For years, the search bar was the starting point of every purchase decision. In 2026, that starting point is shifting: increasingly, people ask an AI assistant like ChatGPT, Gemini, or Perplexity for the right product – and let the agent research, compare, and in some cases order directly. This new purchase channel has a name: Agentic Commerce.
For retailers, this is more than a buzzword. When it's no longer a human reading the product page but an agent evaluating the product data set, the quality of that very data determines visibility and revenue. We break down what's behind the new protocols from OpenAI and Google, how big the channel really is – and what concrete steps small and mid-sized businesses should take now.
1. Agentic Commerce: What the Term Really Means
Agentic Commerce describes purchases in which an autonomous AI agent acts on behalf of a human: it researches products, compares options, checks availability and price, and can carry out parts of the purchase process itself, up to placing the order. The distinction matters: Conversational Commerce supports a human making a purchase in chat; Agentic Commerce lets the agent take on steps independently.
The decisive shift in perspective for marketers: the customer is no longer just the paying person, but increasingly also the agent that decides on that person's behalf. An agent doesn't read a glossy product page and isn't swayed by an attractive banner – it evaluates structured data: title, price, availability, reviews, shipping and return terms. Anyone whose data is incomplete or inconsistent here simply falls out of consideration.
In short: Agentic Commerce turns your product data set into the real storefront. The question is no longer just "How do I convince the human?" but "How do I even become findable and selectable for the AI?"
2. The Numbers: Small Share, Explosive Growth
In absolute terms, the channel is still small – but the momentum is remarkable. According to industry data, AI-driven traffic to US retailers grew by roughly 393% year-over-year in the first quarter of 2026; during the 2025 holiday shopping season, the increase briefly reached around 693%. At the same time, AI traffic still accounts for less than 0.2% of total e-commerce traffic – small, but by far the fastest-growing channel.
What's especially relevant for retailers is the quality of this traffic: users arriving via AI assistants tend to buy with more intent. Through March 2026, Adobe reported a roughly 42% higher conversion rate and about 37% higher revenue per visit compared with non-AI traffic. Analysts such as Bain and McKinsey project a market of several hundred billion US dollars in the US alone by 2030 – globally, estimates run into the trillions. These figures should be read with caution, but the direction is unmistakable.
3. ChatGPT, Instant Checkout & the Agentic Commerce Protocol
OpenAI made the first move with the Agentic Commerce Protocol (ACP), developed together with payment provider Stripe. Through Instant Checkout, US ChatGPT users could buy products directly in the chat – initially from Etsy sellers, with an opening to Shopify merchants as the next step. The promise: from product question to order, without ever leaving the chat.
In practice, the launch was rocky. Only a handful of Shopify merchants went live, and in March 2026 OpenAI largely rolled back the standalone Instant Checkout – officially because it didn't offer the flexibility desired. The new course: focus on product discovery within ChatGPT, while the actual purchase is completed on the retailer's website or app. That's an important lesson – the discovery phase is shifting into the AI, while checkout stays with the retailer for now.
Context: the retreat from a standalone checkout isn't the end of Agentic Commerce, it's a correction of pace. For retailers, that means: your products need to show up in AI discovery – and the handoff to your own checkout has to be seamless.
4. Google's Answer: Universal Commerce Protocol with Shopify
In parallel, a second, more broadly backed camp took shape. Google and Shopify unveiled the Universal Commerce Protocol (UCP) – an open standard covering the entire journey: from discovery through purchase to after-sales support such as returns. UCP was developed with partners including Etsy, Wayfair, Target, and Walmart, and is backed by more than 20 other players, among them Adyen, American Express, Best Buy, Mastercard, Stripe, Visa, and Zalando.
Technically, UCP is deliberately built to connect: it is compatible with existing standards such as the Agent Payments Protocol (AP2), Agent2Agent (A2A), and the Model Context Protocol (MCP). That addresses exactly the weak point where OpenAI's first attempt stalled: secure payments and a consistent process across many retailers.
Schematic comparison of the two approaches. As of July 2026.
5. What Changes for Buyers and the Customer Journey
For consumers, shopping becomes more convenient and faster: instead of comparing ten tabs, they state a need – "a waterproof hiking backpack under €120 with good reviews" – and the agent delivers a curated, justified selection. The classic funnel of awareness, consideration, and decision gets compressed: the AI handles research and shortlisting in seconds.
This shifts the balance of power in the journey. Anyone who previously bought attention through paid ads and clever landing pages now also competes for a spot on an agent's recommendation list – and that list is short. At the same time, trust signals such as genuine reviews, transparent prices, and clear return terms become more important, because an agent weighs exactly these criteria algorithmically.
