From measurement to action: the checklist that gets a brand cited by AI
What it is
The AI Legibility Framework measures. This framework fixes. It maps each of the five metrics to a concrete, prioritized checklist a team can work through to move from Invisible to AI-Legible. It is ordered by leverage, do the top items first; they unblock the most.
Readiness levels (per item): ❌ Not ready · ⚠️ Partial · ✅ Ready.
Tier 0, Front Door (do this first; it's binary)
If the door is shut, nothing below matters.
- [ ] Product and content pages return HTTP 200 with real content to an automated request (not a 403/429 bot-wall).
- [ ] Core content (title, description, price, body) is present in the raw HTML, visible in "View page source", not injected only after JavaScript.
- [ ] No bot-challenge / CAPTCHA interstitial served to crawlers; verified AI/search bots (GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot, Google-Extended) are allowed in robots.txt and the CDN/WAF.
- [ ] No accidental
noindexon pages you want found. - [ ] No soft-404s (a "not found" page returning a 200 status).
Tier 1, Entity Integrity (the highest-impact fix; most-failed metric)
Give AI clean facts to cite.
- [ ] Product structured data on every product page, with price, availability, currency, and a unique identifier (GTIN/MPN/SKU).
- [ ] aggregateRating / reviews in structured data where genuine reviews exist.
- [ ] Organization structured data with the real brand name and working, non-empty
sameAslinks to verified profiles. - [ ] Structured data matches the visible page (no contradictions, same price, same name).
- [ ] Brand identity (name, logo, description) is consistent everywhere a machine reads it.
Tier 2, AI Readability (make the answer extractable)
State it; don't bury it.
- [ ] Key facts (materials, dimensions, fit, shipping, use case) are stated plainly and early, not buried in prose.
- [ ] Collection and category pages have real, descriptive copy that explains what the collection is, not just a grid of products.
- [ ] FAQs answer real questions in a direct question→answer format.
- [ ] Headings describe content accurately (a machine reads the outline first).
Tier 3, Mesh Integrity (connect the map)
Turn islands into a graph.
- [ ] Products link to their collections and to genuinely related products.
- [ ] Collections link to their key products and to related collections.
- [ ] Important pages (about, materials, sustainability, guides) are reachable and linked in context, not orphaned.
- [ ] Internal links use descriptive anchors a machine can learn from.
Tier 4, Authority Signal (earn the confidence to be cited)
Be current, consistent, and trustworthy.
- [ ] Content shows recency where relevant (updated dates, current ranges/stock).
- [ ] Identity, contact, and policy details are consistent and present.
- [ ] Genuine trust markers (reviews, credentials, press) are present and machine-readable.
- [ ] No stale or contradictory information that would make a model hedge.
How to use it
- Score first. Run the AI Visibility Score so you know your band and your weakest metric.
- Fix top-down. Start at Tier 0. Don't optimize Tier 3 while the front door is shut.
- Re-measure. A score that doesn't move means the fix didn't land. Re-scan after each tier.
- Make it a metric. Track the score over time like conversion rate or Core Web Vitals. What gets measured gets managed.
The 60-second self-test (shareable)
Run these five on any site, in order. Each failure is a band you're losing:
- Front Door: Is your product text in "View page source"?
- Entity Integrity: Does Google's Rich Results Test show price + availability + rating?
- AI Readability: Can you find the key answer in one read, or is it buried?
- Mesh: Can you get from a product to its collection to a related product in two clicks?
- Authority: Is your identity consistent and current everywhere?
Five yeses ≈ AI-Legible. Most sites get one or two.
Cite as: Visibility Mesh, AI Citation Readiness Framework v1.0 (2026).