AI-powered natural language processing tool boosts complaint detection by 95 per cent

Entrant: Buckinghamshire Healthcare NHS Trust
Category: HSJ Awards 2025 / Acute Sector Innovation of the Year
Award: Shortlisted SynopsisQuail is an innovative AI-powered solution transforming how NHS trusts process and learn from patient feedback. By applying advanced natural language processing to complaints, PALS enquiries, and incidents, Quail automatically identifies themes, patterns, and improvement opportunities that would otherwise remain hidden in unstructured data. Initially built and implemented at Buckinghamshire Healthcare NHS Trust, the system revealed critical insights across a range of clinical services. The technology improves operational efficiency while enabling more personalised patient communication. Quail’s success has driven adoption across multiple NHS organisations, demonstrating how AI can enhance patient experience, strengthen clinical governance, and support system-wide learning and improvement.AmbitionNHS acute trusts face an overwhelming challenge managing patient feedback through multiple channels, formal complaints, PALS enquiries, compliments and patient safety incident reports. These represent the authentic voice and experience of patients yet trusts struggle to extract meaningful insights from this unstructured data. At Buckinghamshire Healthcare NHS Trust, like most acute providers, staff manually processed thousands of feedback items annually, with inconsistent categorisation and limited ability to detect patterns or systemic issues. Our vision was transformative: harness artificial intelligence to automatically analyse patient narratives, identify themes and trends, and transform reactive complaint, PALS and Incident handling into proactive quality improvement. This approach fundamentally differed from existing practice, which relied on manual coding of limited categories and couldn’t identify connections across feedback channels. Where other systems focused solely on administrative case management, Quail was designed to generate actionable intelligence from patient stories. The innovation was particularly novel in three ways: 1) using advanced natural language processing specifically calibrated for patient safety and experience healthcare contexts; 2) connecting previously siloed data sources (complaints, PALS, compliments, incidents); and 3) providing unprecedented visibility of clinician-specific patterns to support professional development. Solution development followed a rigorous co-design methodology. The development team conducted extensive user research with patient experience teams, interviewing staff about workflow challenges and observing complaint processing. Technology development started with building a proof of concept using natural language processing with a Power BI front end to test value and benefits before moving to the more powerful Federated Data Platform, utilising large language models with an enterprise-grade analytics interface for enhanced capabilities and scalability. Patient involvement was comprehensive and ongoing. We presented the concept to Buckinghamshire’s Patient Experience Group, which includes patient representatives, who provided critical input on ethical considerations and potential benefits. Patient safety charities and patient partners were consulted regarding how the technology could enhance safety learning. Efficacy testing combined technical and user-centred approaches: •   AI accuracy validated against expert human classifications •   Comparison testing between Quail and human analysts •   User acceptance testing with staff •   Sentiment analysis of AI-assisted vs. manual responses •   Tracking of operational metrics In a powerful validation, Buckinghamshire’s CEO commissioned a “mystery patient” initiative to verify Quail’s findings on service accessibility issues. This independent assessment corroborated the system’s analysis, confirming the telephone access problems Quail had identified and reinforcing the value of its data-driven insights for targeted service improvements.OutcomeQuail has delivered transformative impacts on patient care and organisational performance, generating both quantifiable improvements and qualitative benefits that extend across multiple domains of healthcare delivery. At Buckinghamshire Healthcare NHS Trust, Quail’s AI-powered analysis revealed critical blind spots in patient safety reporting. The system identified 44 end-of-life complaints over 21 months where manual methods had found only 2a 95% improvement in detection. This discovery uncovered previously unknown issues with DNR CPR consent forms, directly influencing the trust’s end-of-life care strategy and improving care for vulnerable patients. The impact on operational efficiency has been substantial: •   12% improvement in 25-day complaint response rates •   80% reduction in time spent preparing governance reports •   Time to identify clinician-specific patterns reduced from days to seconds •   At NHS Frimley, complaint response drafting time reduced from 3 hours to 1 minute per letter •   32% decrease in complaints breaching the 60-day deadline These efficiency gains directly translate to improved