The AI Legibility Framework v1.0 defines the five metrics behind the Visibility Mesh AI Visibility Score: Front Door, Entity Integrity, AI Readability, Mesh Integrity and Authority Signal. Together they measure how completely an AI crawler can read a website and understand the brand and products on it, and they roll up into one score from 0 to 100.
This specification explains what the score measures. It describes each metric and why an answer engine depends on it. It does not publish the scoring weights, the breakpoints or the sub criteria. Those stay private, so a site improves its score by becoming easier to read rather than by learning a formula.
What is the AI Visibility Score?
The AI Visibility Score is one number from 0 to 100 for a website, built from the five metrics below. Two scans of an unchanged site usually land within about 2 points of each other. The AI Visibility Maturity Model places each score in a band: Starting Point (0 to 39), Developing (40 to 59), Established (60 to 79) and Leading (80 to 100).
What are the five metrics?
Front Door: can a machine read the page at all?
Front Door tests whether AI and search crawlers can fetch the content. A bot wall, a page that renders only with JavaScript, or an error page sent with a success code all fail it. Every other metric depends on this one. If the Front Door fails, crawlers cannot read the content at all.
Entity Integrity: is each product and the brand a clean fact?
Entity Integrity tests the structured data for the brand and its products. It looks at Product data such as price, availability, ratings and identifiers, at the Organization record and its verified profile links, and at whether all of it matches the visible page. A machine reads a fact stated as structured data more reliably than the same fact in a paragraph. In our June 2026 study of 216 Shopify storefronts, 72% failed this metric, more than any other.
AI Readability: is the answer stated or buried?
AI Readability tests whether key facts appear in a form an engine can lift: plain definitions, direct statements and specs written out. Answer engines lift stated facts, and they rarely dig through marketing prose to find one.
Mesh Integrity: do the pages form one map?
Mesh Integrity tests whether collections, products and key pages link into a graph that shows how they relate. Those links show a model which products belong together and what each one is for. Our guide to the Mesh explains the idea.
Authority Signal: is the identity current and consistent?
Authority Signal tests freshness, consistent identity across the site and its structured data, and the trust markers a model uses to decide how confidently to cite a source.
How do the five metrics combine?
Front Door works as a gate. If it fails, the other four have little to measure. Above it, the score rises only when the entity layer is clean and the content is extractable, connected and current. Most sites have at least one weak metric, and a weak metric pulls the whole score down. The median in our August 2026 study of 245 Shopify storefronts was 46 out of 100, as the Shopify legibility study reports.
What does a typical pattern look like?
Take a pet supplies store on Shopify with 150 products. The theme is fast and the crawlers get in, so Front Door is strong. The Product markup has no barcodes, and an app prints a second Organization block, so Entity Integrity is weak. Collection pages are grids with no text, which drags AI Readability down. The score lands in the high 40s, and the scan points to Entity Integrity as the first fix.
That pattern is common enough that we could almost print it on a mug. It is also good news, since the weakest metric is the one an owner can fix from the Shopify admin.
How can I check each metric myself?
- For Front Door, open a product page, right click and choose View page source. Search for the product name and the price. If they are missing, the content depends on scripts.
- For Entity Integrity, paste a product URL into Google's Rich Results Test and check that it finds one Product item with price and availability.
- For AI Readability, read the first two sentences of a product page and a collection page. Each should say what the item is and who it is for.
- For Mesh Integrity, start on a product and try to reach its collection and a related product in two clicks.
- For Authority Signal, compare your business name, address and contact details on the footer, the About page and the Contact page.
Our AI Citation Readiness checklist turns these metrics into a full task list in priority order, and the guide to what your AI visibility score means explains how to read a result.
How is the framework versioned?
This is version 1.0. A score compares directly with another score only under the same framework version, so published research stays reproducible. Changes ship as version 1.1 or 2.0 with a changelog on the methodology page.
Questions people ask about the AI Legibility Framework
What does the AI Visibility Score measure?
The AI Visibility Score measures how completely an AI crawler can read a website and understand its brand and products. It combines five metrics, Front Door, Entity Integrity, AI Readability, Mesh Integrity and Authority Signal, into one number from 0 to 100.
Does a high score guarantee that AI will recommend my business?
No score can promise that, since each assistant decides its own answers. The score measures what a crawler can read and understand on your site, which is the part you can change.
Why are the scoring weights not published?
The framework publishes what each metric tests so the score is transparent and citable. The weights and breakpoints stay private so the score measures real legibility rather than how well a site has been tuned to a formula.
Cite as: Visibility Mesh, AI Legibility Framework v1.0 (2026).
A free scan of up to 5 key pages returns your score and the result for each of the five metrics.