How we measure AI visibility
Leadee runs each prompt against each AI engine multiple times a week through official vendor APIs, then reports mention rate, citation rate, position, share of voice and sentiment — every metric with a 95% Wilson confidence interval, so a real change is never confused with normal run-to-run variation.
Run cadence and confidence intervals
Each prompt × engine × country combination runs multiple times a week by default (more often on higher plans) rather than once. AI answer engines do not return the same answer every time you ask — the same prompt run repeatedly can surface a different list of brands on different runs.
Every metric Leadee reports carries a 95% confidence interval computed with the Wilson score method, the standard approach for proportions with a small number of trials. A change is only flagged as significant when the confidence intervals of two periods no longer overlap — that is the signal that something real moved, not just normal variation.
Metric definitions
Every metric below is defined the same way everywhere it appears in the product.
- Mention rate
- The percentage of runs in which your brand is mentioned at all, for a given prompt and engine.
- Citation rate
- The percentage of runs in which the engine cites or links a URL on your domain, not just your brand name.
- Position
- The average rank of your brand within a listed answer, when the answer format includes a list or ranking.
- Share of voice
- Your brand's mentions divided by the total mentions of every brand that appears across the same prompt set.
- Sentiment
- Whether the engine's mention of your brand reads as positive, neutral or negative.
Data sources
Leadee reads engines through their official APIs only — never by scraping a consumer chat interface.
| Source | What it is |
|---|---|
| OpenAI Responses web_search | Official OpenAI API with web search grounding |
| Gemini grounding | Official Google Gemini API grounding feature |
| Perplexity | Official Perplexity API |
| Anthropic web search | Official Anthropic API web search tool |
| DataForSEO | Official DataForSEO AI Mode / LLM data endpoints |
What Leadee does not do
- Scrape consumer AI chat interfaces (ChatGPT, Gemini or Claude's web UI) to collect answers.
- Promise a ranking, a mention or any other guaranteed outcome.
- Treat llms.txt as a primary lever — independent research on hundreds of thousands of domains found no measurable effect, so Leadee's own Audit score weights it low.
- Show a single run as if it were a stable result, without a confidence interval.
Methodology questions
- Can GEO guarantee results?
- No. AI answer engines change their output frequently, and no vendor — including Leadee — can guarantee a ranking or a mention. Leadee publishes confidence intervals on every metric so you can see how much a result can vary run to run.
- How does Leadee measure AI visibility?
- Leadee runs your prompt set against each AI engine multiple times per week through official vendor APIs (OpenAI, Gemini, Perplexity, Anthropic, DataForSEO) — never by scraping consumer chat interfaces — and reports mention rate, citation rate, average position, share of voice and sentiment with a 95% confidence interval.
- Who do I talk to?
- Email us at [email protected].
See the methodology in action
Run a free AI visibility scan and see the same signals the full Visibility Tracker measures.