Behind the answer.
Citequo measures a defined sample of AI answers. Every result keeps its question, answer, citations, and testing date.
1. Start with buying questions
Confirm your services, market, exact business name, and optional competitors. Review 3–20 questions before saving your business. Keep your brand name out of questions intended to measure discovery.
2. Run independent searches
Each question is sent independently to OpenAI’s Responses API with web search required, without your target brand added to the prompt. The current configured model is recorded in each scan. These API answers can differ from answers in the consumer ChatGPT app.
3. Read separate measurements
- Brand mention rate: completed answers containing the exact business name, divided by all valid completed answers.
- Website citation rate: completed answers with an actual citation to your website domain or its subdomains, divided by all valid completed answers.
- Coverage: valid completed answers divided by the planned questions. Rates are withheld below 80% coverage.
- Competitor mentions: tracked names and additional business names extracted from each answer, checked against the actual answer text. This is a sampled list, not a complete market census.
A mention can be positive, neutral, or negative. We do not interpret all mentions as recommendations. Exact-name matching may miss abbreviations or confuse similarly named businesses. Read the original answer and matched excerpt to verify it.
4. Treat failures honestly
A timeout, refusal, truncated response, or answer without a completed search is excluded from the visibility denominator. A saved scan uses one allowance for its full question set. Failed answers can be retried within the same scan, subject to operating limits. We never retry an unfavorable result to improve a score.
5. Compare like with like
History compares scans with the same question set, model, business name, domain, and tracked competitors. Each scan is a snapshot. Answer variation, a small sample, and provider changes can affect results. A change after an edit does not prove the edit caused it.
What the free website check covers
We fetch the submitted page’s HTML, following up to three validated redirects. We check page-level indexing directives, title, description, main heading, text availability, and parseable JSON-LD. We do not execute JavaScript, crawl the whole site, verify indexing, measure Google rankings, or confirm that robots.txt permits every crawler. Structured data is optional and does not guarantee AI visibility.
Suggested actions
Actions connect observed page issues and missing mentions to possible improvements. They are recommendations for review, not guaranteed ranking fixes. Google’s guidance explains that normal SEO fundamentals also apply to its AI search features.