AI Visibility Scoring

AI Visibility Glossary · The Lighthouse System

Foundational Term

Signal

A signal is any piece of information an AI system can read, verify, and use to decide whether your business is a credible answer to a question. Signals are not the same as content. A blog post is content. Whether that post is indexed, authored, dated, and structured in a way AI can parse. That is a signal. The AI Visibility Rubric evaluates sixteen signals across five categories. Each is scored 0–3. The total determines whether AI systems treat your business as invisible, flat, emerging, or AI-ready.

The Why: Most businesses focus on whether content exists. AI systems care about whether content is readable, attributable, consistent, and answerable. You can have a full website and score zero on a signal if the structure isn't there. The Lighthouse System diagnoses signals, not just content.

The Five Rubric Categories

Category A

Owned Content on Domain

All business types

Owned content is any text, page, post, or resource that lives on your domain and was created by you, not a directory listing, not a press mention, not a social post. It includes your blog, your service pages, your FAQ, your About page, and any educational material you've published. The AI Visibility Rubric evaluates it on four signals: whether editorial content exists, whether an FAQ or explainer is present, whether individual services or products have their own depth pages, and whether the content has been updated recently.

The Why: AI systems need something on your domain to read before they can cite you. Press and directories point toward you, but if there's nothing substantive to land on, you don't get cited. Owned content is the foundation. Without it, every other category is building on sand.

Category B

Expertise Signals
(E-E-A-T)

All business types

Expertise signals are the evidence that a real, credentialed, experienced person stands behind the content. The framework is E-E-A-T: Experience (have you actually done this?), Expertise (are you trained or qualified?), Authoritativeness (do others recognize you as a source?), and Trustworthiness (are you consistent and verifiable?). The rubric evaluates four signals: named founder credentials in indexed content, authored editorial voice, external authority references, and schema markup that makes the entity machine-readable.

The Why: AI systems do not distinguish between a business and a person the way a human would. They look for named authors, verifiable credentials, and structured data that confirms an entity is real. A beautiful website with no named human behind it scores low on expertise. A scrappy site with a bylined column, a LinkedIn profile, and a press mention scores higher.

Category C

Local Geo Signals

Local service businesses

Distribution & Discoverability

DTC variant · New

For local service businesses, Category C evaluates geographic clarity: is the city named in indexed copy, is there a neighborhood or community presence, and is the NAP (Name, Address, Phone) consistent across every platform AI can cross-reference? For direct-to-consumer brands, the category is adapted: instead of a physical address, the signals become shipping reach, operational base, and consistent platform identity across channels. Both versions ask the same underlying question: does AI know where this business operates and who it serves?

The Why: AI systems are asked "where can I find a good chiropractor in Noe Valley?" and "can I order this wine in Ohio?" in equal measure. If your domain doesn't answer those questions in plain language, you don't appear in the answer. Location and reach are not decorative details; they are retrieval triggers. The DTC variant recognizes that geography for an online brand is not an address but a service boundary, and treats it accordingly.

Category D

Citation Quality

All business types

Citation quality measures how often and how credibly other sources refer to your business. It is the only category you cannot improve by editing your own website. The rubric evaluates three signals: authority inbound references (press, institutions, recognized directories), named testimonials with specific outcomes on your domain, and whether press and media features are actively claimed and linked on your site. A citation from the Associated Press carries more weight than a citation from a local Facebook group. Both count. Neither is within your direct control to manufacture.

The Why: AI systems learn what a business is partly by reading what others say about it. A business mentioned in Quartz, a university alumni profile, and a regional magazine is treated as a real, established entity. A business that only describes itself, no matter how eloquently, is treated with less confidence. Citation quality is the closest thing the rubric has to reputation. You build it over time, through press outreach, community presence, and making sure every mention you earn is claimed and linked from your own domain.

