2026年电子商务AI报告

Introduction environments, forcing businesses to rethink discovery,trust and visibility in an AI-mediated market. The second, Experience AI, focuses on the AIsystems retailers and brands build for customersdirectly. Here, the challenge is redesigning digitalcommerce experiences around simplification,confidence and guidance. AI is increasingly AI has rapidly become the defining conversationin modern commerce, yet many businesses aresimultaneously struggling to separate meaningful This uncertainty exists because AI is not arrivingas a single technology shift, instead reshapingcommerce from multiple directions simulataneously.Consumers are beginning to use external AI toolssuch as ChatGPT, Gemini and Perplexity to discover,compare and evaluate products before they ever The third pillar, Capability AI, examines theoperational transformation taking place insideorganisations. Increasingly, customer experience and Finally, Strategic AI looks beyond individual toolsand deployments towards the wider restructuring ofcommerce itself. As AI reshapes customer behaviour, Internally, businesses are embedding AI intooperations, forecasting, logistics, content productionand decision-making in ways that fundamentallyalter how organisations function. At the same But this report does not argue that AI is a magicsolution. Nor does it suggest that every businessmust become a technology company. What it does These are the four pillars of Commerce AI and it isthese that this report explores. The organisations that succeed are unlikely to bethose with the loudest AI marketing or the mostexperimental pilots; it’ll be those that are capable ofusing AI to create more adaptive operations, more The first pillar, External AI, examines how AI assistants,conversational search and emerging agentic shoppingsystems are changing the balance of power betweenconsumers, retailers and platforms. Increasingly, the Pillar 1: External AI The issue of how consumers use AI –and what AI they use – is foundationalto commerceAI. The data shows thatshoppers have embraced the technology, TOOLS THE CUSTOMER BRINGSThe first pillar of commerce AI is how consumers use the technology as part of their shopping activities– a fundamental that is reshaping retail right here,right now. The first and most important point inthe data is that consumer AI shopping behaviouris not futuristic. It is already here, but it is unevenly This distinction matters because it shows where theimmediate commercial disruption lies. The first waveof external AI in retail is less about robots silentlybuying groceries in the background and more about The picture changes when consumers retainoversight (see chart overleaf). Some 57.8% would behappy for an AI assistant to add items to an onlinebasket so long as they can review and approve thebefore paying. Some 60% would like an AI assistant This is a significant minority willing to experimentwith autonomous replenishment, but it is not yet amainstream licence for AI to buy unchecked. Similarly,43.2% say they would prefer an AI assistant helping The chart opposite shows this tension clearly. Only45.9% of consumers agree or strongly agree thatthey would be comfortable allowing an AI assistant This is the emerging boundary of consumerpermission: consumers are more willing todelegate tasks than decisions. They will allow AIto watch, compare, shortlist, prepare and eventrigger pre-authorised rules, but they remain more This impression of caution is reinforced by the chartson page 5. Some 65% of consumers agree or stronglyagree that they would only allow an AI assistant toact for them if it is managed by a brand they alreadyknow and trust. A further 61.7% would only allow anAI assistant to complete a purchase if the retailer orbrand is one they already know and trust. This tells Trust is no longer only a consumer-facing emotionalasset expressed through advertising, site experienceand service. It becomes a machine-mediatedpermission structure. The trusted brand is more The chart on page 7 underlines the same point. Avery large majority, 75.6%, say they would not wantAI tools to make purchase decisions on their behalf and value discovery. They are not yet whollycomfortable with AI becoming an independentcommercial actor. For retailers, this means that thenext phase of AI commerce is likely to be hybrid.Consumers will use AI agents to reduce effort, but CONSUMER COMFORT ZONEThe chart above shows where consumers are already most comfortable. Some 58.1% are okaywith retailers using their data so AI tools can providemore personalised product recommendations. Some This is the key caveat to much of the currentexcitement around agentic commerce. Consumersare interested in better information, time saving product features and prices across different retailers.Meanwhile 57% use AI to identify product categoriesor brands that might meet their needs when they arenot sure where to start. These are classic discovery The strongest sign