Fast Blades and Heavy Swords: Navigating the Unformed AI Healthcare Arena

Original source: VCBeat, Author: Wang Hang

“How can AI in healthcare succeed?” a reporter asked.

“Healthcare is a vast and complex ecosystem. When you add AI into the mix, every step forward is fraught with pitfalls. Once you have navigated all the inevitable challenges, you will have largely completed the journey and be poised for success.” This was the response provided by Dr. Yu Zhong, Founder and CEO of Jinglun Century.

This scene took place at the 2017 Yangtze River Industry Forum (Autumn) and the Conference on Big Data and Artificial Intelligence in Healthcare. Among the attendees was Dr. Yu Zhong, former Chief Architect at AT&T, an aerodynamics expert, and a specialist in big data applications. He had previously participated in the “Save Galileo Satellites” project. After returning to China, he founded Jinglun Century, dedicating eight years to developing expert systems, navigating numerous pitfalls, and persistently overcoming countless challenges.

Just like Dr. Yu Zhong, many other entrepreneurs in the AI healthcare sector share similar journeys: some are just starting out, others are midway through, and a few are fortunate enough to have moved quickly, already establishing a certain “reputation in the field.” Yet they are still far from reaching their final destination. Within this “arena,” some wield “swift, invisible blades,” while others rely on “heavy, edgeless swords.”

Healthcare represents a multi-layered equilibrium among the state, enterprises, and individuals. The existing healthcare framework has failed to balance individual medical needs due to the severe stratification of the vast healthcare structure and the historically underdeveloped state of medical services in our country. Although the concepts of tiered diagnosis and treatment and smart healthcare were introduced relatively late, there is reason for optimism: over the past decade, we have witnessed a significant acceleration in healthcare capabilities.

According to SCI journal data statistics, from 1996 to 2016, China published 210,000 articles on artificial intelligence, while the United States published 80,000. The majority of China’s articles were published between 2007 and 2016, accounting for 90% of the total. In terms of these figures, China has currently surpassed the United States in the number of academic articles published in the field of artificial intelligence, representing a qualitative leap in capability.

Beyond technological capabilities, China’s demographic advantage has enabled access to medical data on a scale unmatched by any other country, and the impact of artificial intelligence on the current state of healthcare in China will be unprecedented.

Currently, artificial intelligence is primarily assisting junior physicians in training to improve the overall accuracy of diagnosis and treatment, reduce the cost of physician development, and comprehensively elevate China’s overall standards in clinical care and diagnostic interpretation. This represents not merely a preliminary reflection on AI applications in healthcare, but also a forecast of the ultimate trajectory of AI in medicine.

AI Healthcare, also known as Smart Healthcare. At its core, it fully leverages informatization, intelligence, and automation to enable proactive prevention, accurate diagnosis, rapid treatment, and guided rehabilitation, representing a transformation in the healthcare model.

Informatization, intelligence, and automation are all closely linked to the equipment systems or departmental structures within existing hospital frameworks. Medical imaging constitutes a critical component of hospital informatization, and radiologists demonstrate strong adaptability to intelligent technologies. Consequently, smart imaging has emerged as a key area for enhancing AI capabilities in healthcare, with most breakthroughs in AI-driven diagnostic proficiency focusing on this domain. Another significant area is the application of AI in treatment, where AI-enabled radiotherapy systems currently dominate the market. Following these are AI products designed for real-time follow-up and rehabilitation guidance, though such solutions remain relatively scarce at present.

Viewed from the perspective of the healthcare ecosystem, it is evident that the emergence of AI-driven medical products has been relatively slow and limited, underscoring the significant barriers to entry in healthcare entrepreneurship. In simple terms, a “winner-takes-all” scenario is unlikely to materialize in the healthcare sector. From the standpoint of business models, purely consumer-facing (C-end) products struggle to achieve scale, as hospitals and other institutional entities (B-end) remain the primary payers. On the product front, it is unfeasible to capture a large-scale market by applying a single standardized data system across diverse settings, further affirming the inherently hierarchical nature of the healthcare industry.

