The e-commerce industry in 2026 seems more bustling than ever. Since the start of the year, major platforms have simultaneously accelerated their pace, deeply integrating AI models with their own ecosystems: for example, Alibaba’s Qianwen app integrated “Taobao Flash Purchase,” using 3 billion yuan in subsidies to drive over 120 million AI orders; JD.com launched a standalone app “JD AI Shopping,” aiming to create a native AI shopping entrance fully driven by conversation, without traditional shelves; ByteDance’s Doubao began gray-scale testing shopping features, allowing users to complete payment and ordering without jumping to Douyin after a conversation... At the same time, the Ministry of Commerce and five other ministries jointly issued a document elevating “AI + E-commerce” to a strategic level. The wave of “e-commerce AI transformation” is no longer a trial but a direction. So, what impact will this AI-driven industry revolution bring to traditional e-commerce? With this question, the author visited several frontline operators deeply engaged in the e-commerce field. After exchanges, it became clear that in this transformation, some have found new opportunities, while others are being eliminated. AI Shopping Is Fierce, but Search-Based Ordering Remains Mainstream If one only looks at the high-profile moves of giants and public opinion, it’s easy to conclude that AI is disrupting e-commerce. However, the actual situation is that multiple practitioners stated that currently, conversational AI shopping has not caused any substantial impact or user diversion on traditional search-based ordering, let alone replacement. “At present, whether it’s traditional platforms like Tmall and Taobao, or content e-commerce like Douyin and Kuaishou, core orders still come from users’ active searches,” said Muzi (pseudonym), who has been an e-commerce operator in Shandong for nearly 10 years. Data also confirms this. According to public reports, the GMV share of AI conversational ordering in overall domestic e-commerce transactions is still less than 10%, and orders are highly concentrated in low-ticket, high-repurchase scenarios like food delivery and standard FMCG products. Traditional search-based ordering still dominates the industry, and the core transaction landscape has not been shaken. Why hasn’t the fierce AI shopping brought actual impact to traditional e-commerce? The primary reason is that conversational ordering and search-based ordering serve two completely different needs. “Conversational ordering is actually more suitable for users who seek convenience and efficiency and are not very price-sensitive.” Muzi added, but more often, consumers’ decision-making processes require extensive information comparison, review checking, and price weighing. This screening includes not only products but also merchants and platforms. “This is also the limitation of conversational ordering: current mainstream AIs all operate based on their respective platform ecosystems. For example, Doubao’s product supply relies on Douyin e-commerce, and Qianwen only connects to Alibaba’s entire ecosystem, making it difficult for users to achieve one-stop cross-platform comparison.” Consumer psychology research suggests that consumers’ need for autonomy in purchase decisions usually far exceeds their need for the optimal solution, and the pleasure derived from the “cloud shopping” process itself can even surpass the purchase outcome. The one-step solution brought by AI recommendations precisely deprives this process value. Of course, this is not to deny AI shopping. At least from the perspective of category characteristics, the difference between standard and non-standard products directly delineates the boundaries of the two shopping modes. “Products like beverages, snacks, and daily necessities are standardized and branded, so they can be purchased with confidence without comparison, suitable for conversational ordering; products like custom furniture, clothing, shoes, bags, and auto parts that require detailed models, scenarios, and specifications rely more on consumers’ own spending power and aesthetic preferences,” a category specialist gave an example: “But AI-recommended products not only have quantity limits but also still need refinement in accuracy.” Additionally, there is a more practical reason why practitioners generally believe AI shopping will not bring significant impact: AI e-commerce has not escaped the traditional traffic payment logic of e-commerce—buying keywords. “From web ad slots to short-video marketing, and now to AI recommendations, merchants just have an additional traffic investment entrance.” Same Old Story? Merchants Begin to Compete for the First Touchpoint in the AI Era In the past, e-commerce promotion was based on traffic distribution. Whoever occupied the search homepage and secured keyword rankings would grab orders, prompting merchants to invest huge budgets to compete for entrances; in the ideal AI era, users would only need to describe their needs to AI, which would automatically filter out marketing ads and provide products that truly match their needs. But some keen business opportunities have already begun to lay out in the new battlefield. The most obvious change is the rapid rise of white-hat GEO (Generative Engine Optimization) services. Screenshot of a company’s GEO backend “What is GEO service? For example, when a user asks AI ‘Which dried mango is the most delicious?’ As long as you are a merchant on the corresponding platform, the product you get recommended first will be yours.” According to an employee of a major internet company, if your question has not been paid for by anyone, then the content recommended at that time is what AI answers based on comprehensive network information and data. The author learned that this white-hat GEO is fundamentally different from black-hat GEO (AI poisoning). Taking the major company’s package as an example, for an annual fee of only 10,000 to 40,000 yuan, you can purchase 5-20 keywords (such as industry words, regional words, brand words) on mainstream platforms like Doubao, DeepSeek, Qianwen, and WeChat AI, and also get hundreds of effective long-tail keywords as a bonus. The team will manually write articles based on AI platform algorithms and mechanisms, and the fed corpus and articles comply with platform ranking mechanisms, with continuous follow-up optimization. But one point to note is that GEO is not applicable to all industries and categories. Currently, the main applicable types of this service are: first, local life merchants targeting offline, where user needs are clear and decisions are simple, such as restaurants, car washes, and pet shops; second, standardized