Readiness is not a result.
GEO Score (0–100) shows whether AI systems can access, understand and cite your website, across six signal groups checked by Geolify. It does not show whether AI engines mention you, so we never report it as a result.
How We Measure GEO Results
Every result Freeways reports is counted from real AI answers on a fixed prompt set, under the rules below. The data is collected with Geolify, our measurement tool.
PRINCIPLES
How We Think About GEO Data
Results come from real answers.
Mentions, citations and recommendations are counted from what AI engines actually say to a fixed set of buyer prompts.
One answer proves nothing.
AI answers change between runs, so every prompt is repeated and results are read as rates, not single screenshots.
Every number has a denominator.
We show what was counted, what was excluded and why.
WHAT WE REPORT
Four visibility metrics
| Metric | Counts as a hit when… | Denominator |
|---|---|---|
| AI Visibility | The answer names the brand | Valid answers to tracked prompts |
| Citation Coverage | The answer cites an owned or earned page that supports the brand | Valid answers to tracked prompts |
| Top-3 Recommendation Presence | The brand is among the first 3 options the answer recommends | Valid answers to recommendation prompts |
| Answer Accuracy | Key brand facts match the brand's approved fact sheet | Valid answers to branded prompts |
COLLECTION
How each round runs
| Setting | How we run it |
|---|---|
| Platforms | ChatGPT with web search on, Perplexity, Gemini, Google AI Overviews & Claude |
| Market and language | Set per brand (for example Vietnam, in Vietnamese) and fixed across rounds |
| Prompt set | Fixed and versioned, split into branded and non-branded prompts about 20% / 80%. Each group is reported separately. |
| Repeats | Each prompt runs 2 times per platform per round, because AI answers vary between runs |
| Frequency | Baseline, then every 2 weeks |
Freeways GEO Measurement Methodology v1.0, effective January 1, 2025.