The ultimate landscape of the healthcare industry, particularly in AI-driven healthcare, is highly likely to be an oligopolistic structure characterized by a few large-scale enterprises and numerous small- and medium-sized enterprises, as noted by Jiang Tianjiao, Director of the Industrial Finance Department at Founder Securities and Head of Healthcare Industry Investment and M&A.

At this conference, the AI healthcare industry report released by VCBeat’s VBInsight outlined several key directions for future development, including virtual assistants, disease risk diagnosis and prediction, medical imaging, electronic health records, literature analysis, hospital management, smart devices, new drug R&D, health management, and genomics. Currently, companies are actively exploring and advancing in each of these areas.

Enable enterprises to become “swift operators” in the industry, leveraging speed advantages to rapidly rise, gain market feedback, and establish their market position. The fastest way to become a swift operator is to deliver medical capabilities most rapidly; the most efficient approach is to enter sectors amenable to rapid integration. Medical imaging and radiation therapy possess precisely these characteristics.

WingSpan Technology is a “swift and decisive” enterprise. Its clear strategic direction and significant market impact position it with strong potential to become a leading player in the medical imaging industry.

Although Yizhan has been advancing its AI initiatives for only two years, it had already established a highly internet-enabled and informatized platform prior to integrating artificial intelligence with medical imaging. Furthermore, as the exclusive distributor in China for GE Healthcare, an international medical equipment manufacturer, and leveraging its robust capabilities in resource integration and market expansion, Yizhan naturally achieved rapid implementation.

The reporter gained a deeper understanding of Yizhan Technology’s business model from its person in charge. Within Yizhan’s entire ecosystem, a comprehensive closed-loop of capabilities has been established. This ecosystem will continuously enhance its output and processing capacity as third-party medical imaging centers are rapidly deployed.

The entire ecosystem is built on advanced algorithms, leverages data from top-tier hospital departments, and employs highly informatized management systems to rapidly address the weakness of primary healthcare capabilities. Meanwhile, Wingtech has partnered with Yingling to establish the largest radiologist platform, continuously supplying data and enhancing annotation capabilities within the ecological closed loop. The Wingtech team has named this model the “Flywheel Effect.” Each element in the ecosystem represents a flywheel; as these interlocking flywheels incorporate new AI algorithms or data from imaging centers, the entire ecosystem operates with greater speed and momentum.

Yizhan is an exploratory enterprise in the field of AI-driven diagnosis, while Quanyu Medical is an exploratory enterprise in the field of AI-driven therapy.

Over the past two years, Quanyu Medical has been deeply engaged in the field of radiation therapy. Why choose this sector? One key factor is its potential for rapid and efficient implementation. The three major modalities for cancer treatment are surgery, chemotherapy, and radiation therapy. Surgery, by its nature, is difficult to integrate with artificial intelligence and internet technologies. Chemotherapy, on the other hand, faces a relatively chaotic landscape in China. Consequently, radiation therapy has emerged as the preferred direction for the development of oncology care in China.

Third-party independent radiotherapy centers are common abroad, whereas in China there is a severe shortage of both equipment and personnel. Radiotherapy is, in fact, the treatment modality most closely integrated with the internet and IT; building on this digital infrastructure, AI capabilities can be further incorporated.

The process of radiotherapy requires a comprehensive treatment plan to ensure that electronic devices precisely target irregular, patient-specific tumors—a task at which AI excels. Furthermore, while most hospitals are equipped with radiotherapy devices, their utilization rates remain low. To accelerate implementation, remote collaboration systems are needed to enable senior physicians to support junior colleagues, allowing them to direct operations at subordinate hospitals as seamlessly as if they were using their own local equipment.

In addition to this system, Quanyu has developed the Cloud Intelligent Control device and improved its remote training system. With these three systems in place, hospitals at all levels can essentially ensure high-quality treatment planning and effective implementation, while AI addresses capability gaps within the workflow. As a result, departmental radiotherapy data and treatment plans have been revitalized under this rapidly advancing model.

