The Pulse
Tairex Says Agent Hospital Has 54 AI Doctors Across 27 Clinical Departments
Tairex says its Agent Hospital virtual clinic includes 54 AI doctor agents across 27 clinical departments. The company’s multidisciplinary human-agent collaboration framework is intended to bring cross-specialty input into consultations whi
AI.info Team · Written with AI · fact-checked ·
Tairex introduced its Multidisciplinary Human-Agent Collaboration framework, or MHAC, in Beijing on October 21, proposing that AI agents representing different specialties contribute to a case while physicians retain the decision-making role. The company says the system can bring cross-specialty input into an initial consultation rather than waiting for a separate multidisciplinary team meeting.
Three ways to bring agents into a consultation
Tairex describes three arrangements: one doctor working with a personal AI counterpart; one doctor consulting several agents; and multiple doctors working alongside multiple agents. The company identifies the one-doctor, many-agent setup as the framework’s distinctive mode, allowing a physician to call on agents from their own or other specialties during an appointment.
Agents may be selected by the doctor or recommended by the system as a consultation unfolds. They can offer timely prompts or join a consultation when the physician initiates one, with the attending doctor making the final decision. Tairex says the system is designed to show the reasoning behind recommendations and record agent activity for review.
The launch announcement illustrates the intended value with a hypothetical case: an endocrinology agent suggests a beta-blocker for a patient with palpitations and asthma, while a respiratory-medicine agent flags a possible risk of bronchoconstriction from a non-selective beta-blocker. The example shows how the company expects agents to surface competing considerations; it is not evidence that the system has improved patient outcomes.
54 AI doctors, with real-world tests still ahead
Tairex says its Agent Hospital virtual clinic has 54 AI doctor agents across 27 clinical departments. Its October 22 PR Newswire announcement reports more than 96% accuracy on the MedQA dataset and says the simulated environment includes millions of AI patients. Those are company-reported figures; the announcement does not provide clinical-outcome results for the new MHAC framework.
The company says its simulated patients are generated partly from clinical guidelines and medical literature rather than simply reproducing medical records. It also describes a system in which AI doctors train in a virtual setting, while personalized doctor agents learn through collaboration with human physicians in clinical workflows.
Tairex says it plans to begin phased testing of multidisciplinary human-agent collaboration in real-world settings at partner hospitals, subject to regulatory and ethical requirements. The announcement gives no start date for those tests. That leaves the framework’s proposed workflow distinct from demonstrated performance in the new multi-agent setting.
Human doctors remain the decision-makers
The company’s framework says agents should follow medical rules and ethical constraints, reflect the practices of their associated doctors, and yield to human decisions when those principles conflict. Tairex also says its system does not permit decisions made entirely by agents without human-doctor involvement.
That boundary matters because the product is designed to add recommendations to an existing clinical decision, not to replace the clinician overseeing care. Tairex’s own next stated step is staged testing with partner hospitals; evidence from those deployments will show whether the system’s parallel opinions help doctors manage cross-specialty cases in practice.