After nearly a decade in online medical consultations, Shu Hao never expected change to arrive so swiftly.

Shu Hao is a senior attending physician in internal medicine at a top-tier Grade-A hospital in China. He began engaging with online consultations as early as 2016 and registered on virtually every major telemedicine platform. In terms of consultation volume, he consistently ranked among the top providers on several core platforms, averaging around 500 consultations per month.

In late October this year, one of his primary platforms launched an AI-powered health assistant product. At the time, this industry-focused announcement didn’t raise any alarms for him. But by mid-November, he noticed his consultation volume on that platform had dropped by roughly 40%.

“Other platforms have also seen declines—around 20% previously,” Shu Hao said.

Su Yu, who works at an internet healthcare company, spotted this trend even earlier. Her company had long established a dedicated business line centered on online consultations, cultivating strong doctor loyalty and solid user reputation.

However, starting in the second half of 2024, she observed a steady decline in consultation orders. “At first, we couldn’t pinpoint the cause. Only later, as we became more familiar with AI, did we realize that many shallow, health-related inquiries were already being replaced by AI.”

“All routine, standardized consultations will gradually be taken over by AI,” said Xu Xiao, a veteran in internet healthcare, in an interview with Jianwen Consulting. According to him, such replaceable consultation demands account for nearly 80% of all online medical scenarios.

Coincidentally, two recent major developments in China’s digital health sector indirectly confirm this trend. First, Chunyu Doctor sold 78.29% of its equity to Hong Kong-listed Guorui Lifestyle for a total valuation of just RMB 340 million. Second, WeDoctor launched its international cloud pharmacy, seeking a differentiated path in pharmaceutical e-commerce.

Both companies were pioneers in online consultations and among the first to join the “battle of hundreds of medical AI models.” Yet while the ambition of “AI replacing doctors” remains largely aspirational, AI’s infiltration into online consultation services is already unmistakable.

Within the library of large-model-driven medical AI tools—from AI health assistants handling basic inquiries to AI agents supporting deeper diagnostic needs, ultimately connecting users back to human physicians—AI has functionally achieved full coverage of the entire online consultation workflow.

In the era of large language models, the most overt narrative isn’t about disruptive business models—it’s about labor replacement.

I. No One Is Spared

Jia Wei, a 30-year-old urban professional deeply invested in her health, has never hesitated to spend money on medical care.

According to her phone records, she initiated 33 online consultations in 2024, mostly through mainstream internet platforms. Having just become a new mother that year, her questions primarily concerned newborn care. Each consultation cost between RMB 50–150, totaling over RMB 2,500 in expenses.

But when reviewing her habits this year, she realized they’d changed dramatically: her reliance on human doctors dropped to single digits, with most of her health queries now directed toward AI tools like DeepSeek, Doubao, and AQ.

She analyzed her consultation history and identified clear patterns. For personalized yet straightforward health questions—such as “What if my 18-month-old refuses milk?” or “Can medications A and B be taken together?”—she now prefers AI. These queries either have definitive answers or stem from anxiety; what she truly needs is a quick assessment of urgency and severity.

For Jia Wei, this shift didn’t happen overnight—but it clearly correlates with the growing availability of AI-powered consultation tools.

In May 2023, Chunyu Doctor—one of China’s earliest online consultation platforms—officially launched its AI consultation product, “Chunyu Smart Ask.” Two weeks later, Yilian released its self-developed medical large model, MedGPT, branding it as China’s “first AI doctor.” Soon after, JD Health, Ant Group, Tencent Health, and others followed suit with their own AI consultation offerings.

During this period, most products focused on general wellness and light consultations, commonly branded as “AI health assistants.”

“The AI consultation products from 2023 to 2024 were still relatively crude—mostly generic large models fine-tuned with publicly available medical data, functioning more like AI general practitioners,” said an investor who has long tracked the digital health sector, speaking to Jianwen Consulting.

Starting in the second half of 2024, however, as collaborations between companies and hospitals deepened, specialized disease-focused large models began rolling out en masse.

