Nearly six years ago Walmart sent 500 shelf-scanning robots home after concluding that workers picking online orders could see the
Conversations On Retail
July 20, 2026
Artificial intelligence has been part of digital commerce for years. It improved search relevance, refined recommendations, and helped personalize promotions. In every case, the technology supported the shopper, but the shopper still did the work.
That boundary is starting to shift.
A new class of AI systems, commonly described as agentic AI, is designed to carry out multi-step tasks on behalf of users. In a commerce context, this means software that can interpret intent, evaluate options, and complete transactions across services with limited human involvement.
McKinsey characterizes this shift as a move from AI that informs decisions to AI that orchestrates them, particularly in consumer workflows such as shopping and travel. While still early, the transition is no longer theoretical. In China, agent-led commerce is already being deployed inside platforms used by hundreds of millions of consumers.
China’s lead in agentic commerce is not driven by superior AI models alone. It is rooted in platform design.
Chinese consumers are deeply accustomed to super apps that combine messaging, payments, shopping, travel, and local services into a single interface. Platforms such as WeChat, Alipay, and Douyin are embedded into daily routines, giving AI systems access to payment rails, fulfillment infrastructure, and behavioral signals without requiring complex integrations.
Alibaba’s recent update to its Qwen app illustrates this advantage clearly. The AI assistant can now guide users through product discovery and service selection, then complete purchases through direct integrations with Taobao, Fliggy, and Alipay. What once required switching between multiple apps can now happen within a single conversational flow, according to reporting from CNBC and Barron’s.
This is a meaningful shift. The greatest friction in digital commerce has rarely been discovery. It has been execution. By embedding AI directly into transaction infrastructure, platforms increase the likelihood that intent turns into a completed sale.
Alibaba is not moving alone.
Tencent has said it is increasing investment in artificial intelligence, including expanding AI capabilities within WeChat. While the company has not publicly detailed full agent-driven commerce workflows, its focus on AI inside its core consumer platform signals a clear intent to deepen automation over time.
ByteDance is also pushing aggressively. Its Doubao assistant has been tested alongside commerce and local services connected to Douyin, China’s version of TikTok. At the same time, ByteDance has explored deeper integration between AI assistants and smartphone operating systems through prototype devices developed with partners such as ZTE. Some of these experiments have drawn scrutiny around data use and privacy, leading the company to adjust how certain features are deployed.
Across these companies, the pattern is consistent. Rapid experimentation, fast iteration, and deep integration into platforms where consumers already transact daily.
Agentic AI is often discussed alongside robotics, cybersecurity, and enterprise automation. Commerce, however, has emerged as one of the first large-scale applications.
The reason is practical. Shopping decisions tend to follow repeatable patterns. Product data is structured. Outcomes are measurable. For AI systems designed to execute tasks, these conditions are favorable.
McKinsey estimates that by 2030, agentic commerce could help orchestrate up to approximately $1 trillion in annual U.S. B2C retail revenue, with global projections reaching several trillion dollars as adoption expands. These figures are projections rather than guarantees, but they help explain why commerce platforms are prioritizing agent-led workflows early.
Western technology companies are also pursuing agentic commerce, but under different constraints.
OpenAI has introduced Instant Checkout within ChatGPT for U.S. users, beginning with purchases from Etsy sellers and with expansion to Shopify merchants planned. Google has announced the Universal Commerce Protocol, an open standard designed to allow AI agents to connect consumers, merchants, and payment providers across platforms.
These efforts point toward a more federated approach, where AI agents operate across multiple services rather than inside a single, tightly integrated ecosystem.
Privacy regulation, fragmented data environments, and decentralized payment systems have slowed deep integration compared to China’s super apps. At the same time, Western companies emphasize interoperability and governance, reflecting different regulatory and cultural expectations.
For retailers and consumer brands operating globally, this divergence matters. Agentic commerce is unlikely to look identical across regions, even if the underlying technology converges over time.
Agentic commerce is not simply another digital channel.
When AI systems execute purchases, they become intermediaries between brands and consumers. That changes how products are discovered, compared, and selected.
Product data quality becomes critical. Agents rely on structured attributes, availability, pricing, and fulfillment confidence. Brands with incomplete or inconsistent data risk being filtered out before a human shopper ever enters the process.
Differentiation also shifts. Unless guided otherwise, AI agents may optimize for price, delivery speed, or reliability. Brand equity still matters, but it must be expressed in ways machines can interpret and prioritize.
First-party data becomes increasingly valuable. Retailers and brands that understand customer preferences, constraints, and lifetime value will be better positioned to influence how agents behave within their ecosystems.
Agentic commerce is not yet the default shopping experience, even in China. Many consumers still prefer manual control, and regulatory scrutiny around AI autonomy is increasing worldwide.
What is clear is direction.
As AI systems move from assisting decisions to executing them, commerce becomes less about where consumers shop and more about how decisions are made on their behalf. China’s super apps offer an early view of that future, but the implications extend across global retail and consumer goods markets.
For industry leaders, the most important question is not when agentic commerce fully arrives. It is whether their organizations are preparing for a world where software increasingly participates in, and sometimes completes, the buying journey.