patient experience through faster resolution of concerns and targeted service improvements. As Neil Macdonald, CEO of Buckinghamshire Healthcare NHS Trust, states: “Quail has transformed how we understand and respond to patient feedback. The AI-powered insights have allowed us to redirect resources more effectively, respond to concerns faster, and make targeted improvements that matter to our patients.” Perhaps the most significant impact came from Quail’s analysis of service accessibility in Cardiology, which revealed that complaints represented only 1% of actual patient concerns, with PALS enquiries providing the complete picture. This insight prompted targeted interventions including additional administrative staffing and improved telephony systems. Patients have reported significant improvements in their ability to contact departments, reducing anxiety and improving access to care. When presented to the Patient Experience Group, which includes patient representatives, the system received enthusiastic support. Patient partners valued the dual benefits of higher quality, more timely responses to individual complaints and the more accurate theming of issues. They particularly appreciated how proper categorisation of feedback could help identify systemic problems that affect many patients but might otherwise remain hidden. The group recognised that by properly identifying themes across patient experiences, the trust could implement targeted improvements that would benefit entire patient populations rather than just addressing individual complaints in isolation. By redirecting approximately 4 FTE of staff time annually toward quality improvement initiatives, Quail delivers substantial value for money through both direct savings and prevention of recurring issues, creating a positive return-on-investment cycle that benefits patients and staff alike.SpreadQuail’s innovative approach to patient experience management has generated substantial interest across the NHS through systematic knowledge sharing. Learnings from Buckinghamshire and Frimley implementations have been disseminated through blog posts with the Patient Safety Learning Charity and with detailed case studies shared across interested organisations. Buckinghamshire have also hosted several site visits demonstrating how Quail transforms unstructured patient feedback into actionable intelligence. Direct replication has occurred across diverse environments: Buckinghamshire Healthcare NHS Trust (acute care), NHS Lothian in Scotland (different regulatory framework), and NHS Frimley ICB (specialised response automation). This expansion continues with an upcoming pilot in NHS Wales and implementation at Berkshire Healthcare NHS the first mental health trust to adopt demonstrating versatility across fundamentally different organisational types. A key strength driving Quail’s spread is its universal applicability across all clinical specialties and departments. The system’s flexible theme taxonomy adapts to the unique language and concerns of diverse clinical areas from maternity to emergency care, from surgery to mental health services. This cross-specialty application ensures insights contribute to organisation-wide improvement rather than remaining siloed. The technical scalability of Quail is significantly enhanced by its integration with the NHS Federated Data Platform (FDP). This architecture provides a standardised implementation path that dramatically reduces technical barriers to adoption across the NHS ecosystem. As FDP becomes more widely deployed, Quail can be rapidly provisioned as a module within this existing infrastructure, eliminating many integration challenges that typically slow innovation spread. Quail’s modular design allows trusts to implement components addressing their most pressing needs first whether that’s theme identification, clinician pattern analysis, or response automation. This flexibility makes the solution suitable for organisations at different digital maturity levels. Ongoing conversations have explored applying Quail’s aggregate views across larger geographical locations such as Integrated Care Boards or regions to identify patterns spanning multiple organisations. This expanded perspective would enable system-wide learning and coordinated improvement efforts for issues affecting multiple providers along patient pathways. By aggregating anonymised trend data at this level, healthcare systems could proactively address emerging risks and share effective interventions more systematically, representing the next frontier in scaling this transformative approach to patient-centred improvement.ValueQuail has fundamentally transformed patient experience by enabling faster, more thorough responses to concerns while simultaneously identifying systemic issues that affect broader patient populations. The impact on patients is multifaceted: complaints are acknowledged and addressed more rapidly (12% improvement in 25-day response times), responses are more comprehensive through AI-assisted drafting, and most importantly systemic issues are identified and resolved before affecting additional patients. The discovery of previously undetected end-of-life care issues has directly