China's major e-commerce platforms have repositioned AI as the centrepiece of their 618 shopping festival strategy, moving beyond the discount-formula complexity that previously dominated the event. Rather than emphasising traditional promotions, Alibaba, ByteDance, JD.com, and Pinduoduo have integrated AI across the entire customer journey—from product discovery through Qianwen and Doubao assistants, to 24/7 digital human livestreamers generating over 70 million yuan in sales within JD.com's first four hours, to post-purchase logistics optimisation. Taobao's upgraded "Dian Xiaomi" customer service now actively recommends products and intercepts fraudulent refunds with over 20% success rates, whilst JD.com's supply chain AI has improved inventory turnover by 30-40%. Yet beneath these impressive metrics lies a more sobering reality: AI has primarily accelerated existing processes rather than fundamentally reshaping the e-commerce model, and consumer adoption of AI-driven purchasing remains limited, with only 37-48% accepting automated purchase suggestions compared to 65-70% acceptance for information-organisation tools.
The divergence between platform narratives and ground-level experience reveals why this AI deployment matters for CX teams. Merchants report that AI customer service excels at handling standard queries but fails catastrophically during peak periods—precisely when support volume surges and human intervention becomes critical. The responsibility gap is particularly acute: when AI provides incorrect guidance leading to disputes, merchants bear the consequences whilst platforms capture the efficiency gains and data. This creates a structural problem for support operations: as teams implement AI-powered tools, they must simultaneously maintain human escalation pathways that actually connect customers to representatives, not trap them in loops requesting transfers. The question for CX leaders is whether their AI implementations are genuinely reducing support burden or simply shifting costs whilst degrading experience during high-demand periods when customer frustration peaks.
What emerges is a platform strategy less about consumer benefit and more about ecosystem lock-in and competitive moat-building. Alibaba and ByteDance are using proprietary large language models to contain consumption decisions within their ecosystems, whilst JD.com leverages AI to strengthen supply chain advantages. Consumers have not demanded this shift—surveys show they prioritise cost reduction and supply chain reliability over AI-assisted purchasing—yet platforms are collectively betting on AI because the traditional major-promotion model has exhausted its growth potential. For CX professionals, this signals a critical inflection point: the platforms investing most heavily in AI are doing so to consolidate data and decision-making power, not to improve customer outcomes. Teams implementing similar strategies should interrogate whether their AI investments serve customer needs or primarily serve platform interests, and whether they're prepared for the inevitable gap between promised efficiency and actual performance when systems encounter edge cases, high volume, or customer frustration that requires genuine human judgment.
Big Techs' Intense AI Battle in the 618 Shopping Festival 36 Kr