Visibility scores — the headline number, platform breakdown, and persona breakdown — now all count from the same basis: open-ended prompts where an AI could recommend anyone. Branded prompts, where your company name appears in the question and the AI almost always surfaces you, no longer pull any of those numbers upward. Older scans are recalculated the same way, so the dashboard reflects what the score always should have been.
Platforms that ran during a scan but weren't showing up in the Conversations tab or the platform heatmap now appear correctly. If you were wondering why certain platforms looked absent despite running, that's why.
When a prompt fully fails during a scan, the topics score column and the Conversations view now fall back to the most recent scan that has data for that prompt, the same way the prompt detail pane already did. The "—" score and "No response captured" placeholder are gone.