How to measure brand citation in AI systems

27 April 2026·7 min read·Wowseo.ai team

'ChatGPT started mentioning us' is not a metric. We explain how we turn AIO into numbers: share of voice, prompt map and week-by-week dynamics.

The prompt map is the measurement foundation

We take 50–150 formulations people use to ask about the category: selection, comparison, pricing, 'best' roundups. For each we record the intent and priority — high-volume prompts about choosing a vendor weigh more than informational ones.

Share of voice (AI Share of Voice)

Each prompt is run through 4 systems — ChatGPT, Perplexity, Gemini, AI Overviews — multiple times, because answers vary. We calculate the percentage of runs where the brand is mentioned and separately where it is recommended first. This gives us share of voice by prompt category and by each AI system.

What else we track

Mention sentiment: recommendation, neutral listing or cautionary.
Citation source: your page, an external review, or outdated model training data.
Competitive snapshot: who occupies 'your' prompts and through which sources.
Traffic from AI: visits from chatgpt.com, perplexity.ai and summaries in analytics.

Reporting cadence: measurement every two weeks, full prompt map rebuild — quarterly. The AIO score in our dashboard aggregates this into one number so trends are visible at a glance.

You can check your domain's current citation for free — with the AI checker on the homepage.

← All articles Our links →