In the first half of 2025 alone, China saw the launch of 22 new specialty-specific large models covering areas like oncology, cardiovascular diseases, and rare disorders—most co-developed through hospital-enterprise partnerships.

These specialized models enabled the emergence of “digital avatars” of renowned physicians and specialty AI agents. Unlike generalist “AI health assistants,” these tools—trained on real-world data from top hospitals and refined through human oversight—aim to deliver higher-value diagnostic and treatment recommendations within specific disease domains.

China’s first specialty AI agent—the Urology AI Agent from Renji Hospital—reported a revealing statistic eight months after launch: among the 300,000 patients it served, only 3,000 were referred to human doctors for appointment booking. This implies that nearly 99% of user inquiries were resolved entirely by AI.

“For internet healthcare platforms that built their businesses on online consultations, these upgraded, specialty-capable AI agents represent a precise, overwhelming competitive threat—almost impossible to evade,” the investor noted.

II. The Boundary of Replacement

Unlike physicians at elite hospitals who were slow to react, Zhou Ting—a pediatrician at a central hospital in a fourth-tier city—began noticing signs as early as early 2024.

The most obvious change was the rising barrier to platform participation. Previously, during weekends or days off, he could schedule 6–8 hours of online shifts. At 7–8 patients per hour, he earned RMB 400–500 per day.

Since last year, however, platforms have increasingly prioritized doctors from top-tier hospitals in major cities. His available online time shrank to just 1–2 hours per day—mostly after 10 p.m.—and his monthly income plummeted from over RMB 5,000 at its peak to around RMB 1,000.

“It’s the same across all platforms—too many doctors, too few slots. That’s another manifestation of declining consultation volume,” he said.

He follows industry trends and understands AI’s role. On one hand, he acknowledges the trend: about 30% of past consultations involved basic health education or wellness advice—queries AI can answer just as effectively. But for the remaining 70% involving actual medical conditions, he believes AI still “falls short.”

“For example,” he explained, “sometimes I reply politely in text, and users immediately ask, ‘Are you a robot?’ So I switch to voice right away and say, ‘No.’ Their unspoken message is clear: if it’s a bot, I don’t need it.”

Most doctors we interviewed expressed similar views. A physician from a top-tier hospital who participated in developing a specialty large model told Jianwen Consulting: “AI can provide comprehensive analysis of symptoms, but it doesn’t converge on a specific diagnosis. If you’re seeking a clear clinical conclusion, AI still can’t replace human doctors in online consultations—at least not yet.”

Yet AI holds a decisive advantage: it’s free.

As of now, whether from tech giants like JD, Baidu, and Ant, or vertical medical AI startups, all consumer-facing AI health management tools are offered at no cost.

This has been the Achilles’ heel of internet healthcare’s commercialization efforts over the past decade. Nearly every online consultation platform initially relied on heavy subsidies to drive traffic and build user habits. But once funding dried up and platforms shifted to paid models, users rarely returned.

In 2016, Chunyu Doctor—despite boasting a 99%+ user satisfaction rate—tested a membership-based paid model. Within a month, daily consultations plunged from 30,000 to just 3,000. In 2021, Haodf Online, which had struggled with profitability for years, launched a similar subscription service—with equally disappointing results.

Later, Haodf founder Wang Hang summarized the issue succinctly in an interview with Jianwen Consulting: “People simply aren’t ready to pay for services yet.”

AI consultation tools, being free, directly address this pain point.

“Online consultations themselves may not be profitable, but their real value lies in connecting other monetizable segments of internet healthcare—so everyone was willing to sustain them, differing only in how much they invested,” Xu Xiao explained. “But with large models now available, platform operators realized that in light-consultation scenarios, AI can maintain the existing business model at a fraction of the cost.”

Perhaps beyond commercial considerations, AI and human doctors do offer distinct values in online consultations—but the boundary between them remains blurry.

One reasonably clear conclusion is this: as free AI tools divert users with superficial inquiries, the overall volume of physician-led online consultations will continue to shrink. Yet this demand will never disappear entirely. Where it ultimately lands—and who fulfills it—remains a profound question worth further reflection.

