The Agentic Commerce Race: AI Agents Are Learning to Shop for You
From Google's Universal Commerce Protocol to Alibaba's Qwen chatbot completing purchases, a global race is on to let AI agents handle your shopping. Here's what's happening and why it matters.
A new front has opened in the AI wars — and it's happening at checkout.
Over the past few weeks, the world's biggest tech companies have been racing to turn their AI assistants into fully autonomous shopping agents. Not just "here are some recommendations" assistants, but agents that can browse, compare, negotiate, and complete purchases on your behalf. The era of agentic commerce has arrived.
Google Builds the Rails
On January 11, Google launched the Universal Commerce Protocol (UCP), an open-source standard that lets AI agents execute purchases across different retail platforms without needing custom integrations for each merchant.
The initiative has serious backing. Google co-developed UCP with Shopify, Etsy, Wayfair, Target, and Walmart, with endorsements from over 20 additional companies including Visa, Mastercard, Stripe, American Express, Best Buy, and Zalando.
"AI agents will be a big part of how we shop in the not-so-distant future," wrote Google CEO Sundar Pichai on X.
UCP doesn't work alone. It's part of a stack of complementary protocols:
- Model Context Protocol (MCP) — originally developed by Anthropic and donated to the Linux Foundation — connects AI applications to external data sources and tools
- Agent2Agent (A2A) — Google's protocol for AI agents built on different frameworks to collaborate on complex tasks
- Agent Payments Protocol — handles cryptographic transaction authorization
Together, these create the infrastructure for AI agents to discover products, handle checkout, link customer identities, and manage post-purchase workflows across platforms.
China's Super Apps Go Agentic
While Google is building open protocols, Chinese tech giants are taking a different approach: embedding agentic commerce directly into their existing super app ecosystems.
Alibaba recently upgraded its Qwen AI chatbot to let users complete entire transactions without leaving the chat interface — ordering food, booking flights on Fliggy, comparing products on Taobao, and paying through Alipay. Previously, Qwen could make recommendations, but users still had to manually navigate to different platforms to buy.
ByteDance followed suit in December, upgrading its Doubao chatbot to autonomously handle ticket bookings through integrations with Douyin (China's TikTok). The upgraded Doubao was even introduced on a ZTE prototype smartphone as a comprehensive AI assistant that operates across the entire device.
Tencent's president Martin Lau has signaled that AI agents will become core components of the WeChat ecosystem, which already serves over 1 billion users as China's dominant super app.
"AI agents will be foundational to the evolution of super apps, with success depending on deep integration across payments, logistics, and social engagement," Charlie Dai, VP and principal analyst at Forrester, told CNBC.
Two Models, One Goal
The contrast in approaches is striking. Google is playing infrastructure provider — building open standards that any platform can adopt. Amazon, by contrast, is blocking third-party AI agents to protect its advertising revenue, while deploying its own Rufus AI shopping assistant (now with 250 million users).
Chinese companies benefit from integrated ecosystems, rich behavioral data, and consumer familiarity with super apps. Western companies lead in foundational AI models and global reach, but face more fragmented data and stricter privacy regulations.
What This Means for Brands
Here's the part that should worry (or excite) every business: according to Kantar's latest research, 24% of AI users already deploy AI shopping assistants, and 75% of regular AI assistant users seek out AI-driven recommendations.
This creates an entirely new optimization challenge. Forget traditional SEO — welcome to GEO (Generative Engine Optimization). If an AI model doesn't know your brand, it won't recommend your brand. Products and services need to be machine-readable, well-structured, and widely referenced across digital environments.
As Kantar puts it: "If you're not the default recommendation, you'll be optimized out."
The Bottom Line
We're watching the real-time construction of a new commerce layer for the internet. Google is building open protocols, China's super apps are building walled gardens, and Amazon is somewhere in between. The common thread: AI agents that don't just advise, but act.
The winners of this race won't just have the best AI models. They'll have the best plumbing — the protocols, payment rails, and merchant integrations that let agents actually get things done.
For consumers, the promise is compelling: tell your AI what you need, and it handles the rest. For businesses, the message is clear — adapt your digital presence for agent discovery, or risk becoming invisible in the age of agentic commerce.
Sources: CNBC, Google/UCP, PPC Land, Kantar Marketing Trends 2026