Agentic commerce protocols let AI agents browse catalogs, compare offers, and execute purchases autonomously. Google Universal Commerce Protocol (UCP) connects merchants to Google AI Mode and Gemini, while the Agentic Commerce Protocol (ACP) by OpenAI and Stripe powers purchases inside ChatGPT. To participate, websites need machine-readable product feeds, clean offer schemas, and clear return policies.
Enter your website address to check crawler access (GPTBot, ClaudeBot), server-rendered HTML, and JSON-LD schemas.
What are agentic commerce protocols?
The web is rapidly shifting from human browsing to agentic transactions. In early 2026, Google introduced the Universal Commerce Protocol (UCP) to allow agents in Google AI Mode and Gemini to purchase products directly. Simultaneously, OpenAI and Stripe partnered to develop the Agentic Commerce Protocol (ACP) for checkout within ChatGPT. With Shopify reporting that AI-attributed orders have surged 11-fold, businesses must understand the technical handshake required to allow agents to transact.
Comparing UCP and ACP technical requirements
Google UCP relies on Google Merchant Center feeds, real-time inventory endpoints, and valid Schema.org Product and Offer markup delivered in clean HTML. In contrast, OpenAI ACP leverages Stripe infrastructure and Agent Payments Protocols (AP2) to authorize transactions on behalf of verified users. While their backends differ, both protocols share an identical prerequisite: the merchant website must expose unambiguous pricing, availability, and return terms in machine-digestible formats.
- Product visibility: Deliver product specifications and pricing in server-rendered HTML rather than client-rendered JavaScript.
- Structured Offer data: Provide Product and Offer JSON-LD specifying currency, price, and in-stock status.
- Explicit policies: Format refund windows, return fees, and warranty periods as machine-readable data via MerchantReturnPolicy.
- Crawler permissions: Ensure robots.txt permits agent crawlers including ChatGPT-User and OAI-SearchBot.
Why agents skip stores during product evaluation
When an AI agent searches for products to fulfill a user request (such as finding running shoes under 80 dollars), it compares multiple merchants simultaneously. If a store hides prices behind JavaScript hydration, lacks stock availability, or fails to define return terms, the agent flags higher risk of checkout failure. To protect the consumer, the agent automatically skips that store and purchases from a competitor with structured, transparent data.
How SiteReadyFor.SI verifies protocol readiness
SiteReadyFor.SI scans your public store to verify that the observable technical foundations for UCP and ACP are in place. The scanner checks crawler permissions, audits initial server HTML for missing content, inspects Product and Offer schemas, and provides implementation-ready JSON-LD starter templates with return policy blocks that you can download and publish immediately.