6. What Agentic Commerce Means for SMBs and Retailers
The good news for small and mid-sized businesses: in the agent channel, it's not the biggest ad budget that counts, but the cleanest, most complete data foundation. A specialized retailer with excellently maintained product data can hold its own against large corporations in an agent recommendation – because the agent weighs relevance and data quality more heavily than brand recognition alone.
The less good news: gaps are punished mercilessly. If a GTIN is missing, if the price in the feed deviates from the shop, or if the return policy is missing, the product is simply suppressed from agent selection. For SMBs, that means the lever isn't expensive new technology, but data discipline – an effort that pays off twice over, because the same data also feeds classic Shopping ads and free listings.
For retailers in Germany: the purchase features launched US-first, but with backers like Zalando, Adyen, and Mastercard, the European rollout is getting closer. Anyone prepared now won't lose time when the channel launches – and already benefits today from better feeds.
7. Getting Product Data Ready: Feeds, GTINs & native_commerce
The concrete starting point is surprisingly down-to-earth – it's about your product feed. Shopify merchants activate UCP through the Agentic Storefronts sales channel in the Shopify admin; according to Shopify, setup takes less than 15 minutes after activation, and the Shopify catalog automatically syndicates product data along with prices and stock. Those working through Google Merchant Center add the native_commerce attribute per product, giving precise control over which items are eligible for UCP checkout.
Want to get your product data and feeds ready for the agent channel? Schedule a free initial consultation with our team.
8. Visible to Agents: GEO Instead of Pure SEO Reflexes
Beyond clean product data, there's the question of how your brand shows up in AI answers at all. This is where Generative Engine Optimization (GEO) comes in: optimizing content so that AI systems understand, trust, and cite it. Clear structure, factual precision, consistent information across all channels, and solid evidence work better than classic keyword tricks.
A current example worth a sober look is the llms.txt file, often billed as a "map for AI crawlers." Studies, however, show adoption of only around 10% – and that the major crawlers mostly ignore the file so far; Google stated in 2025 that it does not support llms.txt. The lesson: don't bet on individual hype files, but on consistent, well-structured data and content that agents can reliably parse regardless of format.
9. Conclusion: The Data Foundation Today, the Revenue Tomorrow
In 2026, Agentic Commerce is still a small but the fastest-growing channel – and it's shifting the purchase decision from the human-facing product page to the machine-readable data set. OpenAI's cautious retreat on checkout and Google's broadly backed UCP both point in the same direction: discovery is moving into AI, and anyone who wants to show up there needs clean data.
The lever for SMBs isn't expensive technology, it's data discipline: consistent prices, complete attributes, genuine reviews, clear return rules. This work pays off immediately in better Shopping ads – and simultaneously gets your range ready for the moment the agent channels open broadly in Europe too. The advantage doesn't come from watching, it comes from preparing.
Our assessment: 2026 is the year retailers have to choose – between "wait until the channel is big" and "build the data foundation now." Whoever gets their feeds in order today wins the agent's recommendation tomorrow. That's exactly where the quiet but decisive competitive advantage lies.
Also worth reading: Generative Engine Optimization (GEO): How Your Brand Becomes Visible in ChatGPT, Gemini & Perplexity and The Future of Search: SGE & AI Overviews.
10. FAQ: The Most Important Questions About Agentic Commerce
What Is Agentic Commerce?
Agentic Commerce refers to purchases in which an autonomous AI agent – for example within ChatGPT, Google, or a shopping assistant – researches, compares, and sometimes directly buys products on behalf of a human. Unlike classic Conversational Commerce, which only supports a human during a purchase, the agent can take on parts of the purchase process itself. For retailers, this makes the machine-readable product data set the decisive storefront.
What Is the Difference Between ACP and the Universal Commerce Protocol?
The Agentic Commerce Protocol (ACP) was developed by OpenAI together with Stripe and powered ChatGPT Instant Checkout. The Universal Commerce Protocol (UCP) is an open standard from Google and Shopify that covers the entire journey from discovery through purchase to after-sales, and is compatible with protocols such as AP2, A2A, and MCP. Both aim to let AI agents securely process structured product and payment data.
How Do I Make My Product Data Ready for AI Agents?
The foundation is a clean, consistent product data set. Check your feed in Google Merchant Center for disapproved products, missing GTINs, price discrepancies between feed and shop, and incomplete attributes and shipping/return information. A human might overlook a price discrepancy; an AI agent won't. Shopify merchants activate UCP through the Agentic Storefronts sales channel; in Merchant Center, the native_commerce attribute controls participation per product.
Is Agentic Commerce Already Relevant in Germany?
The purchase features launched first in the US market, and the share of AI-driven sessions in total e-commerce is still under half a percent. However, growth is the fastest of any channel, and with backers such as Zalando, Adyen, and Mastercard, the European rollout is getting closer. SMBs that clean up their product data now simultaneously improve their classic Shopping ads and are prepared for when the agent channels open in Europe.
