Docs
Documentation
How RankVyze works, what the numbers mean, and how the workflow fits together.
Last updated
AI Visibility Score
A 0–100 composite computed from the most recent check of every tracked prompt on every engine. It weights mention rate (50%), citation rate (30%) and mention position (20%). Position credit decays from 100% for first place to 25% beyond fifth.
The score is recomputed from raw research results every time the dashboard loads, so a corrected result changes the score immediately.
Research
Each research session checks every active prompt on every engine and records: whether your brand was mentioned, at what position, whether a page of yours was cited, which competitors were named, and a short summary of the answer. Only the latest check per prompt and engine counts toward current metrics; earlier checks remain in history.
In this version, research is performed by RankVyze analysts. The data model records the source of each result so that API-based research can be introduced without changing any calculation.
AEO Audit
Audits score six categories from 0–100: AI Understanding, Content, Structured Data, Technical Accessibility, Entity Signals and Authority. Each audit produces issues with a severity, an impact score (0–10), affected pages and a recommended implementation.
Optimization workflow
- Issue → Optimization: every issue can become a planned optimization with an effort estimate.
- Fix with AI: queues an implementation job. In this version a RankVyze engineer authors the change; the job record is designed for automated agent runs.
- Review: changes appear as diffs. Approve, reject, or create a pull request when a repository is connected.
- Verify: the next audit and research run measure the effect.
Connections
GitHub, Shopify, WordPress and code upload connections are recorded as requests and completed by our team during onboarding. Automated OAuth flows are in development.
Standards we rely on
The structured data we recommend, and generate, follows these specifications rather than anything proprietary to us.