How Your AEO Score Works
Your AEO Score measures how well AI engines like ChatGPT, Claude, Gemini, Perplexity, and Google AI Overviews know and recommend your brand. It’s a 0–100 composite of four dimensions.
The Four Components
Presence
25%How often your brand appears when people ask AI about your category. We only count organic appearances — if your brand was in the question (e.g., “What is [Brand]?”), it doesn’t count toward Presence.
Quality
30%Whether AI mentions are positive recommendations, neutral namechecks, or warnings. A brand that’s mentioned negatively scores lower than one that’s not mentioned at all on this dimension. This is the heaviest-weighted component because what AI says about you matters more than whether it mentions you.
Prominence
25%Whether you’re recommended as a top choice or just listed among many options. We look at your ranking position and how strongly AI recommends you. For comparison queries (“Brand vs Competitor”), we only count recommendation strength, not position — being listed first in a comparison you asked about isn’t meaningful.
Consensus
20%Whether multiple AI engines agree about you. Being recommended by 4 out of 5 engines is a much stronger signal than being recommended by just 1.
Score Bands
| Band | Score Range |
|---|---|
| Invisible | 0 – 10 |
| Minimal | 11 – 25 |
| Emerging | 26 – 40 |
| Moderate | 41 – 55 |
| Established | 56 – 70 |
| Strong | 71 – 85 |
| Dominant | 86 – 100 |
How We Classify Mentions
For each AI response, we analyze: the query’s intent (discovery, comparison, branded, alternative, educational), the sentiment toward your brand (positive, neutral, negative, confused, hallucinated), and the strength of any recommendation (0–100 scale). This classification uses AI to evaluate AI — giving us nuance that simple keyword matching can’t provide.
What About AI Randomness?
AI engines don’t give the same answer every time. Research shows that identical questions produce different brand recommendations on nearly every run. We address this by:
- Scanning across 5 different AI engines (consensus reduces noise)
- Using 10 diverse queries per scan (more data points = more reliable signal)
- Classifying quality not just quantity (a consistent negative mention is worse than an inconsistent positive one)
Your score is a snapshot, not a permanent truth. We recommend re-scanning periodically to track trends.
Score Interpretation Tips
- Don’t fixate on the exact number — focus on which component is weakest
- Low Presence? Create more category content that AI can discover
- Low Quality? Fix your brand narrative — check the Issues section for specifics
- Low Prominence? Build authority through comparison content and case studies
- Low Consensus? Target the specific engines that don’t mention you