ACP
AI search & agentsAgentic Commerce Protocol. An open specification from OpenAI and Stripe for letting a shopping agent complete a purchase through a store's declared checkout. We check whether a store declares it.
CHK-A01Where we measure it ›Short definitions we link to from audits, rules and statistics. Each one names the rule or page where it is measured.
Agentic Commerce Protocol. An open specification from OpenAI and Stripe for letting a shopping agent complete a purchase through a store's declared checkout. We check whether a store declares it.
CHK-A01Where we measure it ›Answer engine optimisation. Making a page the source an AI assistant quotes when it answers a question, by stating the answer plainly with its number, date and sample size.
StatsWhere we measure it ›How well a store works for an AI shopping agent rather than a human: can it read the product, price and stock, hold a cart and reach checkout. Reported as a separate AI Shoppers score, never blended into the main one.
The rulesWhere we measure it ›A challenge that must be solved before continuing. When it is the only way through checkout, it blocks legitimate shopping agents as well as bots.
CHK-A02Where we measure it ›A shopper adds something to the cart and leaves without buying. Unexpected costs, forced accounts and long forms are among the reasons published checkout research cites most.
MethodologyWhere we measure it ›Google's measures of how fast a page loads, how quickly it responds, and how much its layout shifts. A product page that jumps as its images arrive makes shoppers tap the wrong thing.
PDP-032Where we measure it ›Showing shipping, tax and fees early enough that the total at payment is not a surprise. Checked on the product page, in the cart and before payment details.
CHK-031Where we measure it ›The share of the rules that applied to a store which the agent could actually judge. Rules blocked by bot protection or an unreachable page are dropped, never guessed, and lower the coverage figure. Rules the store has nothing for, no variants to picture or no free-shipping offer, are not counted against it: they never applied.
MethodologyWhere we measure it ›Apple Pay, Google Pay, Shop Pay or a PayPal one-touch button. Removes address and card typing, the step where mobile checkouts lose the most shoppers.
CHK-026Where we measure it ›Generative engine optimisation. Another name for AEO: the same practice, aimed at the answers in ChatGPT, Gemini and Perplexity.
StatsWhere we measure it ›Fixed score ranges: A from 90, B from 75, C from 55, D from 35, F below 35. The same bands appear on every report, and a store's grade changes only when its audit is re-run.
MethodologyWhere we measure it ›Paying without creating an account, with the guest path as visible as sign-in. Forced account creation is one of the most cited reasons for abandoning a checkout in Baymard Institute's research.
CHK-001Where we measure it ›A plain-text file at the site root that tells language models what a site is and which pages matter. A young convention, checked alongside whether product pages are open to automated readers at all.
PDP-A03Where we measure it ›A rule outcome meaning the agent could not reach or observe what the rule tests. Excluded from the score and from every statistic.
MethodologyWhere we measure it ›The page for a single product: price, options, stock, delivery and the add-to-cart button. The first of the three pages a UXFix audit walks, and where most buying decisions are made.
How it worksWhere we measure it ›schema.org Product markup in JSON-LD: price, availability, GTIN or SKU, rating and shipping. The first thing an agent reads to quote a product.
PDP-A01Where we measure it ›Full audits whose reports are public: stores we audited as case studies, and owners who agreed to publish. Every UXFix statistic is computed over these and prints how many it covers.
StatsWhere we measure it ›An estimated range of purchases lost to a failing rule, as a share of revenue, or a money range when a store shares its traffic. Always labelled as an estimate.
MethodologyWhere we measure it ›How much a rule matters when it fails: critical, high, medium or low. A critical rule weighs four times as much as a low one in the score.
MethodologyWhere we measure it ›An approximate delivery cost shown before checkout: a flat rate, a free-shipping threshold, or a cost for the shopper's postcode. An estimate labelled as one beats "calculated at checkout".
PDP-004Where we measure it ›Cues that tell a shopper it is safe to pay: payment logos, security notes, clear returns and contact details, placed near the moment of payment where doubt is highest.
CHK-035Where we measure it ›Universal Commerce Protocol. Google's open protocol for agent checkout in its AI shopping experiences. We check for its declarations alongside ACP.
CHK-A01Where we measure it ›The Web Content Accessibility Guidelines, the international standard for pages usable by people with disabilities. Our form rules, such as labels that stay visible, follow it.
CHK-024Where we measure it ›115 rules on your product page, cart and checkout. Nothing to install.