improved care for vulnerable patients through targeted interventions in DNR CPR consent processes. Similarly, identifying widespread telephone access problems in Cardiology led to service redesign that benefited thousands of patients who previously struggled to reach the department, many of whom never formally complained but experienced significant frustration. For staff, Quail has revolutionised their ability to deliver high-quality care and service. Patient experience teams now spend significantly less time manually processing complaints and more time implementing meaningful improvements. Clinical teams receive regular insights about their services rather than only hearing about problems through escalation, enabling proactive quality improvement. The Chief Medical Officer now has unprecedented visibility of clinician-specific patterns, transforming the medical appraisal process from anecdotal feedback to data-informed professional development. Most importantly, staff report feeling empowered to address systemic issues rather than continually managing individual complaints without addressing root causes. Efficiency improvements have been substantial and directly contribute to value for money: •   480 hours of administrative time saved in a 7-week period at Frimley (equivalent to 2 WTE) •   32% decrease in complaints breaching the 60-day deadline •   80% reduction in time spent preparing governance reports •   Response drafting time reduced from 3 hours to 1 minute per letter These efficiency gains translate to approximately 4 FTE of staff time annually redirected from administrative processing to quality improvement activities. The financial return encompasses both direct savings in administrative time and prevented costs through early identification of systemic issues before they generate additional complaints or incidents. The value proposition extends beyond mere cost savings to include improved care quality, enhanced patient experience, and more efficient use of clinical and administrative resources. By automatically generating intelligence from unstructured feedback, Quail transforms what was previously seen as an administrative burden into a strategic asset for quality improvement, delivering exceptional value for money while meaningfully improving outcomes for patients and staff alike.InvolvementQuail’s success stems from an innovation culture deliberately cultivated at Buckinghamshire Healthcare NHS Trust, where leaders recognised that meaningful improvement requires both technological innovation and cultural transformation. The executive team established innovation as a strategic priority, creating forums where frontline staff could share patient experience challenges and collaborate on solutions. This top-down commitment, coupled with bottom-up engagement, created fertile ground for Quail’s development and adoption. Patient involvement was woven throughout the project lifecycle. The initial concept was presented to the Patient Experience Group, where patient representatives scrutinised the proposed approach and provided critical feedback on ethical considerations around AI use in healthcare. These representatives became active participants in discussions about how technology could enhance rather than replace human judgment in complaint handling. To ensure meaningful improvements, the team analysed historical patient complaint data to understand themes and pain points from the patient perspective. This analysis informed the theme taxonomy development and ensured the system would capture issues most important to patients. Patient safety charities, including Patient Safety Learning were consulted to provide independent feedback on how the tool could support broader safety improvement initiatives and ensure learning translated into tangible safety enhancements. The Parliamentary and Health Service Ombudsman reviewed the approach for alignment with national complaint handling standards, while the NHS England Patient Voice and Insight team participated in workshops to ensure the technology enhanced rather than diminished the patient voice in quality improvement. This multi-faceted engagement kept patient perspectives central throughout implementation. Partnership working was essential to success. Buckinghamshire Healthcare NHS Trust and Quantium established a collaborative relationship that transcended the traditional client-supplier dynamic. Quantium embedded team members within the trust during critical phases, working alongside staff to understand challenges and rapidly iterate solutions. This integration enabled development of a solution tailored to NHS requirements rather than imposing a standard commercial product. When expanding to NHS Frimley, the partnership model evolved into a knowledge transfer approach where lessons from Buckinghamshire informed implementation strategies. Frimley’s complaints team participated in iterative testing of response drafting functionality, providing feedback that refined the AI’s capabilities through weekly refinement sessions. This comprehensive involvement strategy created the conditions for successful implementation and sustainable improvement in patient experience across multiple NHS organisations.