Dialogue with Zhang Junjie of Ant Group: How does AI reshape the medical and health industry?

As large models transition from the technological singularity to the industrial foundation, and as intelligent agents move from laboratories to production lines and clinics, the third wave of artificial intelligence is reshaping the global economic fabric with unprecedented sharpness.

China demonstrates dual advantages in this transformation: it is both a testing ground with ultra-large-scale application scenarios and is launching attacks in deep waters such as chip breakthroughs and algorithm open-sourcing. From breaking through single-point technologies to ecosystem-level innovation, from efficiency tools to new-quality productivity engines, an AI development path with Oriental characteristics is emerging at an accelerating pace.

On July 26th, the World Artificial Intelligence Conference (WAIC 2025), themed "In the Intelligent Era, We All Share One Earth", brought together technology giants, academic pioneers, and policy-makers from the AI field. This super feast covering technology, ethics, and art indicates that AI has evolved from an "industrial variable" to a "civilizational constant".

At this grand event about the artificial intelligence industry, 36Kr not only acts as an industry observer but also participates deeply as an industry connector. It set up the "Krypton Star Live Studio" in the exhibition hall to reveal the underlying logic of the advancement of the artificial intelligence industry through dialogues.

During the dialogue, Zhang Junjie, the vice president of Ant Group and the general manager of the Digital Medical and Health Division, said: "The arrival of the AI era will change the engine of the medical industry. It used to be like a car engine, but now it may be directly replaced with an airplane engine, bringing about a huge change in productivity. In the past decade, the Internet and mobile technology have made it more convenient for users to seek medical treatment. The arrival of AI can play a role in user inclusiveness and productivity, and there will also be a huge upgrade in user experience."

The following is the transcript of the dialogue, edited by 36Kr:

36kr: First, could you introduce the business situation of your company?

A: I'm from Ant Group and in charge of the digital medical and health sector. This is my 12th year at Ant Group. Our sector is mainly divided into several parts. One is that we used to solve the problem of the convenience of users' medical payments. 800 million users have activated and used the medical insurance code on Alipay. Recently, we also launched the commercial insurance code. We hope that users can more conveniently use our various payment tools during the process of seeking medical treatment and purchasing medicine.

The second is the Alipay Medical and Health Channel. This channel offers one-stop services such as registration, consultation, medicine purchase, oral vaccination, and even medical aesthetics. Now, more than 800 million users visit the medical and health channel every year to use various services. Through this service, we are connected to nearly 3,600 public hospitals across the country. Users can also conveniently use our services during their offline medical treatment process.

The third is that since 2023, we have focused on exploring the AI medical sector. We hope to provide each user with an AI health butler, making it more convenient for users. At the same time, we also aim to provide AI assistant services for institutions, hospitals, and doctors, enabling doctors, including primary care doctors, to use AI to assist their work.

36kr: Now that artificial intelligence or AI technology is relatively developed, what can Ant do for users by combining these new technologies? Can you give a specific example?

A: In fact, although Internet medical services have developed for many years, we always feel that the overall user experience is still not convenient enough during the process of moving offline services online. For example, as a user, my real experience is that if I feel unwell one day, I will definitely search on the search engine first, and then look for the corresponding specialty or doctor. Then I'll go to platforms like Haodf to find a doctor. After finding a doctor, I have to go to Alipay or WeChat to register at the hospital. Before seeing the doctor, there are actually few opportunities for communication. I'll just talk to the doctor for a few minutes and then go home. After going home, there are no good tools to help users with post-operative health management services or out-of-hospital follow-up services. In addition, in the general health field, it's also difficult for users to obtain professional, authoritative, and easily understandable medical information and health management advice on a reliable platform. As we mentioned before, these are all users' medical needs. For users, the demand for health management services is greater, but the supply is limited. So we hope to solve all these problems through the AI health butler.

First, no matter what questions you have, the AI can answer them professionally. Second, in response to your needs, we not only answer but also have the means to call out relevant services. For example, if you feel unwell one day, after multiple rounds of Q&A with the AI, it determines that you need to go for an offline examination. The AI can directly call out the registration entrance for the most suitable hospital and doctor for the department you need to register for; for example, if you need to seek medical treatment in Shanghai from another place and need to make off-site medical insurance registration, it can directly help you with that. The combination of service accumulation and AI is also very important.

