--- title: "How do shopping agents choose a store? · Specoria" description: "Payment protocols, agentic storefronts and shared commerce standards went live within a year. Getting onto an agent’s shortlist now depends on verifiable data." source_url: "https://specoria.com/insights/how-shopping-agents-choose-a-store/" lang: "en" --- [← All insights](https://specoria.com/insights/) Agentic commerce October 6, 2026 6 min read # How do shopping agents choose a store? Payment protocols, agentic storefronts and shared commerce standards all went live within a year. Getting onto an agent's shortlist now depends on verifiable data, not advertising. [Specoria Team](https://specoria.com/about/) Shoppers no longer type "espresso machine" into a search box and open ten tabs. They tell their assistant what they want: "Find me an espresso machine under €500 that ships in two days and comes with 30-day returns." The assistant searches the web, reads product pages, policies and stock information, then recommends a few sellers and explains why. That becomes the only list the shopper looks at. In this article, we cover what changed over the past year, how agents decide which store to pick, and where stores can start today. ## The infrastructure came together in a year The infrastructure for agentic commerce has been built out quickly since fall 2025: - **Model Context Protocol (MCP), November 2024.** An open standard published by Anthropic that lets AI systems connect to data sources and tools through one shared interface. It has become the common language of agent endpoints that check stock, shipping or policies. - **Agentic Commerce Protocol (ACP), September 2025.** An open standard co-developed by OpenAI and Stripe that made it possible to buy inside ChatGPT without leaving the chat (Instant Checkout). The first phase launched with Etsy sellers in the US. - **Agent Payments Protocol (AP2), September 2025.** An open protocol Google developed with more than 60 payments and technology companies. It lets agents securely initiate transactions on a user's behalf across different payment methods, and it uses signed mandates to prove that the user actually approved the purchase. - **Universal Commerce Protocol (UCP), January 2026.** A standard Google and Shopify announced at NRF. It defines a common language for the entire flow, from discovery and cart to payment, order tracking and returns. - **Shopify Agentic Storefronts, March 2026.** Products from Shopify stores became purchasable inside ChatGPT, Gemini, Microsoft Copilot and Google AI Mode. The number of protocols can look confusing, but they all point the same way: agents are starting to handle the path from finding a product to filling a cart by talking directly to the seller's systems. ## The realistic picture: agents don't buy yet, but they write the list None of this means agents will start buying everything on your behalf tomorrow. According to Forrester's May 2026 assessment, most "agentic" experiences are still chat-based. Product discovery and recommendations are common, but in most cases the decision and the payment still sit with a human. Today, answer engines contribute to commerce mostly through comparison and referral, not direct purchases. In our view, the real turning point isn't payment. It's the **shortlist**. When an agent recommends five sellers, the sixth doesn't exist in that conversation. The shopper never visits your site, never looks at your product, never abandons a cart. That's why the loss doesn't show up in analytics: there's no visit and no bounce rate. > In agentic commerce, visibility isn't a ranking. It's a decision: either you're on the list, or you don't exist in that conversation. ## How does an agent choose a product? Agents don't fill gaps with guesses. They treat anything they can't verify as a risk and move on to the next seller. The reasons a product gets dropped from the comparison fall into five areas: - **Identity.** If GTIN, MPN and brand fields are missing, the agent can't match the same product across your site, your product feed and your marketplace listings. And a product it can't match, it can't compare. - **Price and stock consistency.** If your site, your feed and the marketplace each show a different price, the agent can't tell which one is right. A contradiction alone is reason enough to drop a product. - **Delivery and return terms.** An agent can't compare "fast shipping." It can compare "2 business days, anywhere in Turkey." A return policy buried in a PDF often can't be read at all. - **Evidence.** When claims such as warranty length, certifications and compatibility come with a source, the agent can cite them as justification. Claims without a source are usually ignored. - **Transaction endpoint.