Our methodology

Engines we query

We query OpenAI (ChatGPT), Anthropic (Claude), Perplexity, and Google Gemini through their official public APIs with web search or grounding enabled where each engine supports it, a neutral configuration, and no custom system prompt beyond a locale hint. Google AI Overviews data is sourced via a licensed SERP data provider and is clearly labeled "via licensed SERP data" wherever shown.

Sampling — why we never show a single "rank"

AI answers are nondeterministic: the same question asked twice can name different brands. So we ask each tracked question multiple times per run (3 samples per run on every plan), spaced out in time, and report rates — "mentioned in 2 of 3 samples" — never a fake-precise position. Every metric in the product displays the sample count it is based on.

Visibility Score

score = 100 × mention_rate × (0.6 + 0.4 × position_factor)
where position_factor = 1 / average list position, capped at 1. When a brand is mentioned outside a recommendation-list context, only the base weight applies. The composite score is the mean across your plan-enabled engines.

What we can't promise

Nobody can guarantee placement inside ChatGPT or any other AI engine — anyone who says otherwise is selling snake oil. AnswerStanding measures where you stand, alerts you when it changes, and recommends evidence-grounded actions. The engines decide the rest.