Only 7% of the stores we tested let a page-fetching AI agent reach checkout

Page-fetching agent reached checkout

7%

8 of the 116 stores we could basket-test, with plain page requests as ChatGPT’s agent

Browsing agent reached checkout

83%

96 of 116, in a real browser; 13 more are left out as unverified

Named by AI

38%

84 of 224 brands, named at least once when a shopper asked for their kind of product

Median Agent Checkout

57

where we could see the store’s basket or checkout (208 of 225 stores)

Of the brands AI named that we could basket-test, 34 of 39 opened their checkout to an agent that browses like a person.

By sector

Method

Each store was read as a real browser and as GPTBot, ChatGPT-User, PerplexityBot and ClaudeBot, with product pages sampled from the store itself, and scored against the published Agent Conversion Score specification (v0.5). The basket test added one in-stock product to an anonymous basket as ChatGPT’s agent and opened the checkout page; nothing was ever submitted. Where the checkout turned the agent away, the same steps were repeated as an ordinary browser and, on Shopify, in a real browser: a store that let the browser through but not the agent is marked “agents refused”; a checkout that refused every automated request we could make is marked “unverified” and left out of the basket figures, because we could not tell whether an agent can pay. The recommendation check asked ChatGPT, Claude and Perplexity, with web search on, two questions a shopper would ask for the store’s kind of product, without naming any brand; answers vary between asks, so it is reported beside the score, not inside it. A page-fetching agent here is ChatGPT’s agent making plain page requests, without running a browser; a browsing agent is the same steps in a real browser, as an agent that browses like a person would. “Browser only” marks a store whose checkout opened in a real browser but not to plain requests. A store that asked us to slow down (HTTP 429) is “unverified”, not a failure. Where we could see nothing of a store’s own basket or checkout (no cart or checkout page and no basket test possible), its Agent Checkout is shown as “not measured” and left out of the checkout figures; its overall score still follows the published method. This edition does not report a “recommended but not buyable” figure: for most brands the basket could not be tested, and an Agent Checkout score below 40 is not proof that an agent cannot buy.

Not scored (15): REN Clean Skincare, Beauty Bay, Nothing, Sweetwater, Boden, Pact Coffee, Oddbox, Mindful Chef, Eve Sleep, Ollie, Frugi, JoJo Maman Bébé, Hanna Andersson, Stokke, Fabletics. Most refuse automated readers outright; we list them rather than guess.

Any brand listed can request a free re-scan or a correction at hello@conversionvector.com. Data: JSON (CC BY 4.0).