The Universal Commerce Protocol: When AI Stops Recommending and Starts Buying
If the click was the currency of ecommerce, inflation just hit. AI agents are starting to shop for consumers, collapsing discovery, comparison, and checkout into a single conversation.
Google’s Universal Commerce Protocol (UCP) is the infrastructure layer for that shift. In simple terms, UCP is the system that lets AI assistants act like a shopper — finding products, checking availability, and completing purchases without sending consumers to a website. Instead of one-off integrations for every AI platform, UCP standardizes how AI and commerce systems connect.
For example, UCP will soon power a checkout feature on eligible Google product listings in AI Mode in Search and the Gemini app, allowing shoppers to complete purchases from eligible U.S. retailers as they’re researching. Instead of browsing multiple product pages, an AI assistant could say: “This couch from Target is in stock, fits your budget, and can arrive Friday. Want me to buy it?” If the user agrees, the purchase happens directly within the AI experience.
And this is moving fast. Walmart, Target, and The Home Depot are already aligned, signaling that agentic commerce is shifting from emerging tech to core retail infrastructure.
Clicks are losing their power
AI is no longer just influencing purchase decisions — it can now act on intent. This will fundamentally change how brands approach discovery, conversion, and advertising. Structured, rich product data, clear pricing, and real-time availability are now as important as creative and media spend. Brands that invest in these areas will surface more often in AI-driven shopping experiences, while those that are unprepared risk being invisible in a world where AI intermediates discovery and transactions. This doesn’t eliminate websites, paid media, or PDP optimization — but it adds a new layer where AI becomes the primary decision-maker in the funnel.
Control without compromise
Despite automation, retailers maintain full control as the merchant of record. AI executes transactions within defined guardrails, allowing brands to scale agentic commerce without ceding ownership of the customer relationship. New merchant data attributes also improve discoverability in conversational commerce, making it easier for AI agents to surface the right products at the right time.
What brands can look forward to: Business Agent
Alongside this update, Google is also introducing Business Agent, a branded AI shopping assistant that retailers can activate and customize in Google Merchant Center. Retailers like Lowe’s, Michaels, Poshmark, and Reebok are already live, but for smaller players, this beta is not widely available.
In the coming months, brands will be able to:
- Train the agent on their own product and customer data
- Provide tailored offers and recommendations
- Gain new customer insights from agent interactions
- Enable direct purchases — including agentic checkout inside the AI experience
This is effectively a new owned-commerce surface — an AI-native storefront that lives within search and conversational interfaces.
What advertisers should do now to inform AI
Advertising in an agentic world is less about driving clicks and more about influencing AI decision-making. Success now depends on whether AI systems can understand, trust, and act on your product data and offers in real time. That means optimizing your products for discovery:
- Accurate structured product feeds
- Real-time pricing and inventory
- Clearly defined promotions and eligibility rules
- AI-readable metadata to improve agent discoverability
- Exploring branded AI experiences to guide high-intent shoppers
In practice, the winners will be brands that treat AI discoverability like SEO, retail media, and paid search rolled into one — an always-on optimization discipline, not a campaign tactic.
From clicks to driving real outcomes
UCP isn’t just new tech; it changes who brands are marketing to. The goal is no longer just driving clicks; it’s getting AI to choose your product. For brands, that means clean product data, real-time pricing and inventory, and systems that let AI actually buy your products are as important as creative and media. If AI can’t understand and buy your products, you won’t show up when it matters most.
In a world where AI can shop for customers, marketing is about influencing machines as much as people. Being ready to transact is the new baseline.
Reference: Google Blog. Agentic Commerce: AI Tools and Protocols for Retailers and Platforms. 2026.
