China has quietly turned artificial intelligence into a shopper, a travel agent, and even a health aide. With Meituan’s Xiaomei, Alibaba’s Qwen, and Ant Group’s AQ now handling millions of real transactions each day, the country is becoming the world’s first full-scale test of agentic commerce, where AI does the work users used to do themselves. The shift is already pulling Walmart, Google, and OpenAI into the same race.
At a Glance:
- Alibaba’s Qwen app hit 100 million monthly users in under two months.
- Qwen processed 120 million orders in six days during Lunar New Year 2026.
- Walmart and Google launched agentic shopping in Gemini on January 11, 2026.
- Stanford AI Index: 83% of Chinese adults view AI as net positive, versus 39% in the US.
From Assistance to Real Delegation
For two decades, digital commerce has been about assistance. Search engines, review summaries, and recommendation feeds helped users decide, but humans still clicked, compared, and paid.
Agentic commerce breaks that pattern. The user no longer drives the workflow; the agent does, stepping back to the human only at defined checkpoints. The shift reflects a broader move among some global artificial intelligence firms from a focus on foundational AI models to “agentic AI”, which performs tasks on behalf of users.
China’s platforms are where this idea is being stress-tested at consumer scale. The early signals are clear enough that executives at Walmart, Google, and OpenAI are now openly describing their own playbooks in the same language. Key shifts already visible include:
- Users speaking, not tapping, to place orders and make bookings.
- Agents pulling payment, identity, and logistics into a single flow.
- Brands competing to be chosen by a machine, not just a person.
Inside China’s Agent Playbook
Meituan kicked off the wave in September 2025. Its public beta of Xiaomei began on September 12, 2025, positioned to provide local life services such as food delivery, restaurant recommendations, table reservations, and navigation, fully integrated into Meituan’s core businesses and powered by its self-developed LongCat-Flash-Chat model.
Alibaba answered four months later. It announced a major upgrade to its flagship consumer AI application, Qwen App, with deep integration of core services including Taobao, Taobao Instant Commerce, Alipay, Fliggy, and Amap, allowing users to move seamlessly from intent to completion.
Ant Group has pushed delegation into health care, traditionally a sector where regulators move slowly. AQ connects users with digital services from more than 5,000 hospitals nationwide and online consultations with 300,000 licensed physicians, while more than 1,000 AI Doctor Agents collectively responded to over 27 million inquiries in 2025; as of January 2026, the AQ app has reached 30 million monthly active users and handles more than 10 million health-related queries daily.
| Platform | Agent | Core Tasks |
|---|---|---|
| Meituan | Xiaomei | Food orders, restaurant booking, local services |
| Alibaba | Qwen App | Shopping, travel, payments, quick commerce |
| Ant Group | AQ (Ant A-Fu) | Health Q&A, doctor booking, insurance claims |
| ByteDance | Doubao | Cross-app ticketing and shopping comparisons |
Why China Scales Faster
China’s lead is not about better models. It is about plumbing that already works.
Alipay and WeChat Pay have embedded identity and authorization into daily life. Dense on-demand logistics turn a voice command into a delivered meal within minutes. And a small number of super apps span shopping, payments, maps, and travel inside one account.
Consumer appetite is unusually high. According to Stanford HAI’s 2025 AI Index Report, 83% of respondents in China see AI products and services as more beneficial than harmful, compared with 39% in the United States, 40% in Canada, and 36% in the Netherlands.
Regulation in China also lets products ship before rules fully harden. That sequencing has allowed agents to reach tens of millions of users before governance debates are settled, something US companies rarely enjoy.
The result is visible in spending. In the run up to the 2026 Lunar New Year, China’s largest internet companies spent heavily to accelerate adoption of their consumer AI apps; Alibaba said it would spend up to RMB 3 billion on Qwen promotions during the holiday period, while Tencent and Baidu pledged RMB 1 billion and RMB 500 million respectively for similar campaigns.
100 million monthly users on Qwen within two months of beta launch.
120 million orders placed through Qwen in six days during Lunar New Year.
1.56 million seniors aged 60+ made their first online purchase via the app.
30 million monthly users on Ant Group’s AQ health agent by January 2026.
The New Agent Shelf for Brands
When an agent filters the world for a shopper, traditional marketing funnels bend. The battle moves upstream, from winning attention to earning selection by a machine.
Alibaba’s leadership is framing the change in stark terms. “AI is evolving from intelligence to agency,” said Wu Jia, Vice President of Alibaba Group. “What we are launching today represents a shift from models that understand to systems that act, deeply connected to real-world services.”
“The transition from traditional web or app search to agent-led commerce represents the next great evolution in retail. We aren’t just watching the shift, we are driving it.” — John Furner, incoming CEO, Walmart Inc.
https://x.com/AntGroup/status/1966438897547948300
What US and Global Leaders Must Do Next
The race has crossed the Pacific. On January 11, 2026, Walmart and Google announced their agentic commerce partnership at NRF’s Big Show. Walmart is one of the first major retailers to use Google’s new Universal Commerce Protocol, also announced at NRF 2026, to let AI agents complete checkout.
Rival standards are already emerging. Walmart announced a deal with OpenAI’s ChatGPT in October to allow shoppers to make purchases with “Instant Checkout,” a feature that allows them to buy an item without leaving the AI chatbot; OpenAI recently launched that feature with Walmart and it has Instant Checkout deals with other retailers, including Etsy and several Shopify merchants.
The harder problem for Western retailers is data discipline, not AI talent. As a recent Modern Retail analysis of Alibaba’s strategy noted, turning AI systems into reliable shopping agents remains technically difficult, in part because these assistants need large amounts of structured data not just to recommend products but also to complete purchases, and e-commerce catalogs, prices and inventory levels are constantly changing.
Executives who treat agents as just another channel will likely lose shelf space, because Alibaba’s own Qwen roadmap makes clear that operational reliability, not ad budgets, decides whether an agent includes a brand at all. Firms that invest early in clean catalogs, clear return policies, and predictable fulfillment will be the ones an agent keeps picking.
Frequently Asked Questions
What is agentic commerce?
It is a model where an AI agent handles discovery, comparison, payment, and delivery on a user’s behalf, with humans intervening only at set checkpoints.
Which Chinese apps are leading agentic commerce?
Meituan’s Xiaomei, Alibaba’s Qwen, ByteDance’s Doubao, and Ant Group’s AQ health app are the main consumer-facing agents in China today.
How big is Alibaba’s Qwen app?
Qwen hit 100 million monthly active users within two months of its November 17, 2025 beta launch and processed 120 million orders in six days during Lunar New Year 2026.
What is the Universal Commerce Protocol?
It is an open standard announced by Google at NRF 2026, letting AI agents like Gemini pull catalogs, pricing, and checkout from retailers such as Walmart and Sam’s Club.
Why is China ahead in agentic commerce?
Deep payment rails, dense logistics, tightly integrated super apps, high consumer trust in AI, and flexible regulation let agents execute end-to-end faster than in the US or Europe.
Agentic commerce is no longer a future concept. With 100 million users on Qwen, 30 million on AQ, and Walmart routing shopping carts through Gemini, the center of digital retail is moving from screens to agents. The real question is whether Western firms can build the data, trust, and governance needed to be chosen by a machine. Share your thoughts in the comments.
Leave a Comment