Third, in fact, users often can't clearly express their medical needs. That's why we call it a health butler. When you can't put forward your needs, the product can actively give you health management suggestions. So in addition to users' active questions, we also connect the AI with hospitals' health records, wearable devices, and even household medical devices. We've made a lot of efforts and hope to provide users with real butler-style services.

36kr: So what technical or engineering difficulties have you encountered in the process of implementing these technologies or applications?

A: The difficulties are quite significant. In any innovation and breakthrough in the medical field, the first problem is the issue of standards. For example, when we were doing medical insurance payments before, we had to continuously polish and form standards and then connect institutions and users. This work involves a lot of groundwork. At this stage, we launched the "AI Health Butler" in October 2023, and more than 100 million users have used it by now. During this process, it's very, very important for us to have in-depth cooperation with the Health Commission, large hospitals, and many leading doctors.

AI must have in-depth cooperation with the professional side to better serve users. In addition to the standard issue, the second is how to polish the AI's capabilities, having an authoritative medical knowledge system and a medical expert team for joint annotation. The third is how to make the delivery of AI meet users' needs and expectations, not only providing professionalism but also giving users corresponding emotional comfort and real help and service connections. On the one hand, these are technical difficulties; on the other hand, there are industry standards, including the cooperation of data and doctor resources with relatively high industry barriers. These are all difficulties. But in the medical field, there is also a point that once these difficulties are overcome one by one, you'll find that countless people will help us push this matter forward because every step forward in this matter brings great significance and value.

36kr: So in the process of implementing these technologies or applications, have you made any prior arrangements regarding the privacy of user data?

A: Data and privacy security mainly involve ensuring the actual protection of users' data security during the usage process. So in the process of AI and product design, we focus on how to ensure that every key link of users' health records is authorized by users, how to ensure the minimum scope of use, and how to ensure that users' information security is not leaked, etc. On the one hand, we've done a lot of technical work. On the other hand, in terms of medical ethics and safety construction, Ant is also researching and formulating reference paths together with the industry, integrating the principles of science and technology ethics into the entire process of technology R & D to ensure the safe, effective, and compliant application of medical and health large models.

36kr: Nowadays, people attach great importance to their physical condition, and their health awareness is constantly increasing. The national policies also play a significant role in promoting the development of our industry.

A: Yes, medical AI must involve in-depth cooperation among medical administration, industry, academia, and research. The medical industry is very complex. Maybe a general model only needs to be smart enough, but how can it be made solid enough in terms of medical rigor, professionalism, even including multi-modality, and each vertical specialty? There are a lot of policy and technical difficulties to be overcome one by one.

From 2023 to now, our cooperation with many hospitals and experts has reached a very high level. For example, our cooperation with Shanghai Renji Hospital in urology. In the evaluation in the previous two or three months, its level has actually exceeded that of general practitioners and is very close to that of associate chief physicians. We spent nearly three years, about two and a half years, building this vertical AI model, gradually polishing it with the hospital. Now it has reached a level where it can provide professional consultation services for users and even has the ability to assist some primary care doctors in diagnosis.

I believe that if one specialty can achieve this, other specialties can also. A typical example is Dr. Mao Hongjing, the president of Hangzhou Seventh People's Hospital. He is an expert in mental sleep and a very leading expert in the country. It used to take at least a month to queue up for an appointment with him offline, and he could only see more than 600 patients per month.

In fact, Dr. Mao also wanted to spend more time communicating with patients. But after the AI intelligent agent was introduced, it has served more than 4 million patients in total. More than 30,000 people interact with the AI intelligent agent every day, with more than 110,000 interactions. During this process, the level of the AI intelligent agent will become higher and higher. At the same time, it can do classification. More than 90% of the interactions are simple consultations. For more complex cases, patients can seek help through online Internet hospitals and offline hospitals, so that precious medical resources can be given to those who really need them. At the same time, users' consultations can be synchronized to the doctor's workstation, improving doctors' efficiency in seeing patients.

