Ping An Good Doctor earns 35% more with its medical AI, but the 90% accuracy still needs fine print

By Insurance Asia · October 25, 2026.

Ping An Healthcare and Technology, known as Ping An Good Doctor and listed in Hong Kong (1833.HK), raised its net profit by 35% year on year in the first nine months of 2026, to $37.3 million (250 million yuan). Revenue came to $523.9 million (3.51 billion yuan) for the period ended September 30, according to the company. That puts the net margin at about 7%: profitable, but hardly a profit machine.

The company says its "AI doctor" service reached nearly 13 million users during the period. It also says its AI-assisted multidisciplinary platform, which supports diagnosis and treatment planning for complex diseases, covered four major disease categories by the end of September, with diagnosis and treatment plans about 90% accurate.

The corporate business is its other growth engine. Paying business clients served over the past 12 months topped 8,100, almost 60% more than a year earlier. In September the company also launched a brain health service focused on Alzheimer's with Ping An Life Insurance. It also unveiled Ping An Smart Private Coach, which pairs an AI coach with a human one.

**Our take.** The 90% figure deserves the most caution. In the available material, the company does not explain what that percentage measures, which cases it covers, what it is compared against, or whether a third party has validated it. It is a company figure, not an independently evaluated result, and without that definition it cannot be compared with rigorous clinical studies. In medical AI generally, what has been solidly proven tends to be triage and workload allocation (deciding which cases deserve human time) rather than replacing the specialist.

The business model is the interesting part. Ping An doesn't just sell an app: it ties it into insurance and employee benefits. That is the thesis: healthcare AI thrives where someone has a clear financial incentive to catch problems earlier and spend less, namely insurers and employers. This encourages prevention, but it also raises fair questions about how health data is used and who decides when the algorithm and the payer's interests don't align.

In the short term, the expansion brings friction: uneven quality, sensitive data and regulators who have yet to set the rules. In the long term, the underlying pattern is encouraging. An AI doctor already used by millions of people is an early version of something that could democratize access to healthcare and help us detect diseases such as Alzheimer's much earlier. That a company built around services like these is already profitable, even on a modest margin, suggests it can scale without relying on subsidies. What's missing is the essential part: independent evidence that it works as well as claimed.