Where to start

The integration of artificial intelligence (AI) into the healthcare sector faces a persistent structural challenge: while numerous single-point tools emerge, they often fail to navigate the complexities of the industry's full landscape. As general-purpose large language models enter the medical field, most applications are confined to fragmented uses like diagnostic support or smart customer service, lacking deep integration across the entire chain—from pharmaceutical distribution and business decisions to end-user services. Data silos, disconnected scenarios, and high compliance hurdles have left AI's industrial deployment struggling to move beyond hype.

During the 2024 West China Pharmaceutical Conference (Xipu Hui) on August 13, SINOHEALTH TECH (02361) Chairman Wu Han offered a direct assessment: China's health industry is perched on a fault line between the ebb of old growth drivers and the rise of new ones. The critical question is whether AI can evolve from a mere "tool" into a true "operating system," enabling the industry to bridge this gap.

Why just 10 ASX 200 shares?

When asked about recent deployments in the AI agent space, Wu Han outlined a clear roadmap. Built upon two core models, a comprehensive AI-powered agent platform tailored for the healthcare sector has already taken shape, with applications flourishing across five key scenarios. The business decision-making agent is a particular focus for pharmaceutical and medical supply companies, enabling them to analyze and decide using a unified data language and a single analytical framework, with conclusions generated from massive, integrated datasets for robust support. The R&D decision-making agent functions as a 24/7 intelligence officer, tracking global clinical trial intelligence and progress to offer comprehensive support for pharmaceutical R&D strategy.

In clinical settings, the integration of hardware and software brings tangible changes. Wu Han highlighted the "AI Medical Voice Card," a hardware device that captures real-time conversations between doctors and patients, converting them into medical documentation. This embeds the doctor's agent capabilities into the hardware, offering functions like patient management, evidence-based support, and research assistance. The pharmacy agent's core is the member intelligence module, built on the Zhuomuniao medical large model, merging pharmacy membership management with marketing capabilities to support the transition from "pure drug sales" to a "health station." Additionally, a future-oriented health management agent is positioned as a family's 24/7 health expert, using consumer health reports, diagnostic data, and data from wearable devices to deliver real-time, comprehensive health advice covering exercise, nutrition, and diet. It will later integrate with physician databases for precise referral guidance.

Given the vast scope of the healthcare industry, targeting full-scenario coverage raises the question of whether it's overly ambitious. Wu Han explained that many current AI companies produce excellent single-point products, but their limited functionality prevents deep integration with industry scenarios, and a lack of high-quality medical data often leaves them struggling independently. SINOHEALTH TECH's solution is to build an AI ecosystem center as an industry connector, merging the technological capabilities of the AI world into the healthcare industry's AI application system. "Through a 'one foundation + five scenario agents + one ecosystem connector' architecture, we aim to create an AI operating system specifically for the healthcare industry," he added.

Meanwhile, the pharmaceutical industry is in a transition phase from strategic blueprint to industrial reality. Companies are simultaneously navigating ongoing volume-based procurement (VBP) expansion, rapid growth in out-of-hospital markets, and a shift toward more precise marketing. When asked how the AI strategy aligns with these industry cycle changes, Wu Han stated that the most direct impact of VBP is a sharp decline in gross margins, while the extreme fragmentation and significant regional differences in the out-of-hospital market cause many companies shifting from hospital to retail channels to experience "cultural shock." In this context, the ability to make precise, efficient, and low-risk decisions becomes critical.

In the broader industry picture of retail terminal transformation, the evolution of pharmacies is particularly noteworthy. Wu Han did not avoid this topic, noting that pressures like declining foot traffic, falling margins, and weak member monetization have persisted over the past three years. However, he argued this isn't an industry decline, but rather a sign that old business models no longer match new health needs. China's 600,000-plus pharmacies form the most dense network of health service terminals with the highest frequency of face-to-face consumer interaction. Their future role should shift from "selling drugs and products" to becoming "health stations" for community residents, acting as a community health gateway.

Compelling data supports this: a diabetic patient's annual health spending totals about 10,000 to 13,000 RMB, but only about 2,000 RMB is spent on drugs at the pharmacy. The potential in areas like adjunctive therapies, nutritional supplements, and health services is far greater than merely selling medications—value that was previously overlooked.

It's also noted that to support its business growth, SINOHEALTH TECH is attempting to build a dual ecosystem. First, it is integrating high-quality AI technologies suitable for pharmacy scenarios to create a complete AI technology ecosystem. Second, leveraging its conference platform, it is aggregating resources from pharmaceuticals, devices, health services, medical services, and commercial insurers to build a service system capable of meeting consumers' full-cycle health needs, providing ecological resource support for pharmacies to build lifelong health management capabilities.

Regarding this strategic transformation, Wu Han firmly concluded, "When one window closes, another opens. What we need to do is help the industry find that door faster."