** If the agent can check stock or delivery time in real time (MCP) and checkout can run inside the agent (ACP, UCP), the distance between recommendation and purchase gets shorter. What stands out: none of these five has anything to do with design, advertising or content volume. All of them come down to whether your data is accurate, complete and accessible. Forrester reaches the same conclusion: brands need a content strategy aimed at machines, built on schemas and complete product information. ## What does this mean for stores in Turkey? Some of the payment features launched in the US first. But shoppers already research products with shopping agents in every market, and an agent answering a question in Turkish looks at the same criteria. A store that sells the same product on its own site, on Trendyol and on Hepsiburada (Turkey's largest marketplaces) at different prices, with different delivery times and different return terms, looks to an agent like three separate sellers that contradict each other. However long the payment infrastructure takes to reach Turkey, the race for the shortlist has already started, and it runs on data. ## A readiness checklist for the next 30 days - Complete the GTIN, MPN and brand fields for your 50 best-selling products. - State shipping time and return terms on the product page in clear, measurable language, and provide the same information in structured data (`shippingDetails` and `hasMerchantReturnPolicy` inside the schema.org `Offer`). - Publish your policy pages as HTML, not PDF. - Check price and stock consistency across your site, product feed and marketplace listings every week. - Ask ChatGPT, Gemini, Claude and Perplexity the ten questions your customers actually ask, and note who gets recommended and why. - Review the agent channel settings in your e-commerce platform, and know which data goes to which agent. ## The bottom line Before it's a promise of "autonomous shopping," agentic commerce is a question of **decidability**: can an agent reach a decision about you? Agents decide who makes the list based on information they can verify. The protocols are ready. What makes the difference is the data flowing into them. At Specoria, we measure how stores perform in this channel using real buyer tasks, trace every gap back to its source, and re-measure the impact of each fix with the same tasks. If you want to know how your store looks to a shopping agent today, request a [free readiness report](https://specoria.com/#contact). ### Sources - Anthropic, [Introducing the Model Context Protocol](https://www.anthropic.com/news/model-context-protocol), November 25, 2024. - Stripe, [Stripe powers Instant Checkout in ChatGPT and releases Agentic Commerce Protocol co-developed with OpenAI](https://stripe.com/newsroom/news/stripe-openai-instant-checkout), September 2025. - Google Cloud, [Powering AI commerce with the new Agent Payments Protocol (AP2)](https://cloud.google.com/blog/products/ai-machine-learning/announcing-agents-to-payments-ap2-protocol), September 16, 2025. - FashionNetwork, [Google rolls out its universal standard for AI-powered commerce](https://us.fashionnetwork.com/news/Google-rolls-out-its-universal-standard-for-ai-powered-commerce,1797226.html), January 13, 2026. - The Agile Brand Guide, [Shopify launches Agentic Storefronts to enable direct commerce within ChatGPT and Google Gemini](https://agilebrandguide.com/shopify-launches-agentic-storefronts-to-enable-direct-commerce-within-chatgpt-and-google-gemini/), 2026. - Forrester, [The State Of Agentic Commerce In Mid-2026](https://www.forrester.com/blogs/the-state-of-agentic-commerce-in-mid-2026/), May 28, 2026. ## Related - [Shopify opened checkout to browser agents. What WebMCP means for every other store](https://specoria.com/insights/shopify-opens-checkout-to-browser-agents-webmcp/) - [The Turkish checkout through an AI agent's eyes: seven obstacles and how to fix them](https://specoria.com/insights/turkish-checkout-through-an-ai-agents-eyes/) - [What tires a shopping agent: 33 frictions that slow AI agents down in your store](https://specoria.com/insights/what-tires-a-shopping-agent/) - [Agent protocol map: UCP, ACP, AP2, MCP, WebMCP](https://specoria.com/agent-protocols/) What each protocol does and what the free test checks - [Agentic commerce statistics](https://specoria.com/agentic-commerce-statistics/) Sourced, dated figures you can cite FREE TEST ## Is your store ready for shopping agents? See in seconds how a shopping agent reads your store: free, instant, no email needed. The free test runs in your browser and needs JavaScript. [Request the free deep audit by email instead →](https://specoria.com/#contact) Your result opens on its own page. Want the full picture? The free deep audit and panel come next, by email.