AI visibility

Measuring AI visibility: components

• Engine coverage — are all the important engines included? (ChatGPT, Gemini, Perplexity, Claude and AIO at a minimum.)
• Query set — a representative set of prioritised phrases, ideally weighted by real search volume.
• Citation and mention separately — combine them with weighting or report them separately.
• Majority vote — mandatory, otherwise the data is noise.
• Consistent language and region configuration — DE queries from a DE IP, EN queries from a US IP.
• Freshness of measurement — weekly or fortnightly, not ad hoc.

AI visibility as a reporting KPI

There are three sensible reporting approaches: (1) Overall AI visibility as a 0–100 index, primarily for observing trends over time. (2) Citation rate per engine individually, for optimisation diagnostics (”where am I losing, where am I gaining?”). (3) AI visibility per topic cluster — if the domain covers several topics, report separately per cluster in order to recognise where the optimisation levers take effect.

AI visibility and classic SEO in reporting

Ideally, modern SEO reporting presents both metrics side by side: the classic visibility index (Sistrix/Searchmetrics) for Google SERP performance, and AI visibility for LLM answer performance. The two correlate in part (good rankings favour AI citation), but they also diverge: a brand can be strong in SEO and weak in AI (or the other way round) if only one of the two disciplines is actively optimised.