Assuta Develops AI Doctor Assistant to Support Physicians in Patient Care
Assuta Medical Centers, in collaboration with Microsoft, has initiated the development of a personal artificial intelligence (AI) agent designed to assist doctors before, during, and after patient visits. This AI system will analyze medical records to provide physicians with concise summaries of previous diagnoses, medications, tests, and treatments prior to consultations, aiming to reduce time spent searching through extensive documents and minimize the risk of missing critical information.
During patient appointments, with appropriate privacy approvals, the AI will transcribe conversations between doctors and patients and generate a comprehensive draft medical document including patient complaints, medical history, examination findings, physician impressions, and recommendations. Physicians will review and approve these documents before they are added to medical records. The AI is intended solely as a decision-support tool, with medical responsibility and judgment remaining with the doctor.
Post-visit, the AI agent will help identify needs for follow-up tests or appointments, assist in scheduling based on medical necessity and system availability, and potentially expand its functions along the care continuum under strict medical oversight and data security protocols. The project addresses a major challenge in healthcare: the significant time doctors spend on documentation and administrative tasks, often extending beyond clinic hours. The AI is expected to save hundreds of physician hours monthly, enabling shorter visit durations without reducing patient interaction time, increasing patient throughput, and reducing wait times.
Developing this system involves extensive training of the AI to understand medical documents, summaries, test results, and terminology while preserving patient privacy by removing identifying information. Medical teams continuously evaluate the AI's outputs to correct errors and improve information presentation. Preventing typical AI errors such as omission, patient data confusion, or fabrication is a key focus, ensuring that AI-generated content undergoes human review before integration into medical records.
The rollout will be gradual, starting with a limited pilot within a year to assess accuracy, safety, time savings, and acceptance by doctors and patients. Alongside the AI agent, Assuta and Microsoft plan to transition to Microsoft’s cloud infrastructure and establish an organizational platform for developing and deploying additional AI agents in medicine, service, and operations, all under stringent privacy and data protection standards.
