Artificial Intelligence (AI) Scribes in Primary Care: A Multi-Country, Mixed-Methods, Simulation-based Workshop Study

Background: AI-powered Ambient Voice Technology (AVT) is increasingly being adopted in primary care to support General Practitioners (GPs) with clinical documentation. However, their implementation and influence on care delivery, including the implications of interoperability and cross-system deployment, remain insufficiently explored across different healthcare contexts. This study aims to explore perceptions of the implementation needs and impact of AVT on primary care practice across diverse healthcare systems.<br><br>Methods: A convergent mixed-methods design was used, combining a structured questionnaire with a facilitated group discussion guided by modified Nominal Group Technique principles. Data were collected during an in-person workshop in an international conference, which started with a simulated primary care consultation using an AVT tool. Quantitative data were analysed descriptively, while qualitative data were analysed using a combined deductive–inductive analytical approach. Findings were integrated to generate overall study outcomes.<br><br>Findings: Fifty-four healthcare professionals participated, of whom 83% were GPs, practising in 21 different countries. Most participants (n=46) expressed an intention to use AVT, and approximately four in five perceived them as easy to use without substantial technical support. AVT was viewed as potentially improving efficiency, patient-centredness, and the overall consultation experience, while reducing administrative workload. Participants shared concerns related to ethical, governance, and regulatory requirements across healthcare systems, as well as the variability in electronic health record integration. Patient safety was a key consideration, particularly regarding documentation accuracy and the continued accountability of GPs for final clinical records.<br><br>Interpretation: AVT is widely perceived by primary care clinicians as a potentially transformative intervention for both patients and clinicians. However, its safe, equitable, and effective use depends on context-specific implementation, robust governance frameworks, and careful integration within existing primary care systems.