Google Just Made Its Play for AI Commerce—And It Impacts More Than Retail

Universal Commerce Protocol

Google announced something at the National Retail Federation conference this weekend that deserves more attention than it's getting.

They're calling it the Universal Commerce Protocol. It's an open-source standard designed to let AI agents handle shopping from start to finish; finding products, checking out, handling post-purchase questions, all within Google's ecosystem.

Here's what that actually means and why it matters beyond retail.

What Google Built

Google developed UCP with Shopify, Etsy, Wayfair, Target, and Walmart. That's not a random group of beta testers. That's Google signaling they want this to become the default infrastructure for how AI-powered commerce works.

The protocol sits between the AI interfaces people use (Google's AI Mode in search, Gemini) and the backend systems retailers already have. Instead of every retailer building custom integrations for every AI platform, UCP creates a common language.

This week, Google also rolled out checkout directly in AI Mode.

A user can ask about a product, get recommendations, and complete the purchase without leaving the conversation. Payment runs through Google Wallet for now, with PayPal coming later.

Google's VP of Ads and Commerce put it simply: they want to standardize agentic commerce so everyone can prepare for where this is headed.

A Standards War Is Already Underway

Here's what Google isn't saying out loud: OpenAI already has its own protocol.

When OpenAI launched Instant Checkout inside ChatGPT last September, they built it on the Agentic Commerce Protocol, developed with Stripe. That's a different standard. Not compatible with UCP.

The UCP vs ACP Battle

We're watching a standards battle play out in real time. Google wants UCP to become the default. OpenAI wants ACP. Both are open-source, both have major partners, and whoever wins shapes how AI commerce infrastructure works for everyone.

If this feels familiar, it should. We saw the same thing with mobile payments, streaming formats, and social login standards. The pattern is the same; multiple players launch competing protocols, the market fragments, and eventually one wins or they're forced to interoperate.

For businesses trying to plan around this, the practical reality is frustrating. You may need to optimize for both until the dust settles, or focus on the fundamentals that work across any protocol:

• clean product data

• structured content

• real-time inventory accuracy

AI-Powered Commerce Is Already Here

I've written about this shift before:

• OpenAI launched Instant Checkout inside ChatGPT last September, letting users buy products mid-conversation

• Perplexity partnered with PayPal in November to do the same thing

• Amazon rolled out "Buy for Me" in April 2025, where their AI agent purchases products from other websites on a customer's behalf without them ever leaving Amazon.

The pattern is very clear. Every major platform is building toward a world where AI handles the shopping process. Not just recommendations but transactions.

McKinsey published research in October projecting this market could reach $3 to $5 trillion globally by 2030. That's not aspirational forecasting. That's based on how fast consumers are already shifting discovery and research into AI interfaces.

For anyone selling online, the question has changed. It's no longer whether your SEO is strong. It's whether AI agents can find you, understand what you sell, and recommend you when the question fits.

The Amazon Situation Is Worth Watching

Speaking of AI agents making purchases; Amazon is in an interesting position right now.

In November, Amazon sued Perplexity, alleging that the startup's Comet browser used AI agents to scrape Amazon's site and make purchases without permission. Amazon's statement said third-party shopping agents should operate openly and respect service provider decisions.

Meanwhile, Amazon's own Buy for Me feature does exactly what they're suing Perplexity over, just to other retailers instead of to Amazon.

Last week, small merchants started pushing back. Over 180 businesses have reported that Amazon listed their products through "Buy for Me" without permission. Some found their products for sale with wrong images or items they don't even carry. One stationery shop discovered Amazon was selling a stress ball they've never stocked. Another brand received orders for products that had been removed from their site months ago.

Amazon says merchants can opt out. But the default is opt-in, and many didn't know they were participating until orders started arriving from a "buyforme.amazon" email address.

This matters beyond the PR problem. It shows how quickly AI commerce can move from "interesting experiment" to "affecting your business without your input." The infrastructure is being built whether individual merchants are ready or not.

But Consumer Trust Isn't a Given

Before anyone rewrites their entire strategy around AI commerce, a reality check.

IAB research shows only 46% of shoppers fully trust AI recommendations. That's not a majority. The infrastructure is being built at full speed, but consumer adoption isn't guaranteed to follow at the same pace.

Industry analysts note that AI-driven purchasing will likely gain traction fastest in categories that are repeatable and planning-oriented:

• outfits

• home goods

• gifts

• seasonal items

Categories tied to identity and aspiration (luxury, specialty goods) are expected to resist full automation longer.

The opportunity is real. The timeline may be longer than the hype suggests.

Direct Offers: Advertising Built for AI Conversations

Google is also testing something called Direct Offers. It lets retailers promote specific deals, a percentage off or free shipping for example, when someone in AI Mode shows buying intent.

