How to measure your AI search visibility (share of model)

To measure your AI search visibility, track your share of model: the percentage of AI answers that name your brand versus your competitors, broken down by engine and by query. It is the core GEO KPI, and you build it from real conversational queries run many times across every major engine.

The KPI: share of model

Forget “where do I rank.” The question now is: how often does the AI recommend me to my ideal customer, and is that trending up? Share of model answers exactly that. Measure it per engine (Claude, ChatGPT, Perplexity, Gemini, AI Overviews) and per query class (branded vs discovery), because the picture changes completely depending on which surface and which intent.

The method, step by step

1. Build a conversational query bank.Real users type full sentences, not keywords. Source the questions from sales calls, support tickets and communities like Reddit, then phrase them the way a buyer talks to an AI. There is no “Search Console for ChatGPT,” so this grounding in real language matters.
2. Run every engine. Each engine cites differently: training-recall engines name what they know, live-retrieval engines cite what they fetch. One number across one engine is misleading.
3. Run each query several times and take the median. Generative answers are non-deterministic; a single run can make you look cited or invisible by luck. A citation should count only if you appear in the majority of runs.
4. Extract more than the boolean. Capture your position, the competitors named, and the exact source domains the engine cited. Those sources are your off-site authority targets.

What good looks like

• Share of model by engine, so you know which surface is weak.
• By segment, so you can see where you win and where you are invisible.
• A competitor leaderboard, your share of voice versus rivals.
• Citation sources, the pages AI reads before it answers.
• Monthly deltas, with alerts on every win and drop.

Common mistakes

• Keyword-style queries. They misrepresent how people actually prompt AI.
• A single run. You are measuring luck, not visibility.
• Promising AI Overview visibility everywhere. For many B2B and niche queries, it does not even trigger. Measure where the answers actually appear.

See it in action

We publish real audits. See how often AI cites Lodgify across four engines (it leads at about 89%, but only 65% on ChatGPT search), and how a specialist brand, Odisea Tours, leads its category while staying invisible in some segments. That is the kind of number Tripcite gives you. If you have not yet, start with What is GEO?