A system-level analysis of challenges and strategies for scaling artificial intelligence in healthcare: A qualitative study of the NHS AI lab
Hajar Mozaffar, Robin Williams, Stuart Anderson, Kathrin CresswellObjective
Despite Artificial Intelligence’s (AI) promise in healthcare, achieving widespread and lasting adoption remains a global challenge. While pilot studies show positive impacts, the reasons for limited uptake in practice are underexplored. This paper examines challenges hindering AI scaling in healthcare and develops a system-level explanation of why these challenges persist, alongside strategies to address them.
Methods
As part of a wider evaluation of England’s National Health Service (NHS) AI Lab, we conducted in-depth semi-structured interviews, documentary analysis, and observations of related meetings and events between April 2024 and April 2025. Data were analyzed using reflexive thematic analysis, followed by a higher-level conceptual synthesis.
Results
We identified 15 interrelated challenges spanning market forces, healthcare adopter dynamics, and macro-environment challenges. Challenges included unclear pathways from pilots to adoption, fragmented regulatory and evaluation processes, limited funding for scaling, procurement and organizational capability gaps, uneven digital maturity, unclear partnership models, uncertain returns on investment, and difficulties navigating fragmented healthcare systems. Rather than operating as discrete issues, these challenges reflect four underlying and interdependent system-level mechanisms: fragmented innovation pathways, fragmented governance and lack of system orchestration, organizational transformation deficits, and market and incentive misalignment. Together, these mechanisms help explain why AI innovations frequently struggle to progress from pilot experimentation to routine and system-wide adoption.
Conclusions
Our findings shift the focus from isolated challenges to the system-level dynamics through which challenges to AI scaling emerge, interact, and persist. The study contributes an empirically grounded explanation of the pilot-to-scale gap and highlights how governance, organizational, market, and innovation dynamics jointly constrain scaling. Building on these findings, we propose a set of coordinated strategic responses that emphasize system orchestration, organizational transformation, coherent innovation pathways, and continuous evaluation across the AI lifecycle.