A measurement method you can inspect.
AI answers change by model, time, market, and phrasing. We preserve those variables instead of turning a volatile observation into a false promise.
1. Public-site readiness
The free scanner checks public, observable foundations that help machines understand and cite a website: secure delivery, page titles and descriptions, primary headings, canonical URLs, declared language, structured data, semantic page structure, robots policy, international alternates, and useful content depth.
This score is a technical and content readiness indicator. It is not a claim that a brand currently appears in any specific AI product.
2. Answer-engine observations
Paid monitoring uses documented APIs and supported search-grounding tools where available. Each observation stores the question, language, market, engine, model, answer, citations, execution time, and relevant configuration. Consumer interfaces are not treated as stable APIs.
3. Metrics
- Mention rate: the share of sampled answers that meaningfully include the brand.
- Citation share: the share of relevant source citations that point to the brand’s controlled web properties.
- Competitive share of voice: brand mentions compared with the named competitor set.
- Factual accuracy: whether material claims about the brand match approved facts.
- Volatility: how much outcomes vary across repeated samples and time periods.
4. What we do not promise
No provider controls the output of ChatGPT, Gemini, Perplexity, or other independent answer engines. FMPPT GEO does not guarantee a mention, citation, ranking, revenue result, or permanent model behavior. We measure evidence, improve eligible signals, and report uncertainty.
5. Responsible access
Production connectors prioritize official APIs, documented access, rate limits, and source attribution requirements. Methods are versioned so a change in provider behavior does not silently rewrite historical results.