Skip to main content
Every time your prompts run across platforms like ChatGPT, Gemini, AI Overviews, and Perplexity AI, we capture the generated response. These responses are what we call chats. All your metrics — visibility, position, sentiment, and even source analysis — are derived from these chats. That’s why it’s important not to skip this layer and rely only on aggregated data. Each chat represents what a real user would see when they ask a similar question. Reviewing chats regularly helps you understand not just whether your brand appears, but how and why it appears.

​
Anatomy of a chat

What each chat includes

Each chat contains multiple elements that contribute to your analysis:
  • Response content: The full AI-generated answer. This is the most important part because it shows how your brand is positioned in context.
  • Brands mentioned: A list of all brands referenced in the response, including your own and competitors.
  • Position: The order in which your brand appears. Being mentioned earlier generally indicates higher authority and relevance.
  • Sentiment: The tone used when referencing your brand — whether it is positive, neutral, or negative.
  • Platform and context: The AI model used and the environment in which the response was generated.
  • Sources: The URLs and domains that influenced the response.
One important detail to keep in mind is that your position is calculated against all brands mentioned in the response, not just the competitors you have added. This gives you a more realistic picture of where you stand in the broader AI-generated landscape.

​
Sources vs citations

Important distinction

Not all sources are visible in the response.
  • Citations are the sources explicitly referenced in the answer.
  • Sources include all content the AI used to generate the response, even if not directly shown.
Both are important. Citations influence visibility and user trust, while background sources influence how AI understands your brand and category.

​
How to approach chat analysis

Instead of focusing on individual responses, it’s better to look for patterns over time. AI outputs can vary slightly day to day, but consistent trends are what matter. When reviewing chats, try to identify:
  • Which competitors appear most frequently
  • Whether your brand is consistently included or missing
  • How your brand is described compared to others
  • Which types of sources are being used

Next → Understanding Your Performance

Learn how chats translate into metrics and trends.