How AI is changing marketing in China

Chinese consumers are among the world's most engaged users of AI. A survey last quarter found that almost 80% of Chinese consumers ask AI tools for help with purchase decisions . This has transformed the age-old marketing goal from being 'seen by users' to being "recommended by AI'. Much of the foundational work marketers do to inform AI recommendations remains similar to pre-AI, such as building credibility through influential online sources . However, there are also some fundamental shifts in how marketers need to think about their target audiences and communicate what they are selling. Some of these changes were highlighted last week at the annual M Summit in Hangzhou , hosted by Alibaba's advertising and marketing arm, Alimama. A shift in how to define your target audience Anyone who has shopped online will be familiar with the old 'you may also like' recommendation, which shows shoppers products purchased by people with similar profiles to their own. This is known as 'behaviour-based matching' and has long influenced how marketers define their target audiences . They use historical data to build a profile of the customers who buy the type of product or service being sold, then target more people who resemble them. There are two main limitations to this approach. First, it takes a rear-view mirror view and struggles to account for rapidly changing consumer behaviour. Consumers develop new interests or abandon old ones, gain or lose partners, become parents, move house or experience another major life change. Second, almost every brand selling similar products ends up targeting similar customers, pushing up acquisition costs without materially growing the market. Behaviour-based matching can still be valuable, but AI has considerably broadened audience profiling by anticipating future needs rather than just repeating historical ones. At the conference, Alimama used bed sales as an example. Furniture companies have typically targeted similar consumers: newlyweds, home renovators and people moving house. Alimama's new tool, Wanxiang Dianjing, identified another potential audience: 18- to 28-year-old online gamers. Gamers are an enormous market in China, representing over half of China's population , and often spend heavily to optimise their gaming setups. This includes expensive ergonomic chairs, large monitors, lighting and other furnishings. So it makes sense that they would also want a comfortable bed after a long night of gaming. This sort of targeting was not unheard of before AI, we did plenty of it at China Skinny. But it is now becoming much more common, particularly among brands already investing heavily in targeted advertising. The shift from product searches to real human situations The way consumers find new products and services is also turning traditional marketing and advertising on its head. Chinese consumers have not been heavy search engine users for a long time. For the past 15 years or so, when searching for a product, they have often gone directly to a platform such as Taobao , selected a category and possibly a brand, then entered a fairly precise phrase describing what they wanted. Increasingly, consumers are simply giving AI shopping assistants loosely defined requests. Rather than searching for a particular bottle, a mother might say, 'I often take my baby hiking and need a bottle suitable for the outdoors.' Traditional keyword targeting does not work as well because the consumer has not named a product feature, model or brand. Search tools that once expected consumers to translate their lives into ecommerce categories are now translating consumers' lives into possible purchases. What brands should do differently to appear in AI search results What we like about AI is that, ironically, it makes online marketing a little more human. This isn't just selling products that aren't perfect , it also means moving away from stuffing marketing messages with keywords intended to fool algorithms. Brands should instead give AI recommendation tools more useful material to work with. This means ensuring that product listings, livestream scripts and short-form videos explain the situations in which a product is useful, the problems it addresses and the trade-offs it removes . A technically detailed product page remains important, particularly when someone is already comparing models. But AI also needs context to identify a real-life usage scenario or an unconventional customer. AI can help identify these opportunities, but so can human intuition and judgement.