AI Visibility Score

The AI Visibility Score is a single 0–100 number on your dashboard that summarizes how well AI engines are citing your brand. It’s the headline metric that everything else on the dashboard ladders up to.

Formula

score = mention_rate × 60 + infra_health × 40

• mention_rate (60% of the score) — share of LLM responses that cited your brand across all 5 LLMs (ChatGPT, Claude, Gemini, Perplexity, Google AI Overviews) in the latest weekly run. If your brand appeared in 3 of 50 responses, that’s 6%.
• infra_health — infrastructure score (40% of the score) — composite of three sub-scores, in order of actual contribution to citation impact:
  1. Schema coverage — what % of your eligible pages have JSON-LD entity-graph deployed. The biggest lever; AI crawlers parse this directly from your HTML.
  2. Content depth — what % of your pages have enough content depth + answer-first structure to be quotable by AI engines (depth score of 40 or higher).
  3. llms.txt deployment status — is it live, when was it last updated. Included for completeness; it’s a small contributor given low current crawler adoption. See llms.txt for where it fits in the priority order.

The 60/40 split reflects what actually drives citations. Mention rate is the outcome we care about, but infrastructure is the leading indicator — fix the infrastructure (especially schema + content depth), mentions follow.

Bands

Score	Band	What it means
80–100	Dominant	Rare. Usually the category-defining brand. AI engines reach for you by default.
60–80	Strong	Your brand is one of the LLMs’ default answers in your category.
40–60	Building	Consistent citations, growing primary-recommendation share.
20–40	Limited	Occasional citations, mostly from informational queries. Where most established brands sit when they start GEO work.
0–20	Critical	LLMs aren’t citing your brand at all. Common for new customers in their first month.

Why most new customers start in Critical

Three reasons most brands score below 20 on day one:

1. No schema infrastructure — most sites have either no JSON-LD or fragmented theme-default schema that doesn’t form a coherent brand entity graph. AI engines can’t confidently cite a brand they can’t structurally identify.
2. No llms.txt file — without this, LLMs scrape your site without prioritization signals.
3. Content depth gaps — pages that are 50–200 words don’t get cited even when discovered. LLMs prefer depth.

These are exactly what MeetGEO fixes, in that order. The first weekly scan diagnoses; subsequent weeks measure the climb.

Why your score might drop week-to-week

Three common causes:

1. Query mix changed. The system regenerates 10 queries each week. If the new mix probes territory you don’t yet rank for, mention_rate drops even though nothing on your site changed.
2. A competitor expanded. They published new content the LLMs picked up, displacing you in answers.
3. An LLM updated its training data. Periodic refreshes from any of the 5 LLMs can change citation patterns wholesale.

Open the Weekly Insights card for the diagnosis specific to your run — it names the queries you lost ground on.

Track the trend, not the snapshot

Week-over-week movement matters more than the absolute number. Going from 18 to 24 over four weeks is a healthier signal than holding flat at 60. Most customers see meaningful movement by week 4 once schema and llms.txt have been deployed and AI engines have re-crawled.