products targeting online, such as some FMCG products, with low unit price and high volume, where efficiency-seeking users will not and usually do not need to compare too much. Although the exploration of AI marketing is still in its infancy, at least for now, the results consumers see on AI can already be controlled by merchants. “AI will not impact the transaction landscape of e-commerce; what is truly impacted is ourselves,” Muzi said with a sigh. Laid Off, Marginalized, Replaced Traditional E-Commerce Organizational Structures Undergo a Major Reshuffle While platforms and merchants are still competing for traffic entrances in the AI era, traditional e-commerce practitioners are experiencing an unprecedented career crisis. The first to bear the brunt are basic positions such as customer service, graphic design, and livestream hosts. Customer service teams that once had dozens of members have become the first roles to be replaced on a large scale by AI. “One Alibaba-affiliated merchant we work with saw its customer service team shrink from 80 to 10 people in half a year,” said an e-commerce operator: “The remaining 10 are now used to handle complex issues that AI cannot answer, but their departure may be only a matter of time.” The reason is simple: many companies’ AI customer service can already handle over 90% of standardized inquiries, such as tracking logistics, urging payment, and handling basic returns and exchanges, available 24/7, at only one-tenth the cost of human labor. When efficiency and cost gaps widen, replacement is almost inevitable. The same changes are happening in graphic design roles. According to iResearch’s 2025 retail industry AI application report, 67% of domestic e-commerce companies have deeply integrated AI image generation tools into their daily workflows, with core processes like main images, detail pages, and promotional posters fully covered by AI. In practice, using e-commerce-specific AI tools, multiple compliant design solutions can be generated in about 2 minutes, crushing human efficiency while the marginal cost per design is almost negligible. For many full-time graphic designers, even if they escape layoffs, those who remain are no longer just making images but must transition to high-value work like AI prompt optimization and visual strategy, or they still face the risk of being eliminated. Livestream hosts, who play a core sales role, have also not escaped AI’s impact. As early as last year, AI digital human livestreaming has become a low-cost standard in the e-commerce industry. “Although AI hosts lack emotional communication and flexibility, and their conversion ability is far inferior to real people, they have the advantage of not needing rest, no scheduling, and low monthly operating costs, so merchants are keen to use them for mass distribution and capturing long-tail traffic,” A host from a personal care brand said that AI is currently more used for categories that can be explained in a standardized way, and cannot yet handle livestreams that require high conversion. But this has also led to a significant reduction in demand for novice and part-time hosts, and many small and medium hosts are also stuck with no orders. “They originally earned money through low barriers and time investment, but that is precisely the part most easily replaced by AI.” What is more worrying is that although AI collaboration still has shortcomings, its evolution speed far exceeds the career growth speed of humans. In March 2026, DingTalk officially launched the enterprise-level AI-native work platform “Wukong,” introducing the OPT one-person team solution. In the e-commerce field, Wukong can complete the entire process in one stop, including product selection and sourcing, supplier background checks, cross-platform price comparison, product listing, and operational assistance; Alibaba International’s Accio Work, launched at the same time, directly proposed the “one person + AI = five-person cross-border team” model, extending AI’s labor substitution effect from domestic e-commerce to the entire cross-border chain. Behind this, what changes is not just efficiency but also the redefinition of e-commerce organizational structures—AI is making the “one-person company” feasible. During the visits, some e-commerce bosses have already said that in the future, they will not cultivate traditional operators, not wanting to spend half a year teaching someone to do a set of actions that can already be automated. As repetitive work is continuously taken over by machines, the positions that truly remain are concentrating in two directions: either “those who can design processes” or “those who can amplify results.” As for those positions that remain at the execution level with highly standardized actions, they are being rapidly squeezed out of the system. Final Thoughts In 2026, discussions about traditional e-commerce within the industry seem increasingly pessimistic, and there is no shortage of anxiety swept up by AI. From platforms accelerating the layout of AI shopping closed loops, to merchants competing for AI recommendation slots, to grassroots positions being replaced in batches, this seemingly fierce AI wave is not actually trying to overthrow anything; it is more like a pressure that forces the entire industry to redo its past operational logic. Some treat AI as a tool to amplify existing capabilities and find a place in the new traffic entrances and efficiency systems; others are trapped in old experiences, continuously losing speed in the same division of labor and paths. It is not hard to understand why AI is first erupting in e-commerce scenarios today, because e-commerce itself is a business that is highly digitalized, highly content-driven, and heavily reliant on traffic distribution and efficiency collaboration. For FMCG companies, the impact of AI will also become increasingly common.

  • For brand owners, AI is changing content production, consumer insights, precise targeting, terminal management, and sales forecasting;
  • For distributors, AI can help bosses re-understand their own business, truly utilizing operational data such as finance, procurement, warehousing, business, and stores;
  • For retailers, AI is also entering core processes such as product selection, display, replenishment, membership operations, and store efficiency improvement. If you are also thinking:

Will AI really change the FMCG industry?

Where should brand owners, distributors, and retailers start using AI?

Which positions, processes, and operational actions within the enterprise are most worth redoing with AI first? Welcome to the “2026 China FMCG Conference AI Application Forum” to see cases, learn methods, and exchange ideas.