The rules of e-commerce have changed. Just a day ago, the standard online shopping process looked like this: search → browse → add to cart → checkout. Today, Google has streamlined it into: AI → AI → AI… going all-in on AI. It's clear Google has figured out the e-commerce game.
CEO Sundar Pichai officially announced the launch of Google's Agentic e-commerce solution—ranging from a new protocol standard called UCP to Gemini CX for enterprises, with AI Agents directly integrated throughout the entire shopping journey. Great, just great. This means during future Singles' Day sales, we can finally avoid manually comparing prices across multiple platforms and collecting coupons to meet discount thresholds, freeing up our hands to be noble, intelligent single individuals!
Elon Musk was quick to comment: "Interesting!" So, does this mean we could soon order a Tesla directly via AI?
So, what exactly is happening? Let's break it down one by one. First, let's look at the most critical part of Google's announcement—the UCP (Universal Commerce Protocol). This is an open protocol specifically designed for Agentic e-commerce, aiming to enable seamless collaboration between AI Agents, merchants, and e-commerce platforms to handle the entire process from product discovery and purchase to after-sales service. It will later be integrated into Google Search's AI Overviews and the Gemini app.
For example, if you want Jensen Huang's signature leather jacket, you can simply type your request into a dialog box. The Agent will automatically use UCP to query e-commerce platforms for relevant information and present the recommended results to you. Once you click to confirm the purchase, the Agent will continue by using UCP interfaces to call the merchant's order placement API and complete the transaction using your authorized Google Pay. Throughout the entire process, you won't need to navigate to any external web pages; you can just sit back and wait for delivery.
By the way, it will also help you claim coupons automatically, ensuring you get the lowest possible price. UCP's initial release focuses on three core functions: checkout, identity linking, and order management. Specifically: Checkout: Supports complex shopping cart logic (like multi-item combinations and inventory synchronization), dynamic pricing (such as member prices and real-time discounts), and tax calculations (including adjustments for cross-border tax rates), catering to both small merchants and large enterprises. Identity Linking: Allows agent platforms, including Google's, to legally obtain user permissions to perform corresponding operations on their behalf. Order Management: Provides users with detailed order information, including purchased items, delivery methods, and subsequent real-time shipping updates, while also supporting returns and exchanges.
UCP is compatible with existing industry protocols, such as Agent2Agent (A2A), Agent Payments Protocol (AP2), and Model Context Protocol (MCP). It has already integrated several major retailers and payment platforms, including Walmart, Shopify, and Visa. Going forward, Google also plans to launch a Business Agent, allowing users to interact directly with brands within Search, similar to a virtual customer service agent, to help users learn more detailed product information.
On the enterprise side, Google simultaneously launched Gemini Enterprise for Customer Experience (Gemini CX for short). It integrates Google's latest Gemini models and AI technologies, enabling businesses to quickly deploy Agents that cover the entire customer service lifecycle—from product discovery and ordering to after-sales support and cross-channel assistance.
It comes equipped with a brand-new Shopping Agent. Unlike simple chatbots, the Shopping Agent can directly connect front-end interfaces like chat and voice to back-end tools, including: Complex Reasoning Capabilities: It doesn't just filter products based on user keywords; it can understand more nuanced user needs (such as size, durability, budget) and automatically select the most suitable products. Multimodal Interaction: Supports text, voice, image, and video inputs. For instance, a user can take a photo of a handwritten recipe, and the AI can accurately identify the required ingredients and add them to the cart automatically. Authorized Action Execution: After obtaining explicit user consent, the AI can directly perform actions like adding items to the cart and completing payment.
By consolidating fragmented retail needs into a single platform, it reduces the burden on merchants who previously had to manage multiple tools simultaneously. It supports various Agentic scenarios, including intelligent shopping assistance, business operations, and automated customer service bots, and can be freely integrated into a company's own systems. Currently, major companies like McDonald's and The Home Depot have already implemented it to optimize their customer service quality and better meet consumer demands.
It's not just Google; domestic e-commerce platforms are also targeting this sector. A prime example is Alibaba. This year's Singles' Day also marked the first large-scale application of generative AI by Taobao and Tmall. On one hand, they upgraded their search and recommendation engines based on the Qwen model, using AI-driven mechanisms to distribute coupons and optimize the shopping experience for consumers. On the other hand, they created AI teams for merchants, where AI designers assist in mass-producing marketing materials, and AI business advisors provide customized suggestions like dynamic pricing and inventory optimization based on a merchant's operational situation.
Furthermore, Alibaba's 1688.com has gone all-in on AI, recently launching the cross-border e-commerce AI Agent "Ao Xia," which focuses on "AI + supply chain" to assist merchants with the entire process from product selection and procurement to listing. During the Singles' Day period, JD.com also made over twenty AI tools available to merchants, including digital avatar livestreaming, covering various aspects like store management and marketing promotion. Douyin has also revamped its e-commerce entry point based on its Doubao large model. Users can simply ask shopping-related questions in a chat window, and Doubao can accurately match corresponding products and provide links redirecting to the Douyin mall, significantly shortening the path from user inquiry to purchase.
The wave of "AI + Everything" is no longer just a slogan.
Comments