The goal is not to determine whether an AI model has heard of your brand. The goal is to determine whether it recommends your brand when potential customers ask category-related questions.

AEOGeoAI tests your brand against three major AI systems — Claude (Anthropic), Gemini (Google), and ChatGPT (OpenAI) — and returns a 0–100 AI Visibility Score per model. Here is the current methodology used to generate those scores.

Why We Don't Search Your Brand Name

Most buyers discover brands through category questions, not direct brand lookups.

Someone asking "What's the best CRM for a small business?" is a potential customer in discovery mode.

Someone asking "What is HubSpot?" already knows the brand exists.

AEOGeoAI measures whether AI systems recommend your brand during discovery-stage searches — which is where visibility matters most. Queries are constructed to simulate how a real buyer would find your brand, not how someone who already knows you would look you up.

Step 1: Query Generation

Step 1 of 4
Constructing buyer discovery questions

When you enter a brand name and optional keyword, the tool constructs the questions a real buyer might ask an AI.

If you provide a keyword (for example: "project management software"), three natural-language queries are generated:

If you enter a question directly (starting with "what", "how", "which", or ending with "?"), it is used as-is alongside two variations.

If no keyword is provided, Claude generates three category-appropriate questions based on your brand name alone — questions a buyer would plausibly ask that would result in your brand being recommended if the AI knows it well. The brand name is never included in those questions.

Step 2: Live AI Queries

Step 2 of 4
Querying each model independently

Each generated query is sent simultaneously to three AI systems. No model sees another model's answer. All requests run in parallel so results return in seconds.

Claude
claude-haiku-4-5
Anthropic
Gemini
gemini-2.5-flash-lite
Google
ChatGPT
gpt-4o-mini
OpenAI

Step 3: Scoring

Step 3 of 4
Measuring brand presence and prominence

Each AI response is analysed for brand mentions. The AI Visibility Score measures brand presence and prominence within AI responses. It does not currently measure sentiment or recommendation quality — a brand mentioned frequently in a negative context would still receive a higher presence score than one not mentioned at all.

The tool handles brand name variants automatically. If you enter "booking.com", it checks for "booking.com", "booking", and spaced variants — so partial mentions are captured alongside exact matches.

Signal Points
Brand appears anywhere in the response +35
Exact brand name or domain match +15
Brand mentioned two or more times +25
Brand appears in the first 30% of the response +10
Brand mentioned three or more times +10
Domain extension present and exact match +10
Brand followed by a colon (product recommendation format) +15

Raw scores are capped at 100. Not all signals can apply simultaneously, but where they do, the cap applies.

Step 4: Excerpt Extraction

Step 4 of 4
Capturing what the AI actually said

When a mention is found, the tool captures the surrounding sentence context — typically 200–400 characters around the first mention. This is the excerpt displayed in your results, showing you exactly how the AI described or referenced your brand in its own words.

Score Interpretation

90–100 · Dominant

AI prominently and repeatedly includes your brand when asked relevant category questions.

70–89 · Visible

AI consistently includes your brand in relevant recommendations.

40–69 · Emerging

AI mentions your brand but weakly or infrequently.

AI does not mention your brand when asked relevant category questions.

Why Scores Differ Between Models

Claude, Gemini, and ChatGPT are trained on different datasets, updated at different times, and use different retrieval systems. As a result, the same brand may receive different scores across models. These differences are often as valuable as the scores themselves — they highlight where your brand's evidence footprint is strongest or weakest across the AI ecosystem.

A high score on ChatGPT and a low score on Gemini may indicate differences in how your brand appears across Google's ecosystem and Google-indexed content, though many factors can contribute to model-to-model variation.

Score gaps between models are a diagnostic tool, not just a number. They tell you where your brand's evidence is missing.

What the Score Does Not Measure

AI Visibility Report (Pro)

Pro users receive a written AI Visibility Report generated by Claude after each scan. The report interprets your specific scores, identifies likely causes, names the platforms probably influencing your citation footprint, and gives three specific improvement actions. Reports are generated fresh for each scan — not templated.

See your AI Visibility Score

Free check across ChatGPT, Claude and Gemini — no account required.

Check your brand free →

Infrastructure

Cloudflare Workers All API requests and scoring logic run at the edge
Cloudflare AI Gateway Routes requests to Anthropic, Google and OpenAI with reliability and rate control
Cloudflare D1 Scan results stored for Pro scan history and CSV export
Cloudflare KV Rate limiting and Pro account status
Data retention No personally identifiable data is stored for free users

AI Visibility Scores are directional metrics designed to help compare AI brand visibility over time and against competitors. They should not be interpreted as definitive measurements of market share, brand awareness, or commercial performance. The methodology described on this page reflects the current implementation and may be updated as the tool evolves.