BrandCitation Methodology

What we measure

BrandCitation measures AI search visibility by sending buyer-intent prompts to AI engines and analyzing the responses. For each prompt and engine combination, we track:

• Whether your brand is mentioned
• Whether your domain is cited as a source
• How often competitors appear
• Position in recommendation lists
• Sentiment (positive, neutral, negative)
• Source URLs and domains that influence answers
• Visibility changes over time (for monitoring plans)

Important disclaimer

API-based results may differ from consumer UI results in ChatGPT, Gemini, Claude, Perplexity, or Google AI products. BrandCitation uses official APIs because they are scalable and reliable for trend monitoring. We clearly state this in every report.

AI engines

Gemini

Broad AI model visibility and cost-effective testing at scale.

Perplexity

Web-grounded AI search with source citations — the most valuable engine for citation analysis.

OpenAI-powered models

Additional model coverage. API results do not perfectly equal ChatGPT web browsing answers.

Prompt categories

Brand prompts

Test whether AI understands your brand entity: “What is {brand}?”, “Is {brand} legit?”

Category prompts

Test discovery when users do not know your brand: “Best {category} in {market}”

Competitor prompts

Benchmark against alternatives: “{brand} vs {competitor}”, “Alternatives to {competitor}”

Problem/solution prompts

Real buyer problems: “I need a tool to {job-to-be-done}”

Local and MENA prompts

Regional visibility in English and Arabic, including Egyptian dialect variations.

Sampling and cadence

LLMs are probabilistic — one answer does not represent permanent visibility. The free scan runs once. Paid audits run 15–50 prompts across engines. Monitoring plans run weekly (or daily for premium tiers) to track trends.

Brand matching

We detect brand presence via exact name match, case-insensitive match, and domain match in response text.

Citation detection

Citations are counted when:

• The engine returns URL sources directly (Perplexity-style)
• Response text includes your domain URL
• Cited source page domain matches your domain

Competitor share-of-voice

For each prompt set: your brand mentions divided by total mentions of all tracked brands. Example: 2 client mentions vs. 8 competitor A + 5 competitor B = 13.3% share-of-voice.

Recommendations

Action plans are based on missing mentions, citation gaps, competitor source patterns, technical GEO readiness, content gaps, and Arabic/English coverage imbalances.

Try it

Run a free scan or read how we calculate scores.