[ method ]Our method
Seven layers from unclear source to usable website.
Modern search does not reward a single trick. It rewards a website that can be found, understood, trusted, and acted on. The AI Visibility Stack is the framework we use to audit that foundation and decide what to fix first.
// AI Visibility Stack
foundation first - lower layers support higher layers
[ 01 ]The principle
A system cannot use facts it cannot identify, parse, or verify.
Find
The site has to be crawlable, indexable, internally linked, and clear enough to discover.
Understand
The business, services, market, proof, and answers have to be explicit instead of implied.
Act
A visitor who arrives from modern search needs a clear, reliable path to contact, book, or buy.
[ 02 ]The seven layers
The stack runs from identity to action.
Lower layers come first. A gap in identity, service clarity, or crawlability weakens everything above it.
Layer
Entity Clarity
A machine can tell exactly who the business is.
What it means
Business name, category, location, service area, contact details, profiles, and ownership are consistent enough to identify one clear entity.
Weak signal
The site uses vague category language, inconsistent names, no useful About page, and business facts scattered across templates.
Typical fix
Clarify name, category, market, contact, About content, Organization or LocalBusiness schema, and key sameAs profiles.
Layer
Service Specificity
Each offer maps to a real buyer question.
What it means
Services are described on dedicated, crawlable pages with plain names, use cases, geography, eligibility, process, and next steps.
Weak signal
One thin services list tries to cover everything, with little detail and no clear page for high-intent searches.
Typical fix
Build or rewrite service pages around what the buyer asks, what the service includes, who it fits, and why the provider is credible.
Layer
Proof Density
The site supports important claims with inspectable evidence.
What it means
Credentials, reviews, examples, team details, policies, real photos, and experience signals are visible and specific.
Weak signal
The site makes broad claims but hides the evidence, leaving people and systems little public support for those claims.
Typical fix
Surface honest proof near claims: credentials, reviews, case details, before/after context where appropriate, and real team information.
Layer
Local Relevance
The business is tied to the places it actually serves.
What it means
Cities, neighborhoods, service areas, local proof, Google Business Profile alignment, and location details are explicit and consistent.
Weak signal
The site says 'near me' or targets a city without showing real location context, service-area detail, or profile consistency.
Typical fix
Clarify the actual service area, align NAP details, add local context to service pages, and only create location pages that are genuine.
Layer
Answer Coverage
The site answers pre-sale questions directly.
What it means
Pages cover pricing guidance, fit, process, timing, risks, comparisons, FAQs, and what happens next in language humans can use.
Weak signal
The site sells, but does not answer. AI systems have to infer important buyer details from competitors or generic sources.
Typical fix
Add answer-first sections, FAQs, comparison content, process explanations, and objections that match real buyer questions.
Layer
Machine Readability
The site can be crawled, parsed, and connected.
What it means
Semantic HTML, unique metadata, internal links, sitemap, robots, canonical tags, JSON-LD schema, and optional llms.txt all agree.
Weak signal
Important content is hard to crawl, schema is generic or invalid, pages are orphaned, and machine-readable signals contradict each other.
Typical fix
Clean the technical foundation, validate schema, connect pages with internal links, and keep metadata and support files synchronized.
Layer
Conversion Path
A referred visitor knows what to do next.
What it means
Calls to action, forms, booking paths, phone/email options, accessibility, and trust cues make the next step obvious.
Weak signal
Even if modern search sends a visitor, the page buries contact options, uses fragile forms, or fails to match the visitor's intent.
Typical fix
Make the primary action visible, reduce form friction, improve booking/contact reliability, and align CTAs with page intent.
[ 03 ]Scoring
We score for usefulness, not checklist theater.
0-2
Missing
The layer is absent, contradictory, blocked, or too thin to help people or machines.
3-5
Present but weak
Some signals exist, but they are generic, incomplete, inconsistent, or not tied to a useful page.
6-8
Clear enough
The layer works for core pages, but there are gaps in depth, proof, structure, or maintenance.
9-10
Strong
The layer is specific, visible, structured, consistent, and connected to the rest of the site.
Paid audit formula
Each of the seven layers is scored from 0–10 against documented evidence. Overall readiness = layer points earned ÷ 70 × 100, rounded to the nearest whole number. For example, 49/70 becomes 70/100. The layers are equally weighted in a paid report.
Free self-scorecard
The self-scorecard uses question-level weights that total 100 and awards half credit for “Not sure.” It is a planning aid, not a verified audit or a prediction of rankings, traffic, or citations. Rubric version: 2026-08-25.
[ 04 ]How we use it
The stack turns a fuzzy AI-search problem into ordered work.
Audit
We score each layer against observable evidence, then record any answer samples with the engine, prompt, method, and date used.
Prioritize
Lower-layer gaps come first. A conversion tweak cannot fix a site that cannot identify the business or explain its services.
Implement
The roadmap becomes concrete work: service pages, metadata, schema, FAQs, local signals, proof, internal links, and CTAs.
Maintain
Modern-search readiness drifts as services, pages, locations, competitors, and answer engines change. The stack gives us a repeatable re-audit model.
[ 05 ]Failure patterns
Where sites usually lose points.
- A polished homepage that never states the category, service area, or primary services plainly
- Service pages that exist only as a bulleted list with no answers, proof, or next step
- Schema that validates syntactically but describes a different business than the page does
- FAQs that answer easy internal questions instead of buyer questions
- Local pages targeting cities without genuine local relevance
- A contact path that breaks the moment a mobile visitor wants to book
[ 06 ]Related pages
Go deeper on the layers that usually need work.
[ 07 ]Good to know
Common questions.
See your stack scored.
Start with the free audit, or get the full $497 AI Search Audit for seven-layer scoring, labeled answer samples, a prioritized roadmap, and a live walkthrough.