Methodology
How we measure AI visibility
How we ask the questions, score every answer and combine the results, and what can move the numbers.
Read the full methodologyFive AI assistants
Questions are asked to ChatGPT, Gemini, Claude, Perplexity and Grok through their official APIs, with web search on by default.
Your questions
You choose the questions, topics, competitors and markets. Each question has an importance level: high, normal or low.
Any category
Scans work for the categories, markets and languages you track: SaaS, retail, marketplaces, local services and more.
Core metrics
The numbers behind every report.
Visibility score
Each answer scores 0–100: mention (40), early mention (15), your website named or linked (15), list position (up to 30), positive tone (10) and trust (10), capped at 100.
Share of mentions
When an answer mentions you or a tracked competitor, how often it is you.
Link rate
The share of answers that mention you and also name or link your website.
Coverage
The share of questions that received a complete answer. Incomplete answers are left out of scores.
Quality controls
Checks that keep the numbers honest.
Links to your website are matched against the domains you add for your brand.
Duplicate questions are detected, so the same question can't inflate your results.
Cut-off, refused, empty or failed answers are left out of scoring instead of counting as zero.
Scan schedule
Weekly by default
Scans run weekly by default. You can switch to daily scans, or start a scan at any time.
Limitations
What can move the numbers.
Providers update their models, which can change answer patterns overnight.
The intent behind a question can drift over time.
Results vary with location, freshness of information and exact phrasing.
Methodology FAQ
Common questions from teams new to GEO.
See what AI assistants say about you
Start a 14-day free trial, or book a demo and we'll show you how Geovium works for your category.