The AI Commerce Wars: How Google and OpenAI Are Building Two Competing Shopping Ecosystems

On January 11, 2026, Google CEO Sundar Pichai took the stage at the National Retail Federation conference and announced the Universal Commerce Protocol (UCP), an open-source standard that lets shoppers buy products directly inside Google AI Mode and Gemini without ever visiting a retailer's website. Four months earlier, OpenAI and Stripe had launched the Agentic Commerce Protocol (ACP), enabling the same thing inside ChatGPT.

The two largest AI platforms in the world are now building competing commerce ecosystems. Major retailers are picking sides, or hedging and integrating with both. And the implications for brand visibility are significant: if the entire shopping journey (discovery, comparison, and purchase) happens inside an AI conversation, traditional website traffic may become secondary to product feed visibility inside AI platforms.

This article breaks down what each protocol does, who has adopted what, what the early data shows, and what brands should do now.

What Happened: The Race to Own AI Shopping

The timeline moved fast. OpenAI launched ACP and Instant Checkout on September 29, 2025, co-developed with Stripe (Stripe blog). U.S. ChatGPT users could buy from Etsy sellers directly in chat, with over a million Shopify merchants (including Glossier, SKIMS, Spanx, and Vuori) announced as coming soon. It was the first time a major AI assistant offered native checkout.

Google responded at NRF in January 2026 with UCP, co-developed with Shopify, Etsy, Wayfair, Target, and Walmart. Over 20 additional companies endorsed it, including Best Buy, The Home Depot, Mastercard, Visa, American Express, Adyen, and Stripe (Google blog). UCP powers native checkout inside Google AI Mode and the Gemini app, with payments handled through Google Pay.

Then the plot twist. On March 24, 2026, OpenAI pulled back from Instant Checkout. Their statement: "We've found that the initial version of Instant Checkout did not offer the level of flexibility that we aspire to provide, so we're allowing merchants to use their own checkout experiences while we focus our efforts on product discovery." The reason was simple: it was not converting.

Walmart's Data: In-Chat Purchases Convert 3x Worse

Walmart was the highest-profile test case. In November 2025, Walmart made approximately 200,000 products available through OpenAI's Instant Checkout. The results were poor. Daniel Danker, Walmart's EVP of Product and Design, told Search Engine Land that in-chat purchases converted at one-third the rate of click-out transactions that redirected users to Walmart's website. He called the experience "unsatisfying."

The consumer data explains why. ChannelEngine's Marketplace Shopping Behavior Report 2026, surveying 4,500 shoppers across 5 countries, found that while 58% of shoppers have used AI tools to research products and 37% have started a purchase journey through an AI assistant, only 17% feel comfortable completing a purchase through AI. ChannelEngine calls this the "Confidence Economy": shoppers explore widely but buy where they feel most certain about product quality, cost, and delivery reliability.

Walmart's response was to pivot. Instead of using OpenAI's generic checkout, Walmart is now embedding its own chatbot, Sparky, directly into ChatGPT. Sparky supports account linking, loyalty program use, and Walmart's own payment system. According to Retail Dive, users who access Sparky through ChatGPT complete purchases at roughly 70% of the rate of those using Walmart.com directly, a major improvement over the 33% Instant Checkout conversion rate. Sparky launched the week of March 25, 2026, with a Gemini integration announced for April.

The Two Protocols, Explained

Google's Universal Commerce Protocol (UCP)

UCP is an open-source standard that defines a common language for the entire shopping journey: discovery, purchasing, and post-purchase support. It is designed to connect consumer surfaces (AI Mode, Gemini, third-party apps), businesses, and payment providers. UCP is compatible with Agent2Agent (A2A), the Agent Payments Protocol, and Model Context Protocol (MCP), meaning it can work with AI agents beyond just Google's own products (Google Developers Blog).

For shoppers, the experience works like this: if an AI Mode or Gemini response features an eligible product listing from a U.S. retailer, the shopper taps "Buy" and completes checkout using Google Pay, with payment and shipping info pulled from Google Wallet. The shopper never leaves the conversation. PayPal support is planned (Google Pay Help).

Google also introduced Direct Offers, which allow advertisers to present exclusive deals (like a 20% discount) directly in AI Mode to shoppers who are ready to buy. This is the first ad format designed specifically for AI-powered shopping conversations.

OpenAI's Agentic Commerce Protocol (ACP)

ACP, co-developed by OpenAI and Stripe, is also an open-source standard (currently in beta, available on GitHub). It defines how buyers, their AI agents, and businesses interact to complete purchases. Merchants share product feeds and promotions in ChatGPT through ACP. Orders, payments, and fulfillment are handled by the merchant using their existing systems. ChatGPT acts as the user's agent, passing information securely between user and merchant. Merchants pay a small fee on completed purchases; the service is free for users.

After pulling back from Instant Checkout, OpenAI refocused ChatGPT shopping on product discovery. The new experience lets users browse products visually, compare options side-by-side with pricing and reviews, refine results conversationally, and search by uploading images or describing items with criteria like budget and preferences. This is available to all Free, Go, Plus, and Pro users (OpenAI blog).

Who Is Adopting What

Shopify is the bridge between both ecosystems. Shopify co-developed UCP with Google and integrates with ACP through Shopify Catalog, requiring no additional setup from merchants. Shopify also has a Microsoft Copilot integration with embedded checkout. All connections are managed from the Shopify Admin through what Shopify calls "Agentic Storefronts" (Shopify blog). AI-driven traffic to Shopify stores is up 7x since January 2025, and AI-attributed orders are up 11x over the same period.