After nearly a decade in online medical consultations, Shu Hao never expected change to arrive so swiftly.

Shu Hao is a senior attending physician in internal medicine at a top-tier Grade-A hospital in China. He began engaging with online consultations as early as 2016 and registered on virtually every major telemedicine platform. In terms of consultation volume, he consistently ranked among the top providers on several core platforms, averaging around 500 consultations per month.

In late October this year, one of his primary platforms launched an AI-powered health assistant product. At the time, this industry-focused announcement didn’t raise any alarms for him. But by mid-November, he noticed his consultation volume on that platform had dropped by roughly 40%.

“Other platforms have also seen declines—around 20% previously,” Shu Hao said.

Su Yu, who works at an internet healthcare company, spotted this trend even earlier. Her company had long established a dedicated business line centered on online consultations, cultivating strong doctor loyalty and solid user reputation.

However, starting in the second half of 2024, she observed a steady decline in consultation orders. “At first, we couldn’t pinpoint the cause. Only later, as we became more familiar with AI, did we realize that many shallow, health-related inquiries were already being replaced by AI.”

“All routine, standardized consultations will gradually be taken over by AI,” said Xu Xiao, a veteran in internet healthcare, in an interview with Jianwen Consulting. According to him, such replaceable consultation demands account for nearly 80% of all online medical scenarios.

Coincidentally, two recent major developments in China’s digital health sector indirectly confirm this trend. First, Chunyu Doctor sold 78.29% of its equity to Hong Kong-listed Guorui Lifestyle for a total valuation of just RMB 340 million. Second, WeDoctor launched its international cloud pharmacy, seeking a differentiated path in pharmaceutical e-commerce.

Both companies were pioneers in online consultations and among the first to join the “battle of hundreds of medical AI models.” Yet while the ambition of “AI replacing doctors” remains largely aspirational, AI’s infiltration into online consultation services is already unmistakable.

Within the library of large-model-driven medical AI tools—from AI health assistants handling basic inquiries to AI agents supporting deeper diagnostic needs, ultimately connecting users back to human physicians—AI has functionally achieved full coverage of the entire online consultation workflow.

In the era of large language models, the most overt narrative isn’t about disruptive business models—it’s about labor replacement.

I. No One Is Spared

Jia Wei, a 30-year-old urban professional deeply invested in her health, has never hesitated to spend money on medical care.

According to her phone records, she initiated 33 online consultations in 2024, mostly through mainstream internet platforms. Having just become a new mother that year, her questions primarily concerned newborn care. Each consultation cost between RMB 50–150, totaling over RMB 2,500 in expenses.

But when reviewing her habits this year, she realized they’d changed dramatically: her reliance on human doctors dropped to single digits, with most of her health queries now directed toward AI tools like DeepSeek, Doubao, and AQ.

She analyzed her consultation history and identified clear patterns. For personalized yet straightforward health questions—such as “What if my 18-month-old refuses milk?” or “Can medications A and B be taken together?”—she now prefers AI. These queries either have definitive answers or stem from anxiety; what she truly needs is a quick assessment of urgency and severity.

For Jia Wei, this shift didn’t happen overnight—but it clearly correlates with the growing availability of AI-powered consultation tools.

In May 2023, Chunyu Doctor—one of China’s earliest online consultation platforms—officially launched its AI consultation product, “Chunyu Smart Ask.” Two weeks later, Yilian released its self-developed medical large model, MedGPT, branding it as China’s “first AI doctor.” Soon after, JD Health, Ant Group, Tencent Health, and others followed suit with their own AI consultation offerings.

During this period, most products focused on general wellness and light consultations, commonly branded as “AI health assistants.”

“The AI consultation products from 2023 to 2024 were still relatively crude—mostly generic large models fine-tuned with publicly available medical data, functioning more like AI general practitioners,” said an investor who has long tracked the digital health sector, speaking to Jianwen Consulting.

Starting in the second half of 2024, however, as collaborations between companies and hospitals deepened, specialized disease-focused large models began rolling out en masse.

