As digital systems become more embedded in day to day work and data is used more actively, it becomes easier to see where implementing AI strategically could have a real impact. Particularly in areas where clinicians and operational teams spend a disproportionate amount of effort on process rather than care.
There is also a shift in what AI is being asked to do. Attention is moving away from broad promises and towards specific support inside existing workflows, focused on removing friction rather than adding new demands for teams already balancing competing priorities.
Looking ahead to 2026, the areas that matter most are those grounded in how healthcare actually works. Where tools fit naturally into clinical and operational systems, build on data teams can trust and are introduced with the right safety checks and governance around them. AI has the potential to reduce friction in everyday delivery and give time back to the people providing care.
Focus area 1: Using AI to reduce administrative burden
AI is starting to find a more practical footing in healthcare. Rather than being positioned as something services need to reorganise around, it is increasingly being used to support existing systems in areas where administrative work absorbs time without improving care.
Administrative work has a direct impact on how healthcare is delivered. Time spent on paperwork competes with time spent with patients and coordinating care across teams, and delays in documentation can ripple through a service, affecting discharge, handover and follow up. Reducing that burden does not require changing clinical judgement or redefining roles. Even modest improvements can ease pressure at key points in the pathway.
Documentation is where this focus area is most visible. Tools such as ambient documentation assistants are being tested to help draft routine content, including discharge summaries and clinical notes, using information already held in electronic health records. Clinicians remain responsible for reviewing and approving what is produced, but starting from a draft rather than a blank page reduces the effort required to complete repeatable tasks.
AI is finding a more practical footing in healthcare, particularly in areas like clinical documentation and discharge tasks. Rather than being positioned as something services need to reorganise around, it is increasingly being used to support existing systems in areas where administrative work absorbs time without improving care.
Administrative work has a direct impact on the delivery of care. It reduces the time available for patients and slows the movement of information between teams. When documentation is delayed or incomplete, it can create avoidable handover gaps, hold up discharge and add extra follow up work across the pathway. Reducing that burden does not require changing clinical judgement or redefining roles. Even modest improvements can ease pressure at key points in the service.
Tools such as ambient documentation assistants are being tested to help draft routine content, including discharge summaries and clinical notes, using information already held in electronic health records. Clinicians remain responsible for reviewing and approving what is produced, but starting from a draft rather than a blank page reduces the effort required to complete repeatable tasks.
What teams are paying attention to
- How well AI tools integrate into existing clinical systems
- Whether improvements are felt by staff, not just measured in time saved
- How accuracy, accountability and clinical oversight are maintained
- The impact on patient flow, particularly around discharge
What to watch as this develops
- Wider rollout of documentation support beyond pilot teams
- Clearer evidence of time saved at discharge and handover points
- More consistent clinical safety approaches for administrative AI
Focus area 2: Using digital services to help patients complete tasks, not just find information
Digital access in healthcare is starting to move beyond signposting towards task completion. National services such as the NHS App are being positioned as places where patients can resolve common needs directly, rather than being redirected across phone lines, forms and follow-up appointments.
AI is what makes this shift more realistic at scale. Rather than relying on fixed digital journeys, AI can help interpret free-text input, structure requests and prepare information before it reaches clinical or operational teams. This is where ideas such as AI clinicians or agent-style support are starting to appear in the research, focused on executing defined steps within a workflow and escalating to humans when risk or uncertainty increases
For many services, pressure on staff is driven less by complex clinical demand and more by high volumes of avoidable contact. When access routes are fragmented, it becomes harder to:
- Reduce repeat calls and repeated triage
- Route patients into the right pathway early
- Keep demand aligned with local service capacity
- Avoid unnecessary handoffs between channels and teams
When AI-supported digital services are designed around real patient tasks, they can reduce duplication and improve routing, while keeping accountability and oversight with clinicians and operational teams.
What teams are paying attention to
- Whether digital services are designed around real patient tasks
- How digital triage connects to local care capacity
- Where human support is still required in digital journeys
- How non-digital access routes are maintained
What to watch as this develops
- Increased use of the NHS App as a default access route
- Clearer boundaries between automated and human supported journeys
- Better alignment between digital access and local service availability
Focus area 3: Data infrastructure expanding, with AI built on top
AI in healthcare is increasingly being shaped by the data foundations underneath it. Rather than starting with new tools, attention is turning to how data is connected, shared and made usable across organisations:Patient journeys routinely cross trusts, care settings and professional boundaries. Fragmented data makes it harder to understand demand, plan services and coordinate care. Shared data infrastructure is increasingly seen as a way to address this by connecting systems rather than replacing them.