The Zero-Click Commerce Reality

This is where AI commerce intersects with AI visibility. We have written about how AI platforms are changing content discovery. AI commerce takes it a step further: the transaction itself can now happen without the shopper ever visiting a website.

Semrush analyzed nearly 69 million Google Search sessions between May and July 2025 and found that 92-94% of AI Mode searches end without a click to an external website (Semrush). That is more than twice the zero-click rate of AI Overviews (43%). If AI Mode adds native checkout on top of that, the incentive for shoppers to click through to a brand's website drops even further.

Gap's approach to Gemini checkout illustrates the shift. According to CNBC, Gap's product information is not crawled from their website. It is provided directly to Gemini in advance by the retailer. Gap controls accuracy, continues to collect customer data, and handles shipping and logistics. Gap CTO Sven Gerjets told CNBC that Gemini's approach gives retailers "better control" over the shopping experience than OpenAI's ChatGPT.

This is a fundamental inversion of the SEO model. In traditional search, you optimize your website so crawlers find and rank it. In AI commerce, you supply structured product data directly to the platform so it can present your products inside conversations. The website may still matter for brand building and direct traffic, but the AI platform becomes the storefront.

Product Feeds Are the New SEO

If AI shopping agents pull from structured product feeds rather than crawling websites, then product data quality becomes the primary visibility lever. This is not speculation. Alhena.ai found that stores with 99.9% attribute completion (what they call a "Golden Record") see 3-4x higher visibility in AI recommendations compared to stores with sparse data (Alhena.ai blog).

What does a complete product feed look like for AI commerce? Based on the UCP and ACP specifications, the critical fields include:

• Product title and description that answer the questions shoppers actually ask AI assistants ("best running shoes for flat feet under $150").
• Accurate, real-time pricing and availability. AI platforms will surface products that are in stock at the stated price. Stale data means lost visibility.
• High-quality images from multiple angles. ChatGPT's new shopping experience lets users search by uploading images, so visual matching matters.
• Complete product attributes (size, color, material, compatibility, dimensions). AI agents use these fields to match products to conversational queries.
• Review data and ratings. AI shopping responses frequently include star ratings and review counts as trust signals.
• Shipping and return policies. These directly affect the "Confidence Economy" that determines whether shoppers convert.

Botify and Kantar research found that 73% of consumers now use AI assistants in some form, 38% have used AI specifically for shopping tasks, and 24% of AI users are already using AI shopping assistants (PPC Land). The demand side is real. The question is whether your product data is good enough to meet it.

What Brands Should Do Now

AI commerce is still early. Conversion rates are low (Walmart's 3x penalty for in-chat checkout, only 17% consumer comfort), and OpenAI has already pivoted once. But the trajectory is clear, and the brands that build the infrastructure now will have a significant advantage as consumer trust grows. Here is what to do:

• If you are on Shopify, enable Agentic Storefronts. This is the lowest-friction path to both ecosystems. Shopify's integration with UCP and ACP is managed from the Shopify Admin and connects you to AI Mode, Gemini, ChatGPT, and Microsoft Copilot with no additional setup.
• If you are not on Shopify, evaluate direct integration with both protocols. UCP and ACP are open-source standards. Direct integration requires engineering resources but gives you control over what product data AI platforms see. Prioritize Google UCP if your audience skews toward Google Search; prioritize ACP if your category does well in conversational discovery.
• Audit your product data completeness. The "Golden Record" standard (99.9% attribute completion) is the benchmark. Incomplete product feeds will be invisible in AI shopping results. Start with your top 20% of SKUs by revenue.
• Update product data in near-real-time. AI platforms penalize stale data (wrong prices, out-of-stock items) more harshly than traditional search. If your product feed updates daily, consider moving to hourly or real-time sync.
• Maintain your website as a conversion destination. Walmart's experience shows that shoppers who click through to a website convert 3x better than those who buy in-chat. Do not abandon your website in favor of AI-native checkout. Use AI platforms for discovery and your site for conversion, at least until consumer trust catches up.
• Track AI-attributed traffic and orders. If you use Shopify, AI-attributed orders are already tracked. If not, monitor referral traffic from chat.openai.com, gemini.google.com, and copilot.microsoft.com in your analytics.

The Bigger Picture

We are watching the early stages of a platform war that could reshape e-commerce in the same way that Amazon, Google Shopping, and social commerce each created new discovery channels. Google and OpenAI are both positioning their protocols as "open standards," but the reality is that each creates a gravitational pull toward its own ecosystem.

For brands focused on AI visibility, this introduces a new dimension. Until now, the conversation has been about getting your content cited in AI responses (we cover the full playbook in our SEO-to-GEO tactics guide). AI commerce adds a second layer: getting your products surfaced, compared, and purchased inside AI conversations. The skills are different. Content visibility depends on editorial quality, sourcing, and structure. Product visibility depends on data completeness, feed accuracy, and protocol integration.

The global AI in e-commerce market was valued at $9.01 billion in 2025 and is projected to reach $74.93 billion by 2035, growing at 23.59% annually (Precedence Research). The infrastructure being built today, UCP, ACP, Shopify Agentic Storefronts, Walmart's Sparky, will determine who captures that growth.

The playbook from every previous platform shift applies here: the brands that show up early, with the best data, while the rules are still being written, tend to win disproportionately. AI commerce is not yet a major revenue channel. But the time to build the foundation is before it becomes one.