By integrating core competencies and assets to align with the most suitable strategic direction, and by focusing intensely on market demands and pain points, a company can achieve rapid growth and eventually secure a prominent position in the industry. Nevertheless, many enterprises that prioritize substantial, long-term development possess an even broader vision.

In the internet era, business models have long been categorized as either “asset-heavy” or “asset-light.” Applying this framework to AI healthcare companies, developing expert systems is considered a highly “asset-heavy” approach due to the significant challenges in acquiring pathological data. Jinglun Century is a company that has dedicated eight years to deepening its expertise in expert systems. An even more “asset-heavy” model than expert systems is the platform-integration approach pursued by iFlytek Healthcare.

“A heavy sword has no edge; it attacks even without sharpness.” This phrase perfectly encapsulates the goal of AI healthcare entrepreneurs pursuing an “asset-heavy” model. Building such a model is a process of establishing formidable barriers to entry, which aligns with the inherent characteristics of the healthcare industry yet, to some extent, fails to meet the expectations of businesses or investors.

What Jinglun Century has been consistently engaged in can be summarized as “Intelligent Medicine,” representing a deep integration of the contents encompassed within both intelligence and medicine. After extensive discussions with numerous medical experts and assisting them in data cleaning and standardization, Jinglun Century has developed a problem-solving model from the experts’ perspective. This approach involves asking physicians: “As professionals, you must have encountered certain challenges during your practice; what specific problems are you most eager to resolve?” In response to these “problems” raised by doctors, we provide input on solution capabilities. For instance, starting from a diagnostic process, we examine global standards, identify unresolved medical issues, and analyze how China currently addresses such issues compared to practices abroad, particularly in Europe and the United States. Subsequently, we address a series of related tasks required to solve these problems, including data acquisition, study design, data organization, machine learning, and predictive modeling. This is an endeavor of considerable scope and sophistication.After eight years of dedicated effort, the current expert system has not yet achieved large-scale practical application. Such an outcome can be quite taxing for any enterprise. However, there is promising news: measures implemented by the National Health and Family Planning Commission (NHFPC) to facilitate implementation have renewed hope for asset-heavy AI healthcare startups. Having just navigated through significant pitfalls, we have now received a “sweet reward.” This journey represents not only the refinement of products and business models but also the mental fortitude of entrepreneurs.

The “heaviness” of expert systems lies in their substantial technical complexity and massive data requirements, whereas another “heavy” model involves platform-based strategies that integrate markets, plan distribution channels, and demand robust foundational technical capabilities. iFlytek, which gained widespread recognition through its breakthroughs in speech technology, is implementing such a strategic layout.

Recently, iFlytek and Anhui Provincial Hospital jointly launched the first AI-powered hospital. Within this AI-driven healthcare facility, iFlytek will integrate various artificial intelligence technologies into every aspect of hospital operations. The company’s initial focus is on establishing an AI-assisted diagnosis and treatment center, with the goal of empowering primary care hospitals through AI technology. Through this collaboration between iFlytek and Anhui Provincial Hospital, 41 county-level hospitals across Anhui Province have been connected. In China, there is a critical need for such advancements in primary care institutions.

In iFlytek’s vision, this AI-assisted diagnosis and treatment center will become an open platform. While many of the technologies developed by iFlytek will be integrated into the platform, the company adheres to the principle that the healthcare industry involves multidisciplinary teams and diverse diagnostic methods. Therefore, iFlytek aims to incorporate technological capabilities from more companies into its open platform, creating an inclusive ecosystem for delivering services. This essentially means that iFlytek will invest substantial human, material, and financial resources to advance the implementation of medical capabilities and enhance its AI prowess. Currently, iFlytek’s AI research team comprises nearly 1,000 members and has achieved significant breakthroughs across various medical domains. If iFlytek’s open-platform model can be rapidly implemented, it will not only position iFlytek as a leading enterprise in the healthcare sector but also serve as a boon for numerous AI healthcare entrepreneurs seeking to deploy their solutions.

