if we’re going to deliver benefit we’ve got to be really clear otherwise we won’t be comparing apples and apples,” Price-Forbes said.

He said that there must be a “robust monitoring system” in place and that NHSE will be asking suppliers to provide real-time analytics that can be monitored and evaluated “so that we can start to understand model data drift, hallucinations etc.”.

Speaking in the same session, Yvette Khozam, lead electronic prescribing and medicines administration pharmacist at West London NHS Trust, shared her concern that AI adoption is outpacing regulation because “the DCB0129 and DCB0160 [digital clinical safety standards] haven’t substantially been updated” for several years.

“Updates have begun recently; we want to be the most AI-enabled care system in the world and we definitely have the capability to do so.

“But we just need to request specific things because existing standards were designed for deterministic software, not learning systems that can change their behaviour over time or optimise for the wrong things,” Khozam said.

Jacqui Cooper, chief nursing information officer at Health Innovation Manchester, said that the key to good digital transformation is that “it’s clinically led, operationally delivered and digitally enabled”.

“I think we’ve all been guilty of putting the tech first and that’s why we’ve failed, because that just doesn’t work, so engaging our people and making our people lead the innovation and the digital transformation is the key to it.

“In terms of AI we run a real risk of not doing that and not doing that effectively.

“We run a real risk even for the simplest of things that we’re using, like the Office 365 Copilot which we use nowhere near the full functionality and capability of that,” Cooper said.

Subscribe To Our Newsletters

Subscribe to our newsletter

Subscribe To Our Newsletter

1 Comments

  • Speaking as chair of the BCS Health & Care AI group, I fully agree that NHS procurement of AI needs to be flexible, according to clinical need, safety, accuracy, whether explanations are important for that use case as well as ease of integration into existing systems and freedom from bias. But a key question is, how to trade off these different factors ? For example should the NHS procure an AI which lacks explanations if it performs at an extra high level of accuracy – as suggested by a small study we are carrying out at Univ. of Southampton. See this LinkedIn post https://www.linkedin.com/feed/update/urn:li:activity:7389249096751935488/ for more

Comments are closed.

Related News

West Midlands NHS imaging programme enters procurement

West Midlands NHS imaging programme enters procurement

An NHS imaging transformation programme serving 6.6 million people across the West Midlands has entered procurement.
FT analysis suggests FDP benefits are very uneven

FT analysis suggests FDP benefits are very uneven

A report in the FT suggests waiting list reductions attributed to the FDP are heavily driven by figures from a small number of NHS hospitals. 
NHS leaders call time on AI pilots and demand national scaling

NHS leaders call time on AI pilots and demand national scaling

Senior NHS digital leaders have called for an end to repeated AI pilots, arguing that the focus must now shift to implementation and scaling.