Category E

Answer-Readiness

All business types

Answer-readiness is the degree to which your content is structured to directly respond to the questions your potential clients are actually asking. It evaluates two signals: whether the domain contains content that answers real pre-purchase or pre-visit queries, and whether headers are formatted as questions rather than marketing slogans. A page titled "Why is European wine lower in additives?" is answer-ready. A page titled "Get Offline" is not, even if the underlying content says the same thing. Research from Search Engine Land (2026) found that pages using question-format H2 and H3 headers are 40% more likely to be cited by AI engines than pages using marketing slogans or generic section labels.

The Why: AI systems are answer engines. When someone asks ChatGPT a question, the model retrieves the most direct, structured answer it can find. If your headers are slogans, your content is invisible to that retrieval process regardless of how good the writing is. Answer-readiness is the lowest-cost, highest-return category to improve; it requires no new content, no developer, and no budget. It requires rewriting headers.

The Next Step: Knowing your Answer-Readiness score tells you that content is missing. The Semantic Topical Map tells you exactly what to write: every question cluster, every topic gap, every page that doesn't exist yet but should, priority-flagged and audience-mapped. It is the architectural floor plan before the build. Available as a standalone deliverable for $360.

Supporting Terms

E-E-A-T

Experience, Expertise, Authoritativeness, Trustworthiness. Google's four-part framework for evaluating whether a source is credible enough to surface in search results. Adopted by AI systems as a proxy for citation-worthiness. All four must be present; strong performance on three does not compensate for a zero on the fourth.

NAP

Name, Address, Phone. The three identifying data points that must be identical across every directory, social platform, and website for a local business. Inconsistency in any field signals to AI systems that the entity is unreliable or possibly duplicated. For DTC brands, the equivalent is Name, Shipping Reach, Platform Presence.

GEO (Generative Engine Optimization)

The practice of structuring content so AI systems like ChatGPT, Gemini, and Perplexity cite your business when answering relevant queries. The evolution of SEO for the AI search era. Where SEO optimized for ranking, GEO optimizes for citation and retrieval.

Topical Authority

What AI systems assign when a domain consistently covers a subject in depth over time. A single article does not build it. A cluster of related, authored, interlinked content does. Topical authority is why a specialist practice gets cited more than a generalist one, even when the generalist has more content.

Entity Fragmentation

What happens when an AI system cannot confidently connect multiple names, profiles, or presences to a single real-world business. Caused by rebrands, NAP inconsistencies, mismatched social handles, and the absence of schema markup that bridges identity across platforms. Fragmentation splits authority instead of concentrating it.

Schema Markup

Machine-readable code added to a webpage that explicitly tells AI systems what a business is, who founded it, where it operates, and how it relates to other entities on the web. Not visible to human readers. Essential for AI citation. Organization schema, LocalBusiness schema, and Person schema are the most relevant types for small businesses.

Aboutness

An information science term extended here to mean the extractable identity of any information object: a website, a document, a business. Strong aboutness is specific, consistent, and structural. Weak aboutness is generic and buried. A site with strong aboutness answers "what is this business?" in the first paragraph of every page.

AI Visibility by Platform

Each AI platform reads your content differently. ChatGPT draws from training data plus optional web browsing — frequency of appearance in its training corpus matters. Perplexity searches the web for every query and always cites sources — strong SEO and structured content directly feeds visibility here. Copilot runs on Bing's index — brands with Bing presence and structured data perform better. Gemini sits inside Google's ecosystem — Google Business Profile, reviews, and structured data all contribute. Google AI Mode combines traditional search ranking with AI reasoning — strong SEO directly transfers. The implication: a brand can be visible on Perplexity and invisible on ChatGPT, or vice versa. Tracking one platform gives an incomplete picture. The AI Visibility Rubric scores the foundational signals that improve visibility across all platforms simultaneously.

Server-Rendered HTML

Content that is fully assembled on the server and delivered as complete HTML before any JavaScript runs. AI crawlers (GPTBot, ClaudeBot, and others) do not execute JavaScript — they read only what arrives in the initial HTML response. If your site relies on client-side JavaScript to load content, menus, or text, AI models never see that content. Server-rendered HTML is the only layer AI crawlers read. This is why a page can look complete in a browser but score zero on content signals in an AI audit.

