Each cell tells you how well your brand surfaces when a particular kind of buyer asks a particular kind of model. Lit cells = strong coverage. Gray cells = no signal yet.
Rows are personas — the audience segments Autopilot is training for. You define these on the brand profile (think: "founder considering buying", "engineering lead evaluating", "marketing manager researching"). Columns are LLMs — each model Autopilot is actively auditing for you, including GPT-4o, Claude Sonnet, Gemini, Perplexity sonar-pro, and Google AI Overviews.
Each cell is the visibility score for that (persona, LLM) pair over the last 28 days. Visibility score is the percentage of audit conversations where your brand was the top recommendation for that persona's queries on that model.
Click any cell to drill into the underlying conversations for that (persona, LLM) pair. You'll see the actual prompts, the model's full response, and whether your brand was cited or what it cited instead. That's the bridge from "this cell is red" to "here's the specific gap that needs to be closed."
Most users use the matrix as a defense layer: glance at it weekly, look for cells that have slipped from green to yellow, and run a fresh Autopilot cycle on those personas' prompts.
Open GEOFixer →