.png?width=275&height=77&name=HETT%20insights%20logo%20RGB-01%20(2).png "HETT insights logo RGB-01 (2)")
The following blog has been adapted from the live session transcript of ‘To AI or Not to AI - Who Decides, What Works, and What Scales’, held at HETT Show 2026.
From AI Ambition to Adoption: What Does an AI-Ready NHS Look Like?
The NHS 10-Year Plan sets out an ambitious vision for the future, including the ambition for hospitals to become AI-ready. But what does that actually mean on the ground?
For clinicians and healthcare organisations, adopting AI is not simply about finding a promising technology and putting it into practice. There are questions around education, governance, culture, cybersecurity, regulation, patient involvement and, ultimately, whether the technology actually improves care.
At HETT Show 2026, moderator Robin Carpenter brought together perspectives from across the healthcare and technology ecosystem to explore what it will take to move from AI ambition to successful adoption. The panel included Alexander Deng, Sarah Hart, Matthew Green, MiRa Jacobs and Yaasha Hasan, each representing a different part of the AI adoption journey.
Starting with the clinician
The journey towards an AI-ready NHS starts on the ground, with clinicians working through the AI lifecycle - from developing and testing products to putting them into practice, monitoring whether they continue to work and eventually replacing them when they no longer do.
But clinicians are surrounded by a much wider ecosystem. They need the appropriate education to do their jobs well, alongside governance frameworks covering areas such as clinical safety, data protection and medical device regulation. These are also influenced by the wider ecosystem, including regulation changes, the NHS 10-Year Plan and the third sector.
As Sarah Hart, Chief Clinical Information Officer at University Hospitals of Morecambe Bay and Clinical Lead for the ICS-wide AVT deployment for Lancashire & South Cumbria, explained, there is already a huge range of AI technology being offered to healthcare organisations.
“A lot of the time we get approached by suppliers, and you think that the product is one thing, but actually it's an AI product that's maybe labelled as something different.”
For Sarah, the first question is whether a product will actually benefit staff.
“The last thing we want is to bring something in that's going to add to that sort of cognitive burden,” she explained.
Staff are not looking for technology for technology’s sake. They want to have good encounters with patients, work well with colleagues and have an efficient time of things.
The same applies to patients. Is the product safe? Will it create efficiency in the care that can be given without adding to the cognitive burden? And perhaps most importantly, will it improve the workflow? “The NHS doesn't lack intelligence. The problem in the NHS isn't a lack of intelligence. It's a lack of proper workflow.”
For Sarah, that means asking whether a new product will actually help solve a workflow problem rather than simply introducing another layer of technology. AI is already changing the patient-clinician relationship. The discussion also explored how AI is changing the way both clinicians and patients interact with technology.
One example is the increasing use of AI-enabled tools, including ambient voice technology and devices such as smart glasses. Clinicians using AI solutions in practice receive training and have the ability to validate the output, identify errors and apply their clinical judgement. Patients may not have the same level of understanding.
Sarah highlighted the emerging situation where patients may attend consultations with their own AI solutions, using their phones or other devices to record the encounter and create their own notes. The question is what happens when those tools get something wrong. “Patients don't necessarily have that.”
If a patient uses a patient-centric AI solution to create a note and it records something incorrectly, what could the consequences be? And what can healthcare organisations do to support patients to use these tools while making sure they are not being used in a dangerous way?
The discussion highlighted that the balance between patients and clinicians is shifting and that the healthcare system needs to understand what that means.
Five Ps for successful AI adoption
Moving from an AI idea to successful delivery requires more than a good product.
Yasha, who leads emerging technology and AI and automation initiatives across University Hospital Leicester and University Hospitals Northamptonshire NHS Trust, outlined what she described as the five Ps: Process. People. Product strategy. Project governance. Promotion.
The first is process.
Organisations need a clear strategy that is aligned with their wider organisational priorities. For Yasha, that means being clear about the return on investment, including cash release, patient impact and staff experience.
There also needs to be a clear operating model for taking an innovation from an initial idea through to an adopted solution that can continue to be optimised. “We're not stopping with any solution these days. Every solution can solve every problem.”
The second P is people.
Organisations need a senior responsible owner who can support and champion initiatives from end to end, as well as the right team around them. There also needs to be capacity and resources, with those resources prioritised according to the organisation’s strategic priorities.
The third is product strategy.
With the number of systems already operating across healthcare organisations, there needs to be a clear platform strategy. Before buying something new, organisations should ask whether the capability already exists within their existing systems, whether a national application could be used, whether something could be built internally, or whether an off-the-shelf product is genuinely the right option. The key is to think about integration from the beginning. “A lot of the learnings we have had is that we've hit the ground running with something that we couldn't really integrate at the end.”
The fourth P is project governance.
Yasha described the importance of having the right governance in place from the point a project is prioritised, through to delivery, assurance and compliance. Her organisation established an AI governance office to evaluate AI solutions and help determine whether the organisation had the risk appetite to become an early adopter of a particular technology. Cybersecurity, AI governance and clinical safety all need to be considered, with the relevant leads involved from day one.
And finally, there is promotion - or, more importantly, communication. “I live by the motto of over-communicating is better than under-communicating.”
For Yaasha, communication should start well before a technology is deployed.
The earlier people understand what is happening, the more opportunity there is to find champions who can support deployment and adoption.
Culture comes before technology
Matthew Green, Chief Information Officer at Barnardo’s, brought a perspective from outside the NHS and highlighted the importance of culture. Every time a new technology is introduced, the conversation ultimately comes back to people.,
“No piece of technology will ever save you from poor processes in an organisation, so you have to put your energy and effort into processes first.”
