What Your ARI Score Means
ARI (AI Readability Index): 1–100 scale measuring how well AI agents understand you.
Score Ranges
80–100: Excellent - AI agents understand you clearly - You show up in recommendations - Your positioning survives compression - Agents can match you to customer needs
60–79: Good - AI agents get most of it - You show up in recommendations (but not top) - Some clarity gaps remain - Fix biggest gaps for 80+
40–59: Fair - AI agents get the gist but unclear details - You might appear in recommendations (depends on customer query) - Major clarity needed - Fix 3–4 things to get to 60+
20–39: Weak - AI agents flatten you to category average - You're not recommended unless customer asks by name - Complete positioning overhaul needed
<20: Severe - AI agents can't understand what you do - You're invisible to agent-mediated discovery - Urgent: rebuild product pages, add specs, clarify positioning
Score Breakdown: The 5 Dimensions
Your ARI report breaks down into 5 dimensions:
- Opinion Density (Do you take positions?)
- Factual Provability (Are claims backed by data?)
- DOM Extractability (Can AI easily pull your data?)
- Semantic Specificity (Is your positioning specific?)
- Differentiation Signal (Do you stand out or blend in?)
Each dimension is scored 1–100.
Your overall ARI is the average.
Improvement Path
If you score 40–59: - Fix 1 dimension (biggest gap) to move from 40 → 50 - Fix another dimension to move from 50 → 60 - Takes 2–4 weeks per dimension
If you score 60–79: - 1–2 quick wins get you to 70+ - Usually: add schema + clarify positioning - Takes 1–2 weeks
If you score 80+: - Maintenance only - Monthly reviews to keep score high
Quick Wins by Dimension
Low Opinion Density? → Rewrite product description to be more opinionated about who you serve
Low Factual Provability? → Replace vague claims with data (add case studies, metrics, testing results)
Low DOM Extractability? → Add structure (headers, bullets), add schema markup
Low Semantic Specificity? → Narrow your audience ("for teams 5–50" not "for teams")
Low Differentiation Signal? → Add comparison section, explain how you're different