How GEO Scoring Works
Every score GEOREX shows is built from a real scan result, not a plausible-looking default. If there's no real evidence yet, GEOREX says so rather than quietly filling in a number that looks fine. Here's the actual pipeline behind that.
Mention vs. citation
Each scan records, per engine, whether a response mentions your brand in passing or cites it as the source of the answer. These are tracked as two separate numbers, not blended into one, because they're different outcomes with different fixes.
Category Health: 12 dimensions
Technical Foundation, Content & Architecture, and AI & Authority Signals are scored and reported independently, so a strong AI-readiness score never hides a real technical problem underneath it. Full breakdown in Overview & GEO Health Score.

From score to task
A low score on its own doesn't tell you what to do. Every gap the scoring pipeline finds is passed through a Rule Engine and a set of Generators, deduplicated and quality-gated, before it becomes a task in your pipeline, see Tasks for what happens next.
Why this matters more than a single score
