Autopilot no longer requires you to manually find and add questions you're missing. After each weekly scan, up to three new questions are pulled into your topic areas — sourced from Search Console data, AI fanout signals, and your profile — as long as you're within your plan's question limit. The next scan measures them immediately, so your coverage expands without extra work.
The Create agent now matches your request to a better starting workflow. If your latest scan has nothing relevant to add, it runs on "Answer from what you know" rather than forcing a scan-based workflow. A request about one of your own services now correctly routes to "Explain a service" and picks up that service automatically. Workflows you've saved yourself still take priority.
Research articles now label each brand's score "AI Visibility" instead of "AI Sentiment," which previously looked like it contradicted the sentiment score on a brand's own AI report. Brand write-ups that were running several bolded points together now break into separate paragraphs, and brand cards on research articles and city guides pull in the brand's own share image from its website when one is available.