From Dialogue to Feedback: A GMC-Aligned Generative AI Simulation Platform for Clinical Communication Training

Healthcare education requires repeated practice in consultation and history-taking skills, yet standardised-patient programmes and faculty-led feedback remain resource-intensive and difficult to scale. This paper presents an integrated generative AI platform for clinical consultation skills training. The platform combines structured persona and case authoring, text- and voice-enabled virtual patient interaction, and transcript-linked formative assessment informed by professional communication expectations from the General Medical Council (GMC) and Medical Licensing Assessment (MLA) guidelines.

We describe the platform architecture and report an ethics-approved expert workshop evaluation focused on the consultation experience. The evaluation involved 12 senior clinicians and medical educators who used the platform to complete structured consultation scenarios and then provided quantitative ratings and free-text feedback. Participants rated the simulation positively for clinical plausibility, internal consistency, clarity of language, respectful behaviour, perceived safety, and usefulness for structured history taking. They also identified current limitations, particularly occasional mechanical phrasing, uneven emotional range, weak nonverbal representation, and the need for stronger validation across learners and clinical scenarios.

Taken together, these findings suggest that the platform can support scalable clinical consultation-skills practice as a supplement to human-delivered communication teaching, particularly when embedded within curricula and paired with transparent GMC- and MLA-aligned feedback and educator oversight. The work should be interpreted as an early expert evaluation rather than a full validation study, with future work required to assess learner outcomes, automated assessor reliability, deployment costs, and integration into existing educational workflows.