20-21 AI in NHS

Healthcare systems around the world are under growing pressure to deliver more with limited resources.

Costs continue to rise and demand shows little sign of slowing. Clinicians are being asked to manage increasing workloads while still delivering high-quality patient care.

It is easy to see why AI has become such a priority. Its potential spans everything from reducing administrative work and freeing up clinical capacity to supporting diagnostics, research and more personalised care.

In the NHS, we are already seeing what this could mean in practice. Following a recent trial, more than 500,000 staff are being given access to AI tools after participants reported saving an average of two days a month on administrative work.

But the ability to get value from AI depends heavily on the information it can access. Across healthcare, patient, clinical and operational data is often spread across disconnected systems. Until that fragmentation is addressed, it will be difficult for healthcare organisations to use AI to its full potential.

This is why the next phase of healthcare AI needs to focus as much on the foundations as the technology itself. Healthcare organisations need to connect data across systems, rethink how workflows are designed and prepare their people to work effectively alongside AI.

These changes need to be driven by a clear understanding of the outcome they are trying to achieve. Getting those foundations right will determine whether today’s AI pilots become another layer of technology or translate into meaningful improvements for healthcare systems and their patients.

The first wave of generative AI created enormous excitement across industries including healthcare. Organisations moved quickly to experiment, giving employees access to new tools and testing individual use cases across different parts of the business.

That experimentation was important, but it did not always translate into a clear strategic outcome. As the initial excitement around AI begins to mature, healthcare leaders need to become much clearer about the problem they are trying to solve.

Are we trying to reduce the cost of care, give clinicians more time with patients, improve patient access or improve the quality of care? Once the outcome is clear, healthcare organisations can work backwards to determine what needs to change to achieve it.

This requires a change in mindset. Too many healthcare transformation programmes start with a clear operational or patient challenge, but the focus can gradually shift towards implementing the technology itself.

Simply adding AI to an existing process will only take us so far. Healthcare organisations need to consider how that process could be redesigned around the capabilities now available.

We have seen this with previous waves of technology. When computers first arrived, organisations often used them like more sophisticated typewriters. The technology changed, but the underlying processes largely remained the same.
Healthcare cannot afford to repeat that approach with AI.

The challenge becomes clearer when we look at how healthcare information is currently managed.

A patient’s clinical procedures might sit in one system, while pharmacy information, staffing and operational costs sit elsewhere.

When those systems cannot connect the dots, clinicians can be left without a complete view of the patient. It also becomes much harder to understand the overall journey, cost and quality of care.

This is a very real challenge for the NHS. The UK government’s plans for a Single Patient Record are intended to address fragmented health information across England.

Healthcare professionals currently work across multiple systems, often with incomplete or inconsistent information, while patients can find themselves repeatedly explaining their medical history.

Addressing this fragmentation requires more than bringing information into one place. Healthcare organisations need to connect data across existing clinical and operational systems, establish common standards for how that information is accessed and shared, and put the right governance around it.

Interoperability will be critical, allowing different systems to work together and creating a more connected view of the patient without requiring organisations to replace every system they already use.

As AI becomes more embedded in healthcare, these foundations become even more important. Connected, trusted data gives AI the context it needs to work effectively.

This creates the foundation for AI to support patient access and wider administrative and operational workflows. It also creates the foundation for AI applications and agents to work across different parts of the organisation.

People will continue to play a critical role in healthcare, but the way they work alongside technology will evolve. An AI agent might handle routine administrative steps while a person focuses on exceptions or decisions requiring human judgement.

Recently, NHS trials of AI notetaking technology across more than 17,000 patient encounters found a 23.5% increase in direct patient interaction time and an 8.2% reduction in overall appointment length.

In A&E, the number of patients seen per shift increased by 13.4%.

For a health system under significant capacity pressure, giving clinicians more time with patients is a meaningful outcome. It also demonstrates why the conversation needs to move beyond adopting AI tools towards redesigning how work gets done.

Reskilling and change management must be part of that conversation too. Data, technology, processes and people are interconnected and treating them as separate transformation projects risks limiting what AI can achieve.

Healthcare systems operate differently around the world. The NHS is structurally different from the payer-provider model in North America, while other countries have their own funding and delivery models.

Yet many of the challenges inside healthcare delivery are remarkably similar. Patient journeys often span different parts of the healthcare system, making the ability to access the right information at the right time particularly important.

Clinicians need accurate information at the right time. Organisations need to manage capacity and understand the patient’s overall journey.

That creates an opportunity for healthcare systems to learn from each other, particularly as they explore how AI can reduce administrative work and improve the way information moves across services.

Ultimately, the success of healthcare AI should be judged from the patient’s perspective. Over the next three to five years, progress should mean more accurate diagnosis, increasingly personalised treatment and clinicians having more time to focus on the people in front of them.

There is no shortage of AI technology available. The challenge is turning it into meaningful outcomes. For healthcare leaders, that means starting with the problem they want to solve and aligning the data, workflows and people around it.

That is how today’s AI pilots can translate into better healthcare and better patient outcomes. ing them as separate transformation projects risks limiting what AI can achieve.

AI needs the right data behind it

AI will change how healthcare works

Keep the patient at the centre