tencent health ai big model commercialization accelerates with over 1,300 applications
2024-09-23
한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina
in the sixth year of its transformation from to c to to b, tencent health has explored a business model suitable for the vertical field of medical health.
at the recently concluded tencent global digital ecosystem conference, tencent health president wu wenda pointed out in his speech that tencent health's solutions based on big models have been implemented in more than 1,300 companies, hospitals and other institutions. "in the medical field, we have already done a lot of good ai application scenarios. the next 1-3 years will be a process of coverage, allowing more hospitals, pharmacies and companies to better serve customers with the help of ai."
figure 1: wenda wu, president of tencent health
as the internet enters its mature stage, a number of leading internet companies such as baidu, tencent, alibaba, and jd.com have begun to deploy in the medical industry, trying to create a second curve and betting on the next era after the internet.
from medical services to medical e-commerce, from biomedicine to big models...internet giants have almost spread their footprints to every corner of the medical industry. after more than ten years of growth and trial and error, the medical and health business has gradually become an important exploration area for the giants, but commercialization remains a major problem.
starting from the launch of wechat smart hospital in 2014, tencent's layout in the medical and health field has entered its tenth year. it has made many attempts in various directions, including tencent ai doctor and penguin clinic in 2016, tencent miying and tencent medical dictionary in 2017, and now after the start of to b transformation and the explosion of big models, tencent health's development route has gradually become clear, opening up a b-side paid business model.
wu wenda told health news, "the medical business cannot rely on tencent group's unlimited investment. a business that keeps losing money and needs subsidies is meaningless. we need to find a sustainable development model to expand our influence on society."
using big models as a starting point to explore b-side business models
at last year's tencent global digital ecosystem conference, the medical big model based on the hunyuan big model made its first public appearance, and relying on the big model, tencent health's previously released ai products, including smart question and answer, family doctor assistant, digital medical imaging platform, and drug discovery platform, underwent a new round of upgrades.
at that time, wu wenda said that tencent health "is more concerned about how to use big models and make good use of big models in real-world medical scenarios."
at the tencent global digital ecosystem conference two years ago, tencent group senior executive vice president and tencent cloud and smart industry group ceo tang daosheng clarified tencent's positioning and shift in the industrial internet, namely focusing on "key product tracks" and "building technical barriers" to assist partners in completing integration and delivery. the shift in the "being integrated" strategy is also reflected in the planning of the medical big model: encapsulating the big model as a production tool in different businesses and products, and then delivering it to partners such as isvs (independent software developers).
now, one year later, tencent health shared its interim results at the new digital ecosystem conference: its medical ai and related products have been deployed in more than 1,300 institutions, including hospitals, pharmaceutical and medical device companies, scientific research institutions, medical technology companies, etc., and has worked with more than 1,000 partners to serve more than 36,000 medical and health industry institutions.
figure 2: tencent health’s ai products are embedded in the entire physical examination process
according to zhang yu, vice president of tencent health, the products and services provided by tencent health to customers can be summarized into three scenarios: first, improving the level of intelligence in the r&d and production links; second, solving efficiency issues in internal office collaboration; and third, improving the construction of a digital marketing system to achieve performance breakthroughs.
taking the fully automatic cell morphology analyzer (ai reader) developed by tencent health and mindray medical as an example, after loading the ai algorithm and large model, the ai reading accuracy rate exceeded that of young doctors, reaching more than 95%, and the reading time was also shortened from the original 25 to 30 minutes to less than half a minute.
ai pre-diagnosis is one of the long-term directions of tencent health. after the release of the medical big model, ai pre-diagnosis products loaded with the big model have also begun to be implemented on a large scale.
as early as four years ago, in an exclusive interview with "health news consulting", wu wenda emphasized that tencent health must solve the core problems of users, strengthen the relationship between patients and doctors, use technical means to help patients find suitable doctors, and establish a trusting relationship.
at that time, tencent health was still exploring and trying to break through information asymmetry from the bottom up. after several years of experience in ai-guided diagnosis, the current ai pre-diagnosis system can conduct detailed pre-diagnosis answers with patients before the doctor's consultation, and provide information such as the chief complaint, past medical history, and medication contraindications in advance, thereby helping doctors improve the efficiency and accuracy of consultation.
