How to Check Agentic Commerce Readiness in AmICited

The furthest step in agent-readiness is letting AI agents transact: browse, check out and pay on your site on a user’s behalf. That capability has a name, and understanding it is the first step to knowing whether your site is ready for it.

Quick Steps

• Open Audit → Agent Accessibility and find the Agentic Commerce section.
• It checks whether your homepage advertises a protocol (ACP, UCP) that lets AI agents transact directly.
• “Not detected” is the expected, default result for most sites today since these protocols are still emerging.
• Ecommerce and booking sites should watch this closely; other brands get more immediate value from the rest of the audit.
• Adopt a supported protocol and re-run the audit to confirm detection whenever your setup changes.

What is agentic commerce?

Agentic commerce is the umbrella term for AI agents completing a purchase on a person’s behalf (from product discovery through checkout and payment) without the human clicking through a traditional storefront. Instead of a shopper browsing your category pages, comparing prices and filling in a checkout form, an agentic AI system does that work: it interprets a request like “find me a waterproof jacket under €150 and order the one with the best reviews,” queries multiple merchants, and completes the transaction using a standardized, machine-readable contract rather than by scraping your HTML and guessing at button positions.

This is distinct from earlier phases of AI commerce. Conversational commerce is a chatbot helping someone research or narrow down options inside a chat window; the human still does the clicking. AI-mediated commerce covers a broader set of AI-assisted shopping behaviors, from recommendation to comparison. Agentic commerce is the endpoint of that progression: the agent doesn’t just recommend a product, it executes the transaction. The clearest real-world example is ChatGPT Instant Checkout , which lets a user complete a purchase inside the chat interface via a merchant-provided, machine-readable checkout flow rather than a redirect to a web page.

The mechanism that makes this possible is a protocol: a defined, structured way for an agent to discover what a merchant sells, what it costs, and how to pay, without needing to parse a rendered webpage the way a human browser does. Two names come up most often today: ACP (Agentic Commerce Protocol) and UCP (Universal Commerce Protocol). Both are early-stage specifications that let a site publish, via its homepage HTML or a manifest file, a machine-readable declaration that says “AI agents can transact here, and here is how.” A site that advertises one of these protocols is signaling that it has built the plumbing for agent-led checkout, not just agent-led reading.

Why this matters for AI search visibility: as more shopping journeys start with a prompt instead of a search box, the merchants an agent can actually transact with have a structural advantage over merchants an agent can merely describe. A brand can have excellent AI visibility (frequently cited, favorably described) and still lose the sale if the agent can’t complete a purchase without leaving the conversation. Agentic commerce readiness is the layer that turns being recommended into being bought from.

Where to find it

It’s the Agentic Commerce section of Audit → Agent Accessibility, the same audit area that covers crawler access, structured data, and WebMCP readiness. Agentic Commerce sits alongside those checks as the most forward-looking one: where the rest of the Agent Accessibility audit asks “can an AI agent read and understand this site,” this section asks “can an AI agent complete a transaction on it.”

What it checks

As the section explains: “Whether the homepage advertises a protocol that lets AI agents transact directly, browse, check out, and pay, via a machine-readable contract (ACP, UCP). Detected from homepage HTML and manifest probes.” AmICited looks for these agentic-commerce signals on your site and reports what it finds, giving you a single readiness tile alongside the rest of your AI accessibility audit results rather than requiring you to manually inspect page source or manifest files yourself.

Detection works by probing the same surfaces an AI shopping agent would probe: the homepage HTML for protocol declarations, and any manifest file a merchant might expose to advertise commerce capabilities. If neither signal is present, the tile reports “not detected,” which, given how young ACP and UCP still are, is the default state for the overwhelming majority of ecommerce sites checked today. That’s a fundamentally different kind of check from most of the rest of the audit: robots.txt configuration, structured data, and crawlability checks are graded against mature, widely adopted standards, so most sites can meaningfully pass or fail them today. Agentic commerce is graded against a standard still being written, so the honest reading of a “not detected” result is “not yet applicable” rather than “broken.”

Why it matters

• Agent-driven purchases. As people delegate shopping tasks to AI agents, sites that support agentic checkout can capture those transactions directly instead of losing the shopper to a competitor the agent can transact with. This is the same dynamic covered in how AI agents will transform online shopping : the merchants that adapt first capture disproportionate share.
• Machine-readable trust. Protocols like ACP/UCP give agents a defined, safe way to transact rather than scraping and guessing at your checkout flow. This mirrors why structured data matters for AI citation generally: machines act more confidently, and more accurately, on content and processes they can parse deterministically rather than infer.
• First-mover upside. Commerce sites that adopt this early are positioned for the shift toward agentic shopping: being one of the few detectable merchants when agent-led shopping volume starts to scale is a meaningfully different position than retrofitting after the fact.
• A new front in brand visibility. Agentic AI making autonomous purchases changes what “being recommended” means: it’s no longer enough for an agent to describe your product favorably if it then can’t complete the purchase without a human taking over.

How to use it

1. Check your status. Most sites will show nothing detected; that’s today’s baseline, not a failure state. Treat “not detected” as informational rather than alarming unless you already run agentic checkout infrastructure elsewhere.
2. Assess the fit. If you run an eshop or booking flow, agentic commerce is worth watching closely: this check is most actionable for sites where a completed transaction, not just a citation, is the end goal. Brands whose value comes from being recommended and researched (rather than transacted with directly) will get more immediate return from the rest of the Agent Accessibility audit (crawler access, structured data, robots.txt configuration) than from this section specifically.
3. Adopt and re-check. If you implement a supported protocol, re-run the audit to confirm detection and update the Agentic commerce readiness tile. Because ACP and UCP are still emerging specifications, expect the exact requirements to evolve; re-checking periodically, rather than treating a single pass as final, is the safer approach while the standards mature.
4. Track it alongside your broader visibility metrics. Agentic commerce readiness isn’t a substitute for monitoring how often your brand is actually cited and recommended in AI answers; pair this audit tile with your ongoing AI rank tracker data so you can see both halves of the funnel: whether AI engines mention you, and whether they could transact with you if they wanted to.

Together with WebMCP, this check rounds out the picture of how ready your site is not just to be read by AI agents, but to be used by them. For ecommerce and booking businesses in particular, agentic commerce readiness is the logical next step once the fundamentals are in place: solid crawler access, clean structured data, and a homepage AI systems can parse without guesswork. If you’re already tracking how often ChatGPT, Perplexity and Gemini mention your brand, this audit closes the loop by asking a harder question: could those same agents actually complete a purchase if a user asked them to? For businesses built specifically around ecommerce and product discovery, it’s worth reading how AI shopping intent and agentic checkout interact with the rest of your GEO strategy in our guide for ecommerce and product brands , so agentic commerce readiness becomes one deliberate piece of a wider AI-visibility program rather than an isolated technical checkbox.