Guide
What Is an AI Visibility Score? Definition, Formula, and Measurement
AI Visibility measures observed brand performance in AI answers. AI Readiness diagnoses one page's structure. This guide shows where each measure starts, what it can prove, and how to use both.
An AI Visibility score measures how a brand performs across a repeatable panel of AI answers. It combines observed mentions, review and comparison presence, category inclusion, sentiment, and readiness evidence. It is a brand-and-workspace outcome measure.
AI Readiness is different. It is a page-scoped Technical Audit measure based on eight structural dimensions. It tells you whether a page is accessible and extractable. Because crawler access alone does not prove citation, the audit does not tell you that ChatGPT, Claude, Perplexity, Gemini, or Google AI Overview actually mentioned or cited the brand.
Foglift keeps the two concepts separate: run a Technical Audit to diagnose a page, then use recurring AI Visibility monitoring to measure what engines actually say. The product's six-tier scoring methodology publishes the exact weights, eligibility rules, and 14-day answer window.
AI Visibility and AI Readiness answer different questions
| Measure | Unit | Evidence | What it can establish |
|---|---|---|---|
| AI Readiness | One audited page | Eight structural dimensions from the Technical Audit | Whether the page exposes clear, accessible, extractable signals |
| AI Visibility | One brand in one workspace | Readiness inputs plus successful observations from tracked prompts | How the brand performs across the monitored answer set |
The distinction changes the action you take. A weak AI Readiness dimension calls for a page-level fix. A weak Brand Mentions or Comparison tier calls for prompt, content, entity, or corroboration work. Treating the audit grade as the outcome hides that diagnosis.
The eight AI Readiness dimensions
A Foglift Technical Audit evaluates these page-scoped dimensions. The report gives each dimension its own findings so the score remains diagnostic.
- Structured Data Richness
- Heading Clarity
- FAQ Quality
- Entity Identity
- Content Depth
- Citation Formatting
- Topical Authority
- AI Crawler Access
How the six-tier AI Visibility formula works
Foglift calculates one 0 to 100 composite from six 0 to 100 tiers. The fixed weights total 100%. Technical and Authority Readiness provide the readiness layer. The other four tiers measure successful AI-answer observations from a closed 14-day window.
round(Technical × 10% + Authority × 12% + Brand Mentions × 15% + Review Presence × 18% + Comparison × 20% + Share of Voice × 25%)
| Tier | Weight | Input | Question answered |
|---|---|---|---|
| Technical Readiness | 10% | Latest workspace-owned root-domain Technical Audit | Can systems access and interpret the site? |
| Authority Readiness | 12% | Off-site mentions, domain authority, and sitemap freshness | Does the wider web corroborate the entity? |
| Brand Mentions | 15% | Successful answers to active brand and product prompts | Does the answer name the brand or product? |
| Review Presence | 18% | Successful answers to active review prompts | How is the brand framed in review-shaped answers? |
| Comparison | 20% | Successful answers to active comparison prompts | Does the brand enter buyer comparisons? |
| Share of Voice | 25% | Successful answers to active industry and category prompts | Does the brand appear in category-level answers? |
Provider errors do not enter the monitored denominator. Industry prompts that name the tracked brand are excluded from Share of Voice, and system-generated review or comparison prompts that name the brand are excluded from those tiers. Those rules prevent a failed request or a self-naming prompt from masquerading as visibility.
How do you measure AI Visibility consistently?
- Define the brand and market. Save the tracked brand, products, competitors, and domain before collecting answers.
- Build prompts by buyer intent. Separate brand and product, review, comparison, and category prompts. Each group feeds a different tier.
- Choose the engine set. Record which engines ran. Free monitoring covers Perplexity while active; Launch and higher plans can cover ChatGPT, Claude, Perplexity, Gemini, and Google AI Overview.
- Exclude failed observations. An engine error is missing evidence, not a negative answer. Preserve the error for operations, but keep it out of the score denominator.
- Separate mentions from citations. A brand can be named without a source URL, and a domain can be cited without a brand mention. Store both facts when the provider exposes citations.
- Close the time window. Compare complete periods rather than a live partial total. Foglift uses 14 closed days for its four monitoring tiers.
- Keep the basis stable. When you change prompts, categories, engines, or competitors, annotate the break before interpreting the trend.
The KDD 2024 GEO paper formalized generative-engine visibility as a measurable black-box outcome. In practice, the useful unit is still your declared panel: prompts, engines, categories, and time window. A score without that denominator is difficult to reproduce.
What is a good AI Visibility score?
