NANJING, CHINA – NOVEMBER 25: Aerial view of Alibaba Jiangsu headquarters at night on November 25, 2025 in Nanjing, Jiangsu Province, China. (Photo provided by Fang Dongxu/VCG, Getty Images)
Video Visual China Group | Getty Images
As companies like Alibaba and ByteDance race to turn chatbots into full-service shopping and payment tools, Chinese tech giants are entering a new phase of the artificial intelligence race called “proxy commerce.”
Alibaba last week updated its Qwen AI chatbot to allow users to complete transactions such as ordering food or booking flight tickets directly within the interface.
The upgrade connects Qwen to Alibaba’s extensive e-commerce ecosystem, allowing users to compare customized product recommendations from platforms such as Taobao and the company’s travel site Fliggy, without leaving the chatbot, before ultimately completing a payment via Alipay.
Previously, Qwen could make recommendations based on user-generated prompts, but users still had to manually navigate across multiple platforms to make purchases.
The update reflects a broader shift among some global artificial intelligence companies from a focus on basic AI models to “agent AI” that performs tasks on behalf of users with limited supervision.
Shaochen Wang, research analyst at Counterpoint Research, refers to enhanced long-term user engagement, saying, “Agent transformation of commercial services enables maximum integration of user services and increases user stickiness.”
This allows companies to build a sustainable competitive advantage, known as a business moat, and protect their profits from competitors, he added.
Commercial applications for agentic AI are expected to range from self-driving cars to cybersecurity, but e-commerce has emerged as one of its earliest and most popular use cases, with US payments and technology giants also rolling out their first iterations in recent months.
Mr. Wang said that Alibaba is well-positioned to be a pioneer in agency commerce in China due to its advanced large language model capabilities and extensive e-commerce network covering clothing, food, housing and transportation.
Intensifying global competition
Alibaba’s strategy could help it compete with rival super apps like Tencent’s WeChat (widely regarded as China’s leading “super app”), which combines messaging, payments, e-commerce and other services into a single platform used by more than 1 billion users.
Other Chinese companies are also moving quickly.
In December, ByteDance upgraded its popular Doubao AI chatbot to handle tasks like ticket booking autonomously through integration with the e-commerce capabilities of Douyin, China’s version of TikTok.
The upgraded Doubao model was introduced in a prototype smartphone developed by ZTE Corp as a comprehensive AI assistant that can perform tasks across users’ mobile devices. However, some of Doubao’s planned features were later scaled back as competitors raised privacy and security concerns.
Meanwhile, Tencent President Martin Lau said during the company’s May 2025 earnings call that AI agents could become a core component of the WeChat ecosystem.
“AI agents will be fundamental to the evolution of super apps, and their success will depend on tight integration across payments, logistics, and social engagement,” Forrester vice president and principal analyst Charlie Dye told CNBC.
Chinese companies like alibaba, tencent Although ByteDance and ByteDance will compete by embedding agents across their platforms, they both benefit from an integrated ecosystem, rich behavioral data and consumer familiarity with super apps, Dai said.
Although Western companies have the lead in fundamental AI models and global reach, they face more fragmented data, stricter privacy regulations, and slower integration between services, he added.
US players pursuing agent commerce include OpenAI, Perplexity, Amazon. google is also exploring ways to position itself as a “middleman” between merchants, consumers, and AI agents.
“China prioritizes domestic integration and strategic expansion in some regions, while U.S. companies focus on global scalability and governance,” Dai said.
According to a 2025 McKinsey study, nearly half of all consumers already use AI when searching online.
The report estimates that AI agents could generate more than $1 trillion in economic value for U.S. businesses by 2030 by streamlining critical and routine steps in consumer decision-making.
