Research Shows AI Adoption in the NHS Will Be Shaped by Governance and Readiness

As part of the 2025 Rackspace Healthcare study, we partnered with Coleman Parkes Research to survey 75 NHS IT and digital leaders across the UK, examining how AI is being adopted and where it is delivering value.

In a previous blog post, we looked at how cloud adoption is advancing across the NHS and the integration challenges shaping those environments. In this post, we focus on how AI is moving into operational use, where it is delivering measurable impact and what is shaping the pace of adoption.

The data points to steady progress. AI is already being applied to improve efficiency, reduce clinician workload and accelerate access to insights. It also highlights a more specific challenge. Adoption is progressing within environments where security, governance and organisational readiness are still developing.

AI adoption is advancing but maturity varies

AI adoption across the NHS is progressing, but it is not uniform. Only 1% of organisations report AI fully embedded into business strategy, while 33% describe usage as minimal or ad hoc. At the same time, 51% are enhancing existing technologies with AI capabilities, and 41% are investing in new AI-enabled solutions.

Most organisations are introducing AI within existing systems rather than building entirely new platforms. As a result, AI is being integrated into existing tools and workflows, with adoption shaped by operational needs, available expertise and the constraints of complex NHS environments.

AI is already delivering measurable operational value

Our research shows that AI is already contributing to day-to-day operations across the NHS. Thirty-seven percent of organisations report reduced clinician workload, while 33% cite faster access to insights through data analytics and visualisation. In practical terms, this is helping teams spend less time gathering and processing information, and more time acting on it.

At this stage of adoption, the impact is most visible in efficiency and data utilisation. AI is supporting clinical and administrative teams by reducing manual effort, improving visibility and enabling more informed decision-making, rather than reshaping care delivery end to end.

These findings reflect how AI is being applied across NHS environments today. Organisations are focusing on targeted use cases where value can be realised quickly and where new capabilities can be introduced without adding unnecessary risk or disruption to existing workflows.

Adoption is shaped by security and organisational readiness

We’re seeing the pace of AI adoption shaped as much by organisational readiness as by the technology itself. Forty-five percent of organisations cite change management and clinical adoption as the primary challenge, while 44% identify security risks and vulnerabilities as a key concern.

That tells us something important. AI adoption depends on how well organisations can introduce new capabilities into clinical workflows, governance structures and existing systems, while maintaining control over data access and use.

For NHS organisations, that work is operational as much as technical. Teams need confidence in the outputs, clarity on how tools fit into established ways of working and oversight that keeps data use visible, governed and accountable.

Security concerns reflect the realities of healthcare data

Security concerns are clear in the data, and they’re grounded in the realities of working with sensitive healthcare information. In our research, 44% of organisations identify security risks and vulnerabilities as a key concern, while only 12% describe their organisation as cyber-resilient. That gap directly influences how AI is evaluated and introduced.

AI systems rely on access to patient records, clinical notes and operational data to generate insights and support decision-making. As these tools are introduced, organisations need a clear understanding of how that data is processed, where it is stored and who can access it, particularly when third-party models or external platforms are involved.

In practice, this creates some very real challenges. Data may be shared across systems, accessed outside established controls or used in ways that are not yet fully governed. In some cases, teams begin experimenting with AI tools before formal oversight is in place, making it harder to maintain visibility and control.

For NHS organisations, this comes back to responsibility. As adoption expands, maintaining control over how data is accessed, processed and governed becomes central to building confidence and scaling use.

AI outcomes depend on secure and governed data environments

Across our findings, a consistent pattern emerges. Organisations that are further along in establishing secure, well-governed data environments are better positioned to expand AI adoption with confidence.

We see this in how AI is deployed and managed. Where identity and access controls are clearly defined, data architectures are structured and governance frameworks are in place, organisations have greater control over how AI tools interact with systems and data. That level of control makes it easier to move beyond isolated use cases and apply AI more broadly across clinical and operational workflows.

Where these foundations are still developing, adoption tends to remain more contained. Teams focus on smaller, lower-risk applications, limiting exposure while governance, oversight and operational confidence continue to mature.

AI performance and scalability are closely linked to the strength of these underlying environments. As organisations strengthen security, governance and data control, they create the conditions needed to expand AI use with greater consistency and confidence.

Confidence in governance will determine how AI scales

AI adoption across the NHS will continue to build, supported by ongoing investment and growing use in day-to-day operations.

How far and how fast AI expands depends on how well organisations can establish control around it. Where governance is clear, data access is well managed and AI is integrated into operating models, organisations are in a stronger position to extend its use across clinical and administrative functions.

Where those foundations are still being established, adoption tends to stay focused on targeted use cases. Teams prioritise oversight, risk management and controlled rollout, expanding capability gradually as governance and operational maturity develop.

What we’re seeing is a more deliberate phase of AI adoption across the NHS. Organisations are moving forward with clear intent, balancing the opportunity to apply AI more widely with the need to maintain trust, compliance and operational stability.

AI adoption will be defined by control and integration

AI is already being applied across clinical and operational environments in the NHS. What varies across organisations is how well it is integrated into existing systems and how effectively it is governed.

For a deeper view of the UK findings, including cloud maturity and cyber resilience, explore the full NHS-focused Rackspace Healthcare survey report.