Methodology

Our Methodology

How we calculate your AI Visibility Index — and why every score has a source label

We are honest about what is measured and what is estimated

Your FoundByAI report contains two types of data: live-tested results and calibrated estimates. We label every score clearly so you always know which is which.

✓ Live tested

Claude API test

We send your 12 test queries to the Claude API and record whether your business appears in each response as Visible, Partial, or Invisible. These results are real, specific to your business, and conducted at the time of your audit.

~ GEO estimated

ChatGPT · Perplexity · Gemini · Copilot

Scores for these engines are calculated using our GEO calibration model, which applies each engine's known ranking signals to your business data. We update this model quarterly. We estimate, and we tell you that clearly.

How the AI Visibility Index is calculated

The master score is a weighted composite of four sub-scores, each calculated differently depending on your business type.

35%
AI Engine Visibility
25%
Entity Consistency
25%
Content Coverage
15%
Review Quality

How each AI engine score is calculated

For the four estimated engines, we apply a calibration model built from published GEO research. Expand each engine to see the signals used.

How sub-scores are assessed

Sub-scores are assessed from the information you provide during your audit. Your score improves as you provide more complete data and implement the recommended actions.

AI Engine VisibilityMax 35 pts
Claude live test (12 queries) — Visible+2.9 per query
Claude live test — Partial mention+1.5 per query
Claude live test — Invisible+0
Other engines calibrated from GEO modelup to 35 pts total
Entity ConsistencyMax 25 pts
Google Business Profile present and verified+6
LinkedIn company page present+4
Crunchbase / Product Hunt listing+3 each
Facebook / Instagram presence+2 each
NAP consistency across listings+4
Content CoverageMax 25 pts
FAQ page present on website+5
Schema markup (LocalBusiness / Article / SoftwareApp)+5
Pillar content / topic cluster present+4
Named authors with bio pages (blog / media)+4
Average article length > 1,200 words+3
Internal linking structure+2
Publish frequency (weekly+)+2
Review QualityMax 15 pts
50+ Google reviews+5
20–49 Google reviews+3
1–19 Google reviews+1
Review sentiment includes service keywords+5
Review response rate > 50%+3
External review platforms (Trustpilot, G2, etc.)+2

How we keep the model accurate

We review and update our GEO calibration model quarterly. Each update incorporates published research, community feedback, and manual validation tests comparing estimated scores against real AI responses.

Version
Date
Notes
v1
April 2026
Initial model launch

Ready to see your score?

Run your free AI Visibility Audit in 2 minutes — no credit card required.