While the West fixates on the conversational capabilities of large language models, China is quietly pioneering the next paradigm of the digital economy: agentic commerce. A new analysis published in the Harvard Business Review argues that China’s unique digital infrastructure—specifically its tightly integrated super-app ecosystem—gives it a profound structural advantage in deploying autonomous AI agents that execute complex, multi-step transactions on behalf of users.
The research, authored by Mark J. Greeven, Fabrice Beaulieu, and Wei Wei, suggests that the future of e-commerce is being written not in Silicon Valley but within the walled gardens of Chinese tech giants such as Meituan, Alibaba, and Tencent.
From Chatbots to Execution Agents
The HBR analysis centers on the evolution of AI from conversational interfaces to execution engines. The authors highlight Meituan’s “Xiaomei” AI agent, launched in late 2025, as a prime example. Meituan executives internally describe Xiaomei not as a chatbot but as an “orchestrator plus execution agent.”
The distinction is critical. A user does not simply ask Xiaomei for restaurant recommendations; they delegate the entire task. A command like, “Order my usual lunch, but deliver it 20 minutes later today,” requires the AI to interpret intent, access historical preference data, interface with the restaurant’s ordering system, process the payment, and coordinate with the delivery logistics network—often with zero screen interaction from the user.
This level of seamless execution is exceedingly difficult to achieve in the West, where the digital ecosystem is highly fragmented across disparate apps for search, payment, delivery, and communication. In China, the super-app model consolidates these functions into a single architecture, providing the ideal environment for AI agents to operate autonomously.
The Open-Source Advantage
The rapid deployment of these agents is being fueled by China’s aggressive embrace of open-source AI. The HBR piece cites estimates from Andreessen Horowitz (a16z) indicating that 80% of developers worldwide who use open-source AI tools are now building with Chinese models.
Alibaba’s Qwen family of models has surpassed 700 million downloads, becoming the most widely adopted open-source AI system globally. In some weeks during 2025, Chinese open models accounted for up to 30% of all AI usage worldwide. This widespread adoption creates a massive feedback loop, accelerating the refinement of the models that power China’s domestic agentic commerce ecosystem.
Structural Moats and Global Implications
The implications of this shift are profound. As AI agents become the primary interface for commerce, the value of traditional search advertising and app-based discovery diminishes. Brands will increasingly need to optimize their visibility not for human consumers, but for the algorithms that make purchasing decisions on their behalf.
The HBR authors argue that China’s structural moat in agentic commerce will be difficult for Western platforms to replicate. The integration of payment rails, logistics networks, and consumer data within platforms like WeChat and Alipay creates a frictionless environment that AI agents require to function effectively. As China continues to subsidize its foundational AI infrastructure and deploy AI across physical and digital domains, the gap in commercial AI applications is widening. The West may hold the edge in raw frontier-model capabilities, but China is rapidly defining how those models will be used to conduct business in the real world.