JXR has launched a self-developed AI system called "IVF Co-Pilot" at Sichuan Jinxin Xinnan Women and Children's Hospital (Bisheng Campus), marking the first clinical AI assistant for doctors in the assisted reproduction field in China. The system, developed by the company's subsidiary Chengdu Jinxin Information Technology, targets the full process of fertility treatment, with an initial focus on assisting evaluations of IVF stimulation protocols and medication initiation.
Assisted reproduction involves a continuous, dynamic, and highly personalized clinical pathway. The start of stimulation is a critical point where doctors must assess ovarian reserve, baseline hormones, past stimulation responses, and current ultrasound and test results to determine the protocol and starting medication. The challenge is not a lack of information, but rather providing effective references within limited time from vast historical cycles and professional data to support doctors in improving success rates and reducing patients' financial and psychological burdens.
"On one hand, the system relies on an IVF knowledge base and hospital historical cycles to offer professional and clinical experience references for doctors. On the other hand, clinical algorithms and a local large language model collaboratively analyze two types of information to generate comparable, traceable, and verifiable auxiliary opinions that are clearly integrated back into the doctor's workflow," said Wang Weipeng, head of R&D at Jinxin Technology.
According to a white paper by Frost & Sullivan, JXR is the largest integrated assisted reproduction service platform in China by number of egg retrieval cycles. In 2025, the company recorded over 23,000 cycles in China, with cumulative egg retrieval cycles exceeding 300,000. Fewer than 20 assisted reproduction institutions nationwide perform over 5,000 cycles annually. The "National Assisted Reproduction Report 2026" indicates that leading Chinese platforms have success rates among the highest globally.
"The vast cycle data and successful cases form the foundation for developing this intelligent system. Under legal and data security requirements, the system can screen similar cases and treatment plans from professional databases, which are then summarized, compared, and analyzed by the local model to assist doctor decision-making," Wang added.
"It's like a co-pilot in a commercial aircraft cockpit or a co-driver holding a route book in a rally race: it doesn't replace the main driver but provides continuous necessary support. For clinical experts, it presents key indicators and references at appropriate points. For junior doctors, it offers audited, traceable best practice references. All treatment plans are ultimately decided by doctors based on complete medical records and patient conditions," Wang explained.
Wang further stated: "We are building not a model that 'talks like a doctor,' but an AI capability that doctors can use, check, and trust. The value of medical AI depends not only on model capability but also on whether it appears at the right work nodes."
Comments