Before Autopilot starts training, GEOFixer runs one full sweep of Perplexity's sonar-pro model with your prompts and records the result. That snapshot — the baseline — never moves. Every subsequent cycle's number is meaningful only against it.
Absolute citation rates are nearly useless on their own. A 30% cited rate sounds bad until you learn that the brand started at 8% and clawed to 30 over six weeks. A 60% cited rate sounds great until you learn it dropped from 78. Without a fixed reference point, every conversation about whether Autopilot is working becomes vibes.
The Perplexity Baseline pins one number in place. From that moment on, the only question is: which direction is the gap moving, and how fast?
The baseline locks after Autopilot's first full Perplexity sweep — typically inside the first 24 hours of activation. Until it locks, the panel reads "Data builds as Autopilot runs. Check back in 24h." Once locked, the number is permanent. If you ever want to re-baseline (after a major brand pivot, for example), it's a manual operation in account settings.
The three columns are baseline cited rate, baseline average position when cited, and the three competitors most frequently surfaced when your brand wasn't. Beneath the panel is a link to the full Perplexity pre-scan report — top prompts tested, sample cited and non-cited responses, per-persona breakdown.
Open GEOFixer →