Publications
Building our evidence base

+400 AI simulations

A JMIR Formative Research study (Mar 2026) evaluated SimFlow.ai in primary care consultation training across 70 UK teaching practices, with 47 survey responses representing 179 participants. Medical content was rated highly, with 97.8% judging scenarios clinically plausible, and educational value was positive. AI realism and feedback were more moderate, highlighting limits in conversational naturalness. The authors conclude SimFlow.ai is a promising supplementary tool for consultation skills training.

An evaluation in Education for Primary Care (Sept 2025) examined SimFlow.ai for Simulated Consultation Assessment preparation with GP trainees and educators (n=22), finding high ratings for clinical authenticity and educational value (median 4.5/5) and strong overall acceptability. The study reports improved accessibility and frequency of practice, significant support from senior educators for curriculum integration, and up to an 84% cost reduction compared with traditional actor-based methods—positioning SimFlow.ai as a practical, scalable complement to GP consultation skills training.

Published in Current Treatment Options in Psychiatry (April 2025), the paper highlights SimFlow.ai as an exemplar of AI for clinician training—citing hundreds of AI-driven simulated patients, including youth cases such as early psychosis, PTSD in a child refugee, and eating disorders. It explains how practising with these agents can sharpen diagnostic reasoning, communication, and decision-making in a low-risk setting, with post-session feedback on language and response effectiveness, while noting the need for further validation.

 
 

We are busy Working on the future

Ongoing project (pre-publication): In commercial partnership with Anglia Ruskin University and the University of Bath’s Centre for Applied Autism Research (CAAR).

ASIST—Autism Screening with Intelligent Supporting Technology—tests a SimFlow.ai conversational tool that guides adults (18+) through a short validated autism screening and generates a GP-ready Patient Clinical Form. It is not a diagnostic tool; the aim is to ease first steps, support referrals and help reduce delays to assessment.

Ethics reference: ETH2526-0428

 
 

Ongoing study: The University of Bristol will evaluate SimFlow.ai’s virtual patients against traditional actor-led simulations. The mixed-methods evaluation will compare realism, usability and educational value, examine feedback quality across scenarios, and generate recommendations for curriculum integration.

Take action,
or don't

Join us in the journey to transform your communication skills training.