at shenzhen people's hospital, tencent health's ai pre-consultation system is called more than 20,000 times per month. "compared with other ai products, the number 20,000 times may not be a lot, but the most important thing for us is what this number represents. the ai pre-consultation system has changed the patient communication experience 20,000 times," said wu wenda.
from a doctor in a small australian desert town to the president of tencent health, wu wenda has always had ideals. in an exclusive interview four years ago, he emphasized that tencent health "is not in a hurry to make quick money." today, he once again expressed his views in the interview, thanking the group for being willing to give tencent health space to devote itself to current technology research and development while attaching great importance to profit income.
wu wenda also pointed out that tencent health will not always rely on the group's investment and support, but will find a sustainable development model in the competition. in the next 1-3 years, tencent will continue to deepen its b-side business and cover more hospitals, pharmaceutical and medical equipment companies, and chain drug stores.
new approach to old business
since 2018, tencent's self-built medical sector has gradually formed three parts: the first is the infrastructure of medical informationization, the second is the medical knowledge service focusing on content, and the third is the medical ai with the help of big data and cloud computing. these three directions later gave birth to the electronic health card, medical insurance electronic certificate, tencent medical dictionary, and tencent miying.
in the commercialization process of the medical and health industry, payment by c-end users has always been a major problem. therefore, in tencent health’s strategy, the first two businesses mentioned above adhere to tencent’s consistent "connection" proposition, have no revenue kpi, and aim to enhance user stickiness and bring huge traffic to other products in the ecosystem.
according to tencent health, tencent medical dictionary currently has 600 million c-end users, and the electronic medical insurance card also has 500-600 million c-end users. c-end users of these inclusive products do not need to pay for the services, but the establishment of the service platform brings greater industrial value to b-end customers: the huge demand of c-end users, in turn, forces b-end service innovation. using technology and services to "connect" c and b ends is tencent's specialty.
for example, at this digital ecosystem conference, tencent health, tencent yuanbao and china resources sanjiu jointly launched the "sanjiu health butler" intelligent body, which is an ai smart assistant built on tencent's medical big model, cloud computing and other new technologies to provide health services to users. among the many tools that support the "sanjiu health butler" health question-and-answer service, there is the popular science knowledge of tencent medical dictionary.
figure 3: tencent health, tencent yuanbao and china resources sanjiu jointly released the "sanjiu health manager" intelligent body
tencent miying has also entered a period of adjustment. this product was born in 2017 when medical imaging ai was at its hottest. tencent health initially chose a path that most similar products were taking, treating it as an independent medical device software product for research and development, clinical trials, and approval and certification.
however, simply copying the business model of medical devices may not allow medical ai to be commercialized, and this star track has quickly become deserted. when tencent miying obtained the first ai class iii certificate for pneumonia ct image-assisted triage and evaluation software in 2021, the industry had already begun to decline, with yitu medical sold, ande medical intelligence bankrupted and reorganized, and many companies' ipos failed. tencent miying is also moving forward and adjusting at the same time. while obtaining class iii certificates for three ai devices for pneumonia, fundus, and colorectal, it is also responsible for the construction of the national new generation of artificial intelligence open innovation platform for medical imaging. in combination with the capabilities of tencent cloud, it has also incubated a "digital medical imaging platform", equipped with an artificial intelligence analysis engine on the cloud, to achieve the interconnection of medical imaging data and business processes, and integrate the "industry-university-research-management" of medical ai. these actions are different from traditional medical device companies, but they may better reflect the value of tencent.
"the medical device industry is a viable path, but it is a long and slow one. tencent is not a pure medical device company, and our channels and ecosystem are different from those of traditional medical device companies," said wu wenda. tencent health decided to change its mindset and focus on long-term exploration of basic technologies for large medical imaging models, "jumping out of the boundaries of traditional vertical models."
for tencent health, at this stage when it still needs to invest heavily in research and development, being able to achieve self-sufficiency and not rely on the group's transfusions is also the company's goal and expectation in terms of performance and profit and loss at this stage.
the arduous mountain of medical care will eventually be climbed.
(this article comes from china business network)