AI in healthcare has been designated as a key priority in the New Generation Artificial Intelligence Development Plan, emerging as a hot sector that has attracted intense capital investment and strategic positioning. However, the trajectory of AI in healthcare is not akin to internet startup ventures, where capital and market conditions drive exceptionally rapid growth. The inherent challenges in commercializing the healthcare industry have kept investors remarkably clear-headed.

Jiang Tianjiao, Managing Director of the Industrial Finance Department at Founder Securities and Head of Healthcare Investment and M&A, offered an interpretation of capital’s perspective on AI in healthcare at the 2017 Yangtze River Industry Forum: “There are many angles from which to categorize and understand AI in healthcare, such as from the perspective of the healthcare industry, the industrial value chain, or the service value chain. However, when considering AI in healthcare comprehensively, numerous issues exist. ‘Difficulty in distinguishing near-term from long-term prospects and differentiating genuine innovations from hype’ may well reflect the reality.”

Furthermore, Mr. Jiang raised several critical questions: Are AI products and technologies truly the absolute driving forces? From a commercial perspective, are the value propositions clearly defined in areas such as health management, smart wearable devices, smart hospital management, intelligent triage, and disease risk prediction? Is there a risk of addressing pseudo-demands? Factors such as usage frequency and willingness to pay are critical determinants for successful implementation. If a product features both low usage frequency and low average transaction value, its viability is questionable. Additional challenges include the difficulty of data acquisition, prolonged implementation cycles, and the limited patience of capital investors.

From a business perspective, the process for solving problems involves product implementation, business model implementation, and profitability realization. Product implementation requires addressing genuine needs and leveraging technical capabilities. This is followed by the implementation of the business model, which involves testing revenue streams and scaling up income generation. Finally, there is the realization of profitability; however, no highly refined profit model has yet emerged, and entrepreneurs in the AI healthcare sector are still continuously exploring viable paths. The patience cycle for capital trends typically spans about two years: the first two years are dedicated to demand validation and technological implementation; the next two to three years focus on testing revenue streams and achieving scalable replication; and another two to three years are needed to generate net profits and extend profit models. Projects that follow this trajectory are considered successful by investors. Clearly, AI healthcare cannot advance in such a manner.

It is encouraging that the government’s growing recognition of AI needs will also allow capital to gain a clearer understanding of the development trajectory of AI in healthcare. As evidenced by various initiatives led by the National Health and Family Planning Commission (NHFPC) to implement AI-based medical products, the state aims to address existing healthcare challenges through market-driven economic mechanisms. The aforementioned companies—Jinglun Century, Yizhan Technology, and Quanyu Medical—are all included in certain implementation plans under the NHFPC.

The landscape of AI in healthcare has yet to take shape, and the current phase of entrepreneurs racing forward together will persist for some time. Some founders describe their exploration of the medical industry as akin to navigating a video game, battling monsters level by level. Whether wielding swift blades or heavy swords, regardless of their current speed or achievements in clearing these stages, all are fighting to resolve the core challenges plaguing healthcare.

As Li Datao, founder of VCBeat, stated: “We are at the starting point of constructing a new medical ecosystem. In the future, healthcare services will be characterized by artificial intelligence, accessed within a context of continuous integration between human–computer interaction and emerging sociocultural dynamics.”

Currently, artificial intelligence is far from having the capability to transform healthcare models; it remains merely in an auxiliary role. Mitigating problems and fundamentally changing them are two distinct concepts. At present, telemedicine represents the most effective and readily implementable model capable of driving changes to existing diagnostic and treatment paradigms. However, realizing future visions for healthcare transformation will likely require further advancements in technological capabilities as well as continued refinement of relevant national policies.

In the AI era, healthcare entrepreneurs are continuously experimenting, repeatedly reconnecting technology, data, devices, people, and other elements, while restructuring and recombining resources. We are confident that we can excel in seizing the strategic high ground of artificial intelligence as outlined in the new generation AI initiative.

However, the foundation of AI in healthcare remains commerce, and ultimately, we must return to contemplating the essence of business. Amidst this profound societal transformation, the landscape will undoubtedly be shaped by intense competition and upheaval, where only the true heroes will emerge.