You can assess your store for AI agent readiness three ways: run a free public scanner, run Lighthouse’s agentic audits yourself, or follow a manual checklist for the transaction itself. They answer different questions, so the useful move is knowing what each one measures before you trust its number.

Option 1: Cloudflare’s Agent Readiness score

Cloudflare launched isitagentready.com with its Agent Readiness announcement. Enter a URL and Cloudflare makes requests to the site to check which standards it supports. The score is based on four dimensions:

  • Discoverability: robots.txt, sitemap.xml, Link headers (RFC 8288)
  • Content: Markdown for Agents
  • Bot Access Control: Content Signals, AI bot rules in robots.txt, Web Bot Auth
  • Capabilities: Agent Skills, API Catalog, OAuth discovery, MCP Server Card, WebMCP

The detail that matters for a shop: Cloudflare also checks for x402, the Universal Commerce Protocol (UCP) and the Agentic Commerce Protocol (ACP), but per the same post these do not currently count towards the score. A store can score well without any way to complete a purchase. We unpack the gap in Cloudflare’s agent readiness score, read for merchants.

Option 2: Lighthouse’s Agentic Browsing category

Chrome’s Lighthouse has an Agentic Browsing category, but it deliberately does not produce a 0-100 score. It shows a fractional pass ratio and pass/fail audits. Chrome’s documentation labels the category experimental, says testing needs Chrome 150 or later, and says WebMCP audits require registering for the WebMCP origin trial.

What it looks at, per the same page:

  • WebMCP tool registration, declarative and imperative
  • A subset of accessibility checks, because agents use the accessibility tree as their primary data model
  • Cumulative Layout Shift, because shifting elements can move between the moment an agent finds them and the moment it acts
  • An llms.txt audit

It is deterministic and CI-friendly, which makes it a good fit for a developer on your team. It does not look at your price, stock or checkout. More in what the Lighthouse agentic score checks.

Option 3: An audit that covers the transaction

Both tools above start from discovery and capability signals. Neither tells you whether an agent can read a price, pick a variant and reach checkout. That is the question AgentReady scores: access, structured product data, catalog and checkout readiness, with the check list on the methodology page. Our scan stops at the checkout boundary and never places an order.

If you would rather do it by hand, the five-step checklist covers the same ground with commands you can run yourself.

Which one to use

Your question Start with
Are the emerging agent standards published on my domain? Cloudflare
Is my front end usable by a browser agent? Lighthouse
Can an agent actually buy here? A transaction-aware audit or the manual checklist

Run more than one. Scores from different tools disagree because they measure different layers, as we showed in why agent readiness scores disagree.

FAQ

Do I need an agency to assess my store for AI agent readiness?

Not for a first pass. Free tools and a manual checklist cover discovery, front-end usability and the path to checkout. An agency or developer becomes useful when you have to fix what the tools find, such as structured data or WebMCP registration.

Does a high Cloudflare Agent Readiness score mean AI agents can buy from my store?

No. Cloudflare’s own announcement says the commerce checks for x402, UCP and ACP do not currently count towards the score. The score reflects discoverability, content format, bot access control and capabilities.

Does Lighthouse give an agent readiness score out of 100?

No. Chrome’s documentation says the Agentic Browsing category has no weighted 0-100 score. It reports a pass ratio and individual pass or fail audits, and it is marked experimental.

Sources

A discovery score tells you the front door is signposted; the transaction is what an agent actually needs to get through.