Glossary

Every term we use, defined clearly.

Mirr reports contain a handful of specialised terms: Coverage, Share of Voice, Identity Gap, Web Presence Score. This glossary defines each one in plain language so anyone on your team can read a report without guessing. Bookmark this page and share it with anyone who touches the output of a Mirr audit.

AI Brand Visibility

How often and how accurately a brand is surfaced by large language models (ChatGPT, Perplexity, Claude, Gemini) when users ask questions in its category. Measured on a 0-100 scale across five perception dimensions.

Generative Engine Optimisation (GEO)

The practice of making your brand findable and accurately described inside AI answer engines. Sits alongside SEO: SEO earns clicks from Google, GEO earns mentions inside ChatGPT, Perplexity, Claude and Gemini answers.

Coverage

The percentage of research prompts in which your brand is mentioned by name at least once. Computed deterministically with regex word boundaries across all LLM responses. A Coverage of 40% means your brand appears in 6 of 15 category prompts.

Share of Voice (SoV)

Of all brand mentions across research responses (your brand plus competitors), the percentage that are yours. SoV = your mentions ÷ (your mentions + competitor mentions) × 100. High SoV means you dominate the AI conversation in your category.

Web Presence Score

A 0-100 composite of nine deterministic signals AI crawlers rely on: Wikipedia presence (25 pts), Schema.org structured data (15), verified LinkedIn company page (15), meta title (10), meta description (10), Open Graph tags (8), review platform listings (up to 8), Twitter card (5), canonical URL (4).

Identity Gap

The difference between what a brand intends to stand for and how AI actually describes it. Scored 0-10 on six dimensions: Tone of Voice, Core Values, Audience Recognition, Market Position, Brand Promise and Emotional Association.

Per-LLM Coverage

Coverage percentage tracked per AI platform (ChatGPT, Perplexity, Claude, Gemini) rather than aggregated. Reveals which platforms know your brand best and which need focused work. A wide spread means your visibility is fragile and dependent on a single training corpus.

Visibility Score

The headline 0-100 score. Average of five perception categories scored by Claude Haiku 4.5 (rubric-constrained) and then synthesised by Claude Opus 4.7: Awareness, Consideration, Decision, Sentiment and Cultural relevance. Split into four bands: Invisible (0-20), Weak-Emerging (20-50), Established (50-75), Dominant (75-100).

Cultural Signals

Qualitative associations AI attaches to a brand. Positive signals help (relevance, differentiation). Negative signals hurt (controversy, outdated positioning). Missing signals are opportunities where the brand wants to be seen but AI does not yet connect the dots.

Prompt Matrix

The full test grid: 15 realistic customer discovery queries × 4 LLMs = 60 API calls per audit. Each cell records whether the brand appeared, its position (first, early, late, not mentioned) and which competitors appeared alongside.

Action Plan

Mirr's signature output: exactly seven ranked moves, each with a rationale tied to a specific audit finding, concrete step-by-step execution instructions (3-5 steps), required tools, a measurable KPI and a 12-week timeline.

llms.txt

An emerging standard (2024) for websites to give AI crawlers a structured markdown index of canonical pages, a short brand description and key product facts. Read by GPTBot, ClaudeBot and PerplexityBot; analogous to robots.txt but for LLM training and retrieval.

Last updated

22 April 2026

Run your first audit

All of these metrics are calculated automatically on your brand in under 5 minutes. €99 for a full report.

Start your AI visibility audit →