Search / Found
Whether OpenAI searched for or found the brand or its official domain while answering buyer-intent questions.
Methodology
How the AI Brand Check measures OpenAI visibility
The AI Brand Check is designed to show how OpenAI currently sees a brand when answering buyer questions. The downloadable artifact can continue to be used as the formal OpenAI AI Visibility Report.
Start AI Brand CheckWhat we measure
Five signals from the same buyer-question run.
Citation
Whether the final answer used the brand's official domain as evidence through answer citations.
Mention
Whether the final answer explicitly named the brand.
Recommendation
Whether the final answer included the brand as a relevant choice, shortlist item, or recommendation.
Competitive position
How often the brand appears relative to submitted competitors and market brands that OpenAI surfaces.
How to read it
Found, cited, mentioned, and recommended are different.
OpenAI can find a website during research without citing it in the final answer. It can cite a source without strongly recommending the brand. It can mention a brand without making it a useful choice for the buyer.
That is why the customer result shows a simple funnel before the plain-language diagnosis and action plan.
Guardrails
Transparent boundaries.
Snapshot, not universal ranking
The report measures API-based OpenAI web-search behavior for one buyer-question run. It does not claim to measure every consumer ChatGPT session and is not a universal ChatGPT ranking.
AI results can change
AI responses are stochastic and can shift over time, so a single report is a point-in-time diagnostic rather than permanent truth.
Failures are not brand misses
Technical or provider failures are excluded from brand-performance measurement where applicable.
No recommendation guarantees
We do not guarantee that optimization work will cause OpenAI to rank, cite, or recommend a brand.
Product stance
Simple for business owners. Transparent enough for professionals.
The main flow stays simple because customers should not need to understand GEO terminology to buy a useful snapshot. The methodology is here for users who want to see that the product measures actual AI behavior rather than running a generic checklist.