At Barnardo’s, rather than immediately creating a group of AI champions, the organisation established an advisory panel that deliberately included people who were sceptical about AI.
The aim was to create a space where people could debate concerns around AI, including job security, bias and problems with language models. It meant the organisation could understand its position before simply moving forward with the latest technology.
Cybersecurity was another important part of the conversation.
For Matthew, organisations should not try to reinvent the wheel when it comes to AI security. They should go back to the basics, use the policies and controls already in place, and make sure they all line up. Barnardo’s also made deliberate choices about which AI platforms could be used within the organisation, creating what Matthew described as a “walled garden” so that potential issues could be contained.
But cybersecurity is not only about technology.
The organisation also focuses on giving staff the skills and critical thinking they need to work in an AI world - helping them understand what they are hearing, question whether something is real and consider how it could affect the organisation.
Building an AI-skilled clinical workforce
The panel also looked at whether clinicians and healthcare leaders are ready for AI and where the education system is currently falling short.
Alex Deng, a Doctor in clinical genetics and consultant in genomic AI, highlighted the huge demand for clinical AI education. The NHS Fellowship in Clinical AI received nearly 600 applications for just 34 competitively funded places.
The challenge is that there is currently no clear, established pathway to becoming a digital doctor or AI clinician. As Alex explained, clinicians who move into this space often describe an “insane random pinballing path” through industry fellowships, time out of training and unpaid work. “There's no legible signal that they hold other than, ‘Here are the things that I've done.’”
That creates a problem for both employers and the profession. Employers do not necessarily know what skills someone has, while clinicians have to navigate a difficult and often luck-based route into the profession.
For Alex, creating clear signals of competence and expertise is critical. “I think those signals and pathways need to be put in there because otherwise, it's a very unscalable solution if everyone getting into that position of influence has to be lucky rather than being able to follow a route that is available to anyone who wants it.”
Rethinking regulation and oversight
AI is not static. Unlike a traditional product, even a small change in its environment can affect its performance.
MiRa Jacobs, Head of Digital Health at the MHRA, argued that this requires a shift in how we think about oversight. Rather than seeing oversight as a series of separate, sequential processes, she described the need to think about it as a parallel process, with different parts of the system working together.
The question is not simply who is responsible for one individual part of the system, but:
Who is responsible for what? Who needs those resources when? And how do we make sure we are reacting to data proactively and on time?
This also raises difficult questions around liability. As AI becomes more embedded in clinical decision-making, responsibility cannot simply sit with one individual who happens to be using the technology. Clinicians must still maintain professional decision-making, but organisations also have a responsibility to ensure that the tools being introduced into healthcare are appropriate, safe and secure.
The panel agreed that these questions will become increasingly important as clinicians are asked to make judgements about the appropriate level of human oversight for different AI tools.
That also brings us back to education. Clinicians are increasingly being asked to understand areas outside their traditional clinical domains - including how AI systems are designed, governed and monitored.
Putting patients at the centre
One question from the audience brought the conversation back to the most important part of healthcare: the patient. How can the patient voice be brought into the AI conversation?
Yasha explained that her organisation had established a patient resource group, inviting local patients to be involved in decisions around AI, EPR and other technologies that would affect them. Patients have even been involved in procurement panels for AI technology. “I think there's more as an organisation we can do to, again, this goes back to the basics, governance.”
For Sarah, keeping patients at the centre is the whole reason the healthcare system exists.
Patients are already using AI tools themselves, but there are also people who are technologically frightened or less confident using these technologies.
The challenge is to make sure patients can use the tools available to them safely, while feeling empowered as part of their care. Mira also highlighted that the traditional assessment of risk and benefit may need to change. A patient’s appetite for risk may be different depending on their particular health condition, raising important questions about how patient voices are represented when decisions are made about access to new technologies.
What is the one change that could help deliver the 10-Year Plan?
To close the session, each panellist was asked one final question:
If you had the Prime Minister in front of you, what is the one change you would make to help deliver the 10-Year Plan?
For Yaasha, the biggest shift needs to be culture. Her recommendation would be to bring senior leaders from healthcare organisations together to work through the risk appetite journey and learn by doing. “You're not going to learn until you actually do it.”
For MiRa, it was about confidence across the system - creating a shared language and giving organisations confidence in what they are hearing from each other so they can build momentum rather than repeatedly doing the same work.
For Matthew, the focus was children and young people. We do not yet know enough about what children and young people are using AI for, how they are using it and what this means for the future workforce. Their voices need to be part of the decisions being made now.
For Sarah, the answer was simple: “Please, can you make the NHS app fit for purpose?”
She described a vision for the NHS App to become a proper bridge between primary care, secondary care, mental health, community and children’s services - a health record in your pocket that allows patients to access seamless care wherever they are.
And for Alex, the priority was to recognise that digital medicine is becoming a distinct and increasingly important profession. His pitch was for the creation of a Royal College of Digital Medicine. Moving from ambition to action.
The discussion at HETT 2026 demonstrated that becoming an AI-ready NHS is about far more than adopting new technology. It requires the right processes, people, governance, education and culture. It requires organisations to understand where AI can genuinely improve workflows, while making sure that patients and clinicians remain at the centre of decisions. It also requires a willingness to learn.
As the panel highlighted, we are still at the beginning of this journey. The technology will continue to evolve, and so will the questions around regulation, liability, workforce, cybersecurity and patient involvement. The challenge now is turning ambition into action and creating the systems, skills and confidence needed to make AI work safely and effectively across healthcare.
Session Speakers
%20(1).png?width=275&height=77&name=HETT%20insights%20logo%20RGB-04%20(1)%20(1).png "HETT insights logo RGB-04 (1) (1)")