In the first half of 2025 alone, China saw the launch of 22 new specialty-specific large models covering areas like oncology, cardiovascular diseases, and rare disorders—most co-developed through hospital-enterprise partnerships.

These specialized models enabled the emergence of “digital avatars” of renowned physicians and specialty AI agents. Unlike generalist “AI health assistants,” these tools—trained on real-world data from top hospitals and refined through human oversight—aim to deliver higher-value diagnostic and treatment recommendations within specific disease domains.

China’s first specialty AI agent—the Urology AI Agent from Renji Hospital—reported a revealing statistic eight months after launch: among the 300,000 patients it served, only 3,000 were referred to human doctors for appointment booking. This implies that nearly 99% of user inquiries were resolved entirely by AI.

“For internet healthcare platforms that built their businesses on online consultations, these upgraded, specialty-capable AI agents represent a precise, overwhelming competitive threat—almost impossible to evade,” the investor noted.

II. The Boundary of Replacement

Unlike physicians at elite hospitals who were slow to react, Zhou Ting—a pediatrician at a central hospital in a fourth-tier city—began noticing signs as early as early 2024.

The most obvious change was the rising barrier to platform participation. Previously, during weekends or days off, he could schedule 6–8 hours of online shifts. At 7–8 patients per hour, he earned RMB 400–500 per day.

Since last year, however, platforms have increasingly prioritized doctors from top-tier hospitals in major cities. His available online time shrank to just 1–2 hours per day—mostly after 10 p.m.—and his monthly income plummeted from over RMB 5,000 at its peak to around RMB 1,000.

“It’s the same across all platforms—too many doctors, too few slots. That’s another manifestation of declining consultation volume,” he said.

He follows industry trends and understands AI’s role. On one hand, he acknowledges the trend: about 30% of past consultations involved basic health education or wellness advice—queries AI can answer just as effectively. But for the remaining 70% involving actual medical conditions, he believes AI still “falls short.”

“For example,” he explained, “sometimes I reply politely in text, and users immediately ask, ‘Are you a robot?’ So I switch to voice right away and say, ‘No.’ Their unspoken message is clear: if it’s a bot, I don’t need it.”

Most doctors we interviewed expressed similar views. A physician from a top-tier hospital who participated in developing a specialty large model told Jianwen Consulting: “AI can provide comprehensive analysis of symptoms, but it doesn’t converge on a specific diagnosis. If you’re seeking a clear clinical conclusion, AI still can’t replace human doctors in online consultations—at least not yet.”

Yet AI holds a decisive advantage: it’s free.

As of now, whether from tech giants like JD, Baidu, and Ant, or vertical medical AI startups, all consumer-facing AI health management tools are offered at no cost.

This has been the Achilles’ heel of internet healthcare’s commercialization efforts over the past decade. Nearly every online consultation platform initially relied on heavy subsidies to drive traffic and build user habits. But once funding dried up and platforms shifted to paid models, users rarely returned.

In 2016, Chunyu Doctor—despite boasting a 99%+ user satisfaction rate—tested a membership-based paid model. Within a month, daily consultations plunged from 30,000 to just 3,000. In 2021, Haodf Online, which had struggled with profitability for years, launched a similar subscription service—with equally disappointing results.

Later, Haodf founder Wang Hang summarized the issue succinctly in an interview with Jianwen Consulting: “People simply aren’t ready to pay for services yet.”

AI consultation tools, being free, directly address this pain point.

“Online consultations themselves may not be profitable, but their real value lies in connecting other monetizable segments of internet healthcare—so everyone was willing to sustain them, differing only in how much they invested,” Xu Xiao explained. “But with large models now available, platform operators realized that in light-consultation scenarios, AI can maintain the existing business model at a fraction of the cost.”

Perhaps beyond commercial considerations, AI and human doctors do offer distinct values in online consultations—but the boundary between them remains blurry.

One reasonably clear conclusion is this: as free AI tools divert users with superficial inquiries, the overall volume of physician-led online consultations will continue to shrink. Yet this demand will never disappear entirely. Where it ultimately lands—and who fulfills it—remains a profound question worth further reflection.