Croydon Health Services NHS Trust inaugural AI strategy has set out a range of use cases from HR processes to clinical coding, and from contract management to diagnostics and clinical decision support.
The trust begins by weighing the benefits and risks of AI use across the organisation, pointing to opportunities to personalise care and speed up accurate diagnosis; but also acknowledging the potential risk of bias or hallucinations, and the risks of over-reliance deskilling staff.
Seven principles will be embedded to align with responsible AI use, it states, keeping it clinically led and outcomes-focused, with every use having a named sponsor and agreed outcome metrics; with regulatory-grade evidence required; open standards and interoperability with cloud-first and open APIs; and with human-in-the-loop remaining the norm to give clinicians and staff ultimate accountability. The NHS T.E.S.T framework will be used to assess and ensure all technologies in use, the trust continues, and plans are to follow the NHS Data Saves Lives strategy, with AI now included in the DPIA.
A structured approach to AI will be split across clinical effectiveness and efficiency, operational productivity, and corporate services, and funding will be sought through frontline productivity and from other sources to support activities. An ethics governance framework and training materials will be developed to guide staff use of AI within the organisation.
According to the trust, priority projects to begin working on during this time include ambient voice technology in ED and outpatients, RTT validation automation using federated data platform products, the use of Patienteer to improve timeliness and quality of clinical decisions and patient flow, AI-assisted prostate cancer screening, and process automation in finance and HR.
Into 2027/28 and beyond, the focus will shift to building on work with Copilot to reduce administration time, and developing an AI and automation centre of excellence with support from a range of clinical and corporate teams to help identify, rapidly adopt, and scale AI technologies. AVT will be scaled with a focus on community services, workflows will be automated in corporate services including clinical coding, and work with radiology and pathology will explore areas where AI can support low-risk clinical activities.
An audit approach for AI systems including clinical safety cases will be developed and implemented beginning in Q4 2026/27, along with a bias audit framework and AI explainability guidance. AI training will be rolled out, and AI directorate champions will be appointed. The first annual AI progress report will also be published during this timeframe.
CHS also outlines initial use cases and their benefits, highlighting “blob analysis” of EPR data in promoting access to complete patient records and unstructured data to improve flow, as well as the potential for virtual monitoring to allow management of community and acute wound care, and for AI to be used in X-ray, CT stroke, and dermatology tele-advice. It also notes the lack of resources to effectively manage all trust contracts, and the possibility of using AI to review and pinpoint areas of under-recovery and penalty clauses not being actioned, and to support the publication of a procurement plan.
Elsewhere, the trust considers the use of AI in the production of board committee assurance reports, setting out guidelines such as the use of formally approved software, and comprehensive human review. “AI may support drafting and summarisation but must not replace evidence-based evaluation of the discussions, challenge, evidence reviewed and assurance provided by each Board Committee,” it says. “Reports should demonstrate critical analysis of the issues considered, the challenge provided by Committee members, the evidence reviewed, the level of assurance obtained and any residual risks. Narrative summaries alone should not be regarded as sufficient assurance reporting.”
Wider trend: Health AI
North East Ambulance Service NHS Foundation Trust is looking to “the road ahead” in its latest trust strategy, recognising the increasing role that tech, data, and AI will play in shaping urgent and emergency care delivery. “Doing more of the same is not a viable option,” the trust reports. “As demand continues to grow and evolve, relying on models designed for a different context would place increasing pressure on patients, colleagues and partners. NEAS must therefore adapt by design rather than by circumstance.”
With the publication of the latest budget for 2026, the government has reaffirmed its commitment on AI for productivity, increasing opportunities for sovereign AI, AI for biosecurity, and exploring quantum technologies. For health specifically, the government will launch the Chancellor’s AI Challenge Prize as an annual competition, with the first round to offer a top award of up to £2.5 million for promising solutions to priority healthcare challenges. An independent review has also been commissioned into NHS productivity and the opportunities offered by AI and technology, reporting in early 2027.
AI software is now installed on more than half of compatible NHS MRI scanners, the UK government has shared. The technology, along with the introduction of 125 new CT and MRI scanners, has been credited with playing a role in delivering one million more scans per year. AI-powered MRI technology is supporting services to carry out an average of 14 percent more scans “while maintaining image quality”, the government reports. This is in turn enabling the creation of thousands of additional appointments. It highlights that AI acceleration software is now installed on 54 percent of compatible NHS MRI scanners.
