YIDU TECH (02158)'s AI Clinical Assistant Boosts Specialty Diagnosis Efficiency with Nearly 1,000 Daily Calls

In an era where large models are reshaping industries, healthcare faces a critical challenge: How can general AI meet the rigorous demands of clinical practice, ensuring algorithms truly understand doctors and integrate into diagnostic workflows? This is the core hurdle for medical AI transitioning from "usable" to "practical."

Recently, Dr. Huang Nan, Senior Vice President and General Manager of Scientific Research at YIDU TECH (02158), and Dr. Li Linfeng, Vice President of Technology Innovation and AI Architect, shared the company's deep-rooted practices at the 2025 Annual Conference of the Digital Health Branch of the China Geriatric Health Medicine Research Association and Tsinghua Cardiovascular Training Program. Their focus: "Empowering Specialty Diagnosis with Large Model Clinical Assistants for Quality and Efficiency."

**From "General" to "Specialized"** The integration of AI and healthcare is accelerating. However, general large models face fundamental issues like "hallucination" risks, lack of transparency, and insufficient medical rigor when directly aiding clinical decisions. YIDU TECH's solution is a systematic reconstruction for clinical practicality, making AI's "thinking" controllable, trustworthy, and traceable.

Centered on its self-developed medical vertical model "YiduCore," the company built a dual-middle-platform solution combining "big data + large models," creating a robust digital foundation for hospitals. The AI platform’s core value lies in transforming raw model capabilities into "intelligent agents" that comply with medical standards, understand clinical workflows, and master specialty knowledge—ensuring AI’s real-world hospital integration.

**Seamless Integration into Workflows** For medical AI to gain trust in clinical settings, bridging the "trust gap" between practitioners and AI is key. YIDU TECH’s AI platform strengthens trust through three core capabilities: 1. A high-quality "knowledge hub" providing evidence-based professional support. 2. A one-stop medical AI operator library with 40+ intelligent operators for transparent, traceable processes aligned with clinical norms. 3. Lightweight, seamless integration into natural workflows for effortless use.

These capabilities ensure AI decisions are evidence-based and traceable while enabling doctors to adopt the technology effortlessly.

YIDU TECH’s "Clinical Copilot for Doctors," embedded as a lightweight plugin in workstations, integrates deeply into core scenarios like medical record documentation and treatment planning. It offers real-time AI assistance, solving the industry challenge of "how to make medical AI clinically viable and doctor-friendly."

To address diverse specialty needs, YIDU TECH developed 130+ scenario-specific agents covering medical record generation, assisted diagnosis, risk assessment, and patient education. This deep workflow integration has turned AI from an "option" into a "high-frequency tool."

**Transforming Specialty Diagnosis Practices** The ultimate value of technology lies in solving real clinical problems. YIDU TECH converts large-model capabilities into task-specific intelligent applications, reshaping diagnosis efficiency.

- **Cardiology**: The AI platform powers a cardiology guideline assistant integrating 20+ cardiovascular guidelines. It analyzes symptoms, ECGs, and lab results via natural language interaction, offering evidence-based decision support for time-sensitive scenarios like emergencies while boosting doctors’ query efficiency. - **Oncology**: In the highly specialized task of "TNM staging," the AI agent, embedded with staging guidelines, anatomical knowledge graphs, and doctor reasoning logic, achieves 90% accuracy—matching senior physicians—while reducing computational demands. - **Digital Specialists**: Collaborating with Wuhan Tongji Hospital, YIDU TECH created a digital thoracic oncology expert. This "digital twin" auto-accesses patient data for personalized consultations, combines 20+ guidelines with expert experience, and uses digital avatars to provide 24/7 professional advice.

At Sun Yat-sen University Cancer Center, AI tools powered by YIDU TECH’s dual-middle-platform earned high praise: The Clinical Copilot sees 800–1,000 daily calls, improves single-disease data reporting efficiency by 95%, boosts tumor staging completion by 5%+, and accelerates research efficiency by 90%, supporting over 3,000 projects. AI has evolved from a "novelty" to a trusted "digital colleague."

**Looking Ahead** As medical AI shifts toward "professional collaboration," its deep alignment with clinical workflows and doctor cognition marks a transition from "external tool" to "embedded capability." Moving forward, YIDU TECH will leverage its dual-middle-platform and "intelligent agent" matrix to advance precision and efficiency across more specialties, driving healthcare toward a smarter future.