🎙 Listen to a summary of this story
For decades, online shopping meant typing a product into a search box, scrolling through pages of results and deciding what to buy. In China, artificial intelligence is beginning to dismantle that familiar process.
Consumers increasingly can tell an AI system what they need — a coat for an upcoming trip, a replacement phone within a particular budget or ingredients for dinner — and allow the technology to search, compare and recommend products. Increasingly, it can also complete the transaction.
The development marks a potentially profound change for the world's largest e-commerce market. Artificial intelligence is moving from a technology that predicts what consumers might click on to an intermediary capable of interpreting what they want and acting on their behalf.
That transition accelerated this year when Alibaba integrated its Qwen artificial intelligence system more deeply into Taobao and Tmall. In May, the company opened Taobao's catalogue of more than 4 billion products to Qwen, allowing shoppers to browse, compare products, place orders and manage deliveries through natural-language conversations rather than conventional keyword searches.
The result is an emerging form of "agentic commerce," in which AI does more than recommend products. It can carry out parts of the shopping process itself, potentially putting an algorithm between retailers and millions of consumers.
From search engines to shopping agents
Chinese e-commerce companies have long been sophisticated users of recommendation algorithms. Platforms analyse searches, purchases, clicks and other signals to determine which products appear on a user's screen. Generative AI and increasingly capable AI agents, however, are pushing that model further.
Alibaba's Qwen integration illustrates the difference. Instead of searching individually for products and comparing listings, a consumer can describe a requirement conversationally and ask the system to narrow the choices. Taobao's AI shopping tools can also provide services such as price tracking, while AI agents connect shopping with logistics and after-sales services.
Alibaba had already begun integrating Qwen with services including Taobao, Alipay, travel platform Fliggy and mapping service Amap. The company described the strategy earlier this year as moving from AI that merely responds to AI that acts — potentially allowing a single request to trigger several services.
That could fundamentally alter competition online. In the traditional e-commerce model, brands compete for prominent search rankings, advertising positions and consumer attention. In an AI-mediated market, they may increasingly compete to become the product selected by an algorithm.
For merchants, that makes the quality and structure of product information more important. Prices, specifications, availability, delivery information, customer reviews and other data must be understandable not only to shoppers but also to the AI systems acting for them.
AI is simultaneously transforming operations behind the storefront. Retailers can use machine learning to forecast demand, manage inventories, generate marketing material and identify regional variations in purchasing behaviour. Virtual try-on technology and AI customer-service agents are making the technology increasingly visible to shoppers as well.
A battle to control the consumer gateway
The stakes are particularly high because whoever controls the AI assistant may gain enormous influence over which products consumers see.
China's biggest technology companies possess an advantage that many standalone AI developers do not: enormous existing digital ecosystems. Alibaba combines AI with e-commerce, payments, logistics, mapping and travel services, giving its systems opportunities to move directly from a consumer's request to a transaction.
That makes the contest over consumer AI partly a battle over the gateway to China's digital economy. If shoppers increasingly begin with an AI conversation rather than an e-commerce search page, traditional techniques for attracting consumers — including search optimisation and display advertising — could become less influential.
The change could be especially significant for smaller merchants. An AI agent capable of examining thousands of listings might theoretically help consumers discover unfamiliar brands. But it could also concentrate attention around products that algorithms regard as reliable, competitively priced or well reviewed.
Trust therefore remains critical. Consumers may be comfortable allowing software to recommend inexpensive everyday goods while remaining reluctant to delegate expensive or consequential purchases. Retailers and technology companies must also demonstrate that recommendations reflect consumers' interests rather than simply directing them toward the products that generate the greatest revenue.
China tries to put boundaries around the algorithm
The increasing power of AI over consumer decisions also collides with China's expanding regulation of algorithms and personal data.
Rules governing algorithmic recommendation services that took effect in 2022 require providers to disclose that recommendation algorithms are being used and give consumers options to disable personalised recommendations. They also prohibit unreasonable differential treatment in prices or other trading conditions based on characteristics such as purchasing habits.
Chinese regulators have specifically identified practices including algorithmic discrimination, excessive consumption and the use of personal information as areas requiring oversight. Providers are expected to assess their algorithms and maintain systems protecting data and personal information.
Those requirements are becoming more consequential as AI systems progress from recommending content to taking action. An algorithm that determines which advertisement appears on a screen has influence; an AI agent that chooses among products, arranges payment and organises delivery has considerably more.
For China's retailers, the commercial opportunity is enormous. AI could reduce the friction between wanting something and buying it, transforming shopping from a sequence of searches and clicks into a conversation followed by an automated transaction.
But it could also make retailers increasingly dependent on the technology companies controlling those conversations. Brands that once worried about appearing on the first page of search results may soon face a different question: whether an AI assistant considers them worth mentioning at all.
That prospect makes artificial intelligence more than another feature being added to Chinese e-commerce. As shopping agents acquire the ability to search, evaluate and eventually purchase goods for consumers, AI is becoming the new gatekeeper between China's shoppers and the companies trying to reach them.