Traditional paid search serves ads based on keywords. Direct Offers serves promotions based on conversational signals. Someone asking Gemini "what's the best carry-on luggage under $300" who then says "I might buy one this week" could see a targeted discount from a participating retailer.

This changes how promotional strategy works. Your deals need to be structured for AI consumption; real-time pricing, clean product data, clear value propositions that translate into conversational contexts. The brands piloting this (including Shopify merchants, e.l.f. Cosmetics, Petco, and Samsonite) are getting early data on what converts in this new environment.

Business Agent: Brands Inside the Conversation

The third piece of this announcement is Business Agent. It lets shoppers talk directly with brands through Google's surfaces, but in the brand's own voice.

Right now, when someone asks an AI about your products, they get whatever the AI synthesizes from available information. Business Agent gives retailers more control over that conversation. Early partners include Lowe's, Michaels, Poshmark, and Reebok.

This matters because one of the biggest risks of AI-mediated commerce is losing the customer relationship entirely. If every interaction happens through an AI intermediary, brands become commodity suppliers.

Business Agent is Google's attempt to give retailers a foothold inside that conversation.

New Merchant Center Attributes for AI Discovery

One tactical detail from the announcement that's easy to miss: Google is adding new product attributes in Merchant Center specifically designed for AI Mode discovery.

These go beyond traditional keywords. Google is asking for things like answers to common product questions, compatible accessories, and substitutes. The goal is giving AI systems richer context to work with when recommending products.

This is rolling out to a limited group of retailers first, with broader expansion planned over the coming months. If you're running ecommerce on Google's platforms, this is worth tracking. It's a concrete signal of what Google thinks AI agents need to surface products accurately.

This Isn't Just About Retail

Google announced all of this at a retail conference, but the implications extend well beyond ecommerce.

The same infrastructure; AI agents handling research, comparison, and transaction initiation, will reshape B2B purchasing and professional services. Decision-makers already use AI tools to research vendors, compare solutions, and narrow options before contacting anyone directly.

If a procurement manager asks an AI agent about solutions in your category, your company either shows up in that answer or you've lost the opportunity before you knew it existed. That's true whether you're selling software, manufacturing equipment, marketing services, or financial advisory.

Business Agent functionality will expand beyond retail. The concept of a branded conversational presence, where prospects interact with an AI representation of your company before scheduling a discovery call, applies equally to B2B sales cycles.

The businesses treating this as a retail-only trend are going to find themselves scrambling when the same dynamics hit their industry.

The Measurement Problem No One Has Solved

Here's the really uncomfortable part:

we still don't have good tools for measuring AI visibility.

There's no Search Console for ChatGPT. No analytics dashboard showing how often Perplexity recommends your products. Google hasn't released reporting on AI Mode impressions or clicks. You can audit your own visibility manually, ask AI systems about your category and see what comes back, but that's not scalable and it's not data.

I've been developing frameworks for AI visibility audits with clients, but the industry is still figuring this out.

The platforms have the data. They're not sharing it yet.

This means any AI commerce strategy requires a tolerance for ambiguity. You can optimize based on best practices (structured data, comprehensive content, recent reviews, third-party mentions) but you can't measure the results with the same precision you're used to in traditional search.

That's not a reason to ignore this shift. It's a reason to get the fundamentals right now, before measurement catches up and the competition intensifies.

What This Means for 2026 Planning

I keep coming back to the same point: AI visibility is now a parallel track to traditional SEO.

The work that makes you visible in AI-powered commerce - structured product data, clear content that answers real questions, recent reviews, third-party mentions - is the same work that matters across Google's AI features, ChatGPT, Perplexity, and whatever Amazon builds next.

Fix it once, show up everywhere.

Google framing themselves as "matchmaker" through ads in AI Mode tells you where the platform is headed. The open-source approach with UCP isn't altruism. Google wants to own the infrastructure that AI commerce runs on.

For mid-market businesses, the practical question is straightforward: can AI agents find you, understand what you offer, and recommend you accurately? If you don't know the answer, that's the gap to close.

Sources:

• Google NRF 2026 Announcements, Sundar Pichai Keynote (January 11, 2026)

• McKinsey: "The Agentic Commerce Opportunity" (October 2025)

• TechCrunch: Google Universal Commerce Protocol Coverage (January 2026)

• OpenAI: Instant Checkout and Agentic Commerce Protocol (September 2025)

• PayPal/Perplexity: Instant Buy Partnership (November 2025)

• Amazon: Buy for Me Announcement (April 2025)

• CNBC: Amazon AI Shopping Tool Backlash (January 2026)

• Modern Retail: Buy for Me Merchant Complaints (January 2026)

• IAB/WinBuzzer: Consumer Trust in AI Recommendations Research (January 2026)

• Constellation Research: Google NRF Announcements Analysis (January 2026)