Patient journeys routinely cross trusts, care settings and professional boundaries. When data is fragmented, it becomes harder to:
- Understand demand across pathways
- Plan services with a clear view of capacity and pressure points
- Coordinate care when information sits in separate systems
The NHS Federated Data Platform reflects this shift, supporting system wide insight without forcing organisations to move away from existing systems. Early use cases focus on operational priorities such as discharge planning and elective care coordination, where connected data can support better decisions without introducing additional clinical risk.
What teams are paying attention to
- How quickly shared data platforms are adopted locally
- The quality and governance of the underlying data
- Staff capability to use connected data for insight
- How AI outputs are understood and trusted
What to watch as this develops
- Greater consistency in operational use cases across systems
- Increased scrutiny of data quality and lineage
- AI tools being layered onto shared data rather than siloed datasets
Focus area 4: Trust, safety and governance frameworks taking shape
Trust, safety and governance are becoming central to adoption as AI moves from experimentation towards routine use. The question is no longer whether AI can be used, but rather under what conditions it should be relied on in live services. Trust, safety and governance are what allow AI to be used with confidence in live services, because they make risk visible and manageable when tools influence prioritisation, access to care or operational decisions under pressure.
Healthcare operates in environments where accountability ultimately sits with people. If AI outputs cannot be explained, reviewed or challenged, they create uncertainty rather than confidence. Governance frameworks help make responsibility visible, clarify how risk is managed and ensure AI supports decision making rather than obscuring it.
National developments such as the National Commission on AI in Healthcare and emerging MHRA guidance point towards more consistent expectations. The emphasis is on transparency, auditability and human oversight, creating clearer guardrails for organisations moving beyond pilots.
What teams are paying attention to
- How AI governance aligns with existing clinical safety structures
- Whether AI outputs can be explained and audited
- How clinicians and operational teams are supported to oversee AI use
- Clarity around accountability when AI influences decisions
What to watch as this develops
- Governance requirements becoming more explicit at the procurement stage
- Greater focus on explainability and audit trails in deployed tools
- Clearer expectations around human oversight in operational use
Focus area 5: Using AI to support proactive risk stratification and preventive care
AI is increasingly being explored as a way to identify risk earlier. Rather than responding once demand has already surfaced in acute settings, services are testing how data driven approaches can help spot patterns that signal higher risk.
Unplanned hospital admissions place sustained pressure on frontline services. Earlier identification of risk creates opportunities for intervention in primary, community and preventative settings, improving outcomes for patients while easing demand on acute care.
The value of AI in this space lies in pattern recognition across populations. Insight can be surfaced at scale, but action still depends on pathways being in place to respond. Without that link, prediction alone does not reduce pressure.
What teams are paying attention to
- How predictive tools align with population health priorities
- Whether care pathways can act on identified risk
- How outcomes and service impact are measured
- How insight is shared across care settings
What to watch as this develops
- Risk models being tied more closely to funded interventions
- Increased use of AI supported insight in community services
- Stronger evaluation of whether early intervention reduces demand
Focus area 6: Moving AI from pilots to operational delivery at scale
As AI activity increases, a gap is emerging between experimentation and sustained impact. Many organisations have pilots underway, but fewer have embedded AI into everyday delivery across services.
Scaling requires more than technical success. Systems need to integrate, ownership must be clear and teams need support to adopt new tools as part of routine work. As tools move beyond pilots, ownership usually needs to sit with the teams responsible for running the service day to day, with innovation functions supporting rather than leading. Without this, benefits remain localised and difficult to sustain.
This highlights that AI adoption is as much about delivery, change and capability as it is about technology. Where AI is treated as part of service improvement rather than innovation alone, it is more likely to persist and deliver value.
What teams are paying attention to
- How AI initiatives align with service priorities
- Whether tools integrate into existing workflows
- How change and training are supported
- How success is measured beyond pilots
What to watch as this develops
- Pilots being retired or expanded based on operational outcomes
- Clearer ownership for AI tools beyond innovation teams
- Success measured against service impact rather than novelty
What to priortise as AI grows
Taken together, these building blocks point to a more mature phase of AI adoption in UK healthcare. The emphasis is shifting away from ambition for its own sake and towards careful, practical use in areas that genuinely affect delivery.
What stands out is not the pace of change, but the increasing focus on judgement. Where AI is introduced with a clear purpose, supported by trusted data and governed with care, it can reduce friction and help services use limited capacity more effectively.
By 2026, the organisations that see lasting value from AI will be those that treat it as part of how healthcare is delivered day to day, not as a separate initiative to manage alongside it. The challenge ahead is not whether to use AI, but how thoughtfully it is applied.