JavaScript Rendering Gap

The invisible content problem created when a site builds its pages in the browser using JavaScript frameworks (React, Vue, Next.js client-side, Shopify sections). What a human sees in a browser and what an AI crawler reads can be completely different documents. A page that appears rich and content-heavy in Chrome may deliver near-empty HTML to GPTBot or ClaudeBot. Auditing your site the way AI crawlers do — by fetching raw server HTML without executing JavaScript — reveals this gap. The fix is server-side rendering, static generation, or ensuring critical content is present in the initial HTML response.

Backlinks

A backlink is any link from an external website pointing to your domain. In traditional SEO, backlinks pass ranking authority. In GEO, they function as citation signals — proof that credible, external sources recognize your business as real. The more authoritative the linking domain, the stronger the signal. Tools like Search Atlas track backlink profiles and help identify where new inbound authority can be built. Backlinks you cannot change can still be bridged to your current entity via schema markup sameAs declarations.

llms.txt

A plain-text file placed at the root of your website (yoursite.com/llms.txt) that tells AI crawlers exactly who you are, what you offer, and which pages matter most. Think of it as a business card handed directly to every AI system that visits your domain. Free to create, five minutes to deploy, zero technical requirements. An emerging standard gaining adoption across major AI platforms.

HITL (Human in the Loop)

A Human in the Loop ensures the AI process, workflow, or outputs are running smoothly. Tess McCarthy applies 15+ years of information architecture judgment page-by-page to what automated tools cannot see. Not every engagement requires it, but no tool can replace it when it does.

What are common key indicators for AI search or GEO?

By Tess McCarthy, MLIS · V.velox Ltd.

AI systems (native answer search in LLMs like ChatGPT, Perplexity, and Claude, for example) evaluate your digital presence across five categories of signals. A signal is any piece of information an AI can read, verify, and use to decide whether your business is a credible answer to a query. The stronger your signals across all five categories, the more likely AI systems are to cite you when someone asks a relevant question.

Owned Content

Do you own what you know about yourself? Does your domain have substantive, authored content that AI can read and retrieve? Blog posts, FAQs, service depth pages, and educational articles all contribute. Recency matters: content that hasn't been updated signals a dormant presence. When we think about recency, we think about currency, or how up to date is the information on your page? In information science, currency refers to the up-to-date relevance of information. In scientific research, having access to the latest findings is essential; in historical or literary research, older sources may be appropriate. For AI visibility, currency is always essential: a dormant site signals a dormant business.

Expertise Signals (E-E-A-T)

Is there a named, credentialed human behind the content? AI systems look for bylined authorship, verifiable credentials, and schema markup that confirms the entity is real. Experience, Expertise, Authoritativeness, and Trustworthiness are the four criteria; all four must be present.

Discoverability

Does AI know how to find you, specifically you? For local businesses this means city, neighborhood, and consistent NAP. For DTC brands it means shipping reach, operational base, and consistent platform identity across every channel AI can cross-reference.

Citation Quality

Do other credible sources mention you? Press features, institutional references, named testimonials, and a press page that claims your coverage all contribute. Backlinks — links from external sites pointing to your domain — function as citation signals in both traditional SEO and GEO. The more authoritative the linking domain, the stronger the signal. This is the only category you cannot improve by editing your own website.

Answer-Readiness

How ready are you to answer a question about what you know? This is what the AI Visibility Scorecard measures. Is your content structured to directly answer the questions your potential clients are actually asking? Question-format H2 headers, FAQ sections with schema markup, and content that leads with the answer before the explanation all increase your retrieval probability. Research from Search Engine Land (2026) found that pages using question-format headers are 40% more likely to be cited by AI engines.

Book an AI Visibility Audit · $180