Start Tracking Your Brand
Learn how to analyze the AI-generated answers that power your visibility insights.
Every metric inside Cite AI starts with a response generated from your tracked prompts. These AI responses form the foundation of:
Understanding how these responses work helps you interpret your analytics more effectively.

Each time Cite AI runs a prompt across platforms like ChatGPT, Gemini, Google AI Overviews, Google AI Mode, or Perplexity, the generated answer is stored as a tracked response.
These responses are analyzed to understand:
You can access:
Each prompt stores previous response history so you can monitor changes over time.
Every tracked response contains multiple components that contribute to your visibility analysis.
Indicates whether the prompt completed successfully.
Shows which model generated the response.
Examples:
Shows the geographic region used for the prompt request.
AI responses can vary significantly by country or region.
This is the full AI-generated answer returned for the prompt.
Cite AI analyzes this response to determine:
Tracks:
This data powers:
Cite AI captures the websites and sources referenced while generating responses.
This helps you understand:
Sources and citations are related, but not identical.
Citations are websites directly referenced inside the AI response itself.
These are explicit references tied to specific parts of the answer.
Citations usually represent stronger visibility because users can directly see the source association.
Sources include all websites used during response generation, even if they were not explicitly referenced inside the final answer.
A source may influence the AI response without being visibly cited.
An AI response may:
All 8 influenced the response, but only 4 appeared visibly within the answer.
Help drive:
Help build:
Both are important for long-term AI visibility growth.
Each AI platform handles responses, citations, and sources differently.
Sometimes uses web search and sometimes responds from existing model knowledge.
When no web search occurs, responses may contain no visible sources.
This is expected behavior.
Typically references many sources during generation but cites fewer directly inside the response.
This often results in:
Each platform has unique behaviors around:
Response patterns naturally vary across models.
When multiple brands appear in a response, Cite AI calculates ranking based on mention order.
Example
If a response mentions:
then Puma receives position 3 for that response.
If another response includes:
then Puma’s position becomes 4.
This ranking includes all detected brands, not just manually tracked competitors.
Earlier mentions generally indicate stronger authority and relevance.
Brands appearing closer to position 1 are more likely to be:
Tracking position over time helps measure whether your brand is gaining or losing prominence across AI platforms.