There is no universal grade band that can prove a brand is visible. A single number can come from very different tier combinations, and every monitoring panel reflects its own prompt mix and market. Use your first complete, well-configured period as a baseline.
| Pattern | Likely reading | Next check |
|---|---|---|
| Composite rises, one tier falls | The headline gain may conceal a specific buyer-intent weakness. | Open the tier drill-down and inspect its prompts and engines. |
| Readiness rises, monitored tiers stay flat | The page became easier to access or interpret, but answer selection has not moved yet. | Check exact-intent coverage and independent corroboration before another technical edit. |
| Mention rate rises, citations stay flat | The entity is entering answers without the domain becoming a supporting source. | Compare cited sources with the relevant first-party page and source layer. |
How do you improve the right part of the score?
- Weak Technical Readiness: fix the exact failed audit dimensions on the affected page, then rerun that page.
- Weak Authority Readiness: improve entity consistency, current source evidence, sitemap freshness, and independent mentions.
- Weak Brand Mentions: answer the tracked brand and product questions with specific capability, price, and plan facts.
- Weak Review Presence: make the review proposition checkable and earn honest third-party reviews where buyers already look.
- Weak Comparison: publish an accurate buyer-facing comparison and close factual gaps against the pages engines select.
- Weak Share of Voice: study the category prompts and cited sources, then decide whether the gap is owned content or off-page corroboration.
Foglift joins these layers in one workflow: unlimited single-page Technical Audits, active-use weekly Perplexity monitoring on Free, and all five engines plus API, CLI, and MCP access from the $49 monthly Launch plan. That lets a team diagnose the page, observe the answer, and carry the evidence into its implementation tools.
Measure observed AI Visibility
Track recurring prompts and keep mentions, citations, sentiment, competitors, and tier evidence separate from the page audit.
Start AI Visibility monitoringFrequently asked questions
What is an AI Visibility score?
An AI Visibility score is a repeatable measure of how a brand performs in monitored AI answers. Foglift combines Technical Readiness, Authority Readiness, Brand Mentions, Review Presence, Comparison, and Share of Voice into one 0 to 100 composite. It is an observed brand-and-workspace outcome, not a page audit grade.
Is AI Readiness the same as AI Visibility?
No. AI Readiness is a page-scoped Technical Audit measure based on eight structural dimensions. AI Visibility is a brand-and-workspace outcome built from readiness inputs plus monitored answer evidence. A page can be technically ready without being mentioned, cited, reviewed, compared, or included in category answers.
How does Foglift calculate AI Visibility?
Foglift multiplies six tier scores by fixed weights, then rounds the sum: Technical Readiness 10%, Authority Readiness 12%, Brand Mentions 15%, Review Presence 18%, Comparison 20%, and Share of Voice 25%. The four answer-based tiers use successful observations from a closed 14-day window.
What is a good AI Visibility score?
A useful interpretation comes from your own stable baseline, not a universal grade band. Hold the prompt set, prompt categories, engines, competitor set, and time window steady. Then compare the composite and each tier over time, because the same headline score can hide very different strengths and gaps.
How do I check AI Readiness?
Run a Foglift Technical Audit for the exact public page you want to inspect. The report evaluates Structured Data Richness, Heading Clarity, FAQ Quality, Entity Identity, Content Depth, Citation Formatting, Topical Authority, and AI Crawler Access. The result diagnoses page structure; it does not claim that an AI engine mentioned or cited the page.
How often should I measure AI Visibility?
Use a recurring cadence that matches how quickly you can act, while keeping the evidence basis comparable. Foglift Free includes active-use weekly Perplexity monitoring. Launch supports all five engines up to daily, Growth up to twice daily, and Enterprise up to hourly.
Primary sources and methodology
- OpenAI: Publishers and Developers FAQ. Defines OAI-SearchBot discovery access separately from GPTBot potential-training controls and explains referral tracking from ChatGPT.
- Google Search Central: Generative AI features. Explains the Search eligibility boundary, query fan-out, supporting links, and provider-native measurement in Search Console.
- Aggarwal et al.: GEO: Generative Engine Optimization. Introduces a black-box framework for defining and evaluating visibility metrics in generative-engine responses.
- Foglift AI Visibility methodology. Exact fixed weights, prompt eligibility, sentiment formulas, closed-window rules, snapshot provenance, and plan coverage.
- Foglift AI Readiness research. Reproducible Technical Audit analysis across 311 domains with full AI Readiness scoring.
Fundamentals: Learn about GEO (Generative Engine Optimization) and AEO (Answer Engine Optimization) (the two frameworks for optimizing your content for AI search engines).