So we think that the arrival of the AI era will change the engine of the medical industry. It used to be like a car engine, but now it may be directly replaced with an airplane engine, bringing about a huge change in productivity. In the past decade, the Internet and mobile technology have made it more convenient for users to seek medical treatment. The arrival of AI can play a role in user inclusiveness and productivity, and there will also be a huge upgrade in user experience.

After two to three years of development, we've seen very significant results. For example, in less than a year from last October to now, nearly 100 million users have used our AI health butler. After the independent app was launched, middle-aged and elderly users aged 40 - 60 use it very frequently, which proves that they are a group with a strong demand for this service.

Especially for third - and fifth - tier cities, it's difficult to have the same medical resources as in Beijing, Shanghai, and Guangzhou. But now our AI can help them well. Of course, there are still many directions for improvement. So we think that medical + AI should be the pearl on the crown among AI vertical industries because of its great social value and is worthy of long - term investment and refinement.

36kr: If a newly graduated medical student is rated as 1 and a top - level medical expert like Dr. Mao is rated as 10, what level do you think the large model can reach now?

A: How to scientifically and authoritatively evaluate the capabilities of a model also needs to be defined by the industry. I'll talk about two aspects. First, from a subjective feeling, for example, Dr. Mao rates his AI every year compared with his own diagnostic level. He gave it 20 points in the first year and 80 points in the second year. The growth rate is very fast. The urology model of Renji Hospital has exceeded the ordinary standard technical model and general practitioners, and is similar to the level of associate chief physicians.

On the other hand, having an authoritative and scientific evaluation standard for medical large models is crucial. So under the guidance of the Health Commission, there are also some works on medical artificial intelligence in Zhejiang. Under the guidance of relevant national ministries and commissions, we are also actively participating in promoting the standard - setting work.

For example, in addition to the capabilities of the basic model in large - language models, dialogue capabilities, and medical examination capabilities, the overall level is relatively high. But it's very important to have an authoritative evaluation standard for medical reasoning, multi - modality, and specialty models.

Under the guidance of the China Academy of Information and Communications Technology, we released the grading standard for medical doctor intelligent agents, which is a bit like the grading from L0 to L5 in autonomous driving. In the future, there will also be grading standards for general practitioner intelligent agents. Specialized imaging, multi - modality vertical series medical models should all have their own evaluation standards. With these guiding standards, companies in the industry can all benefit.

36kr: We all know that Ant's technology is indeed leading, but its greatest advantage is not the technology itself but the integration of technology and its own ecosystem. In your opinion, what are the biggest difficulties in the overall ecosystem integration in the medical sector, or what are the biggest benefits after integration?

A: We've always maintained an open attitude. Ant didn't start doing medical and health services overnight. We started with the most difficult things and focused on the aspects where users have the greatest needs. From the first online payment in 2014, to the first electronic social security card, the first electronic medical insurance certificate, a medical and health channel, and now moving forward with an independent app.

Second, we think users' trust is very important. We cherish users' trust very much. In the past, we quietly played the role of tools and assistants, and users found it convenient to use without too much interference. We hope that after the arrival of the AI era, we can uphold our original intention and bring small but wonderful changes to users' medical and health.

In terms of ecosystem cooperation, from the very beginning of developing AI, we've had in - depth cooperation with the government, doctors, medical institutions, and universities. Because in the medical field, government, industry, academia, and research are inseparable and need to cooperate. AI is not an opponent but provides solid technical support. For example, after AI enters the deep water area in the medical field, it needs to connect with hardware and have in - depth cooperation with specialized fields. There are a lot of partners, and we need to work together.

Ant has done a lot in the medical and health industry in the past 11 years in terms of ecosystem, service accumulation, and user trust accumulation, which can be well applied and utilized in the AI era. There are also many challenges, such as the standards for user data privacy and security, how to appear in the right place at the right time for users and doctors, and the combination of hardware and AI. These are all challenges faced by the industry, and we're working hard to overcome them.