The in-app chat agent can now generate image cards from your real data. Ask who AI platforms name in your category, how your questions break down by win/contested/opening, or what your quick wins are — the card renders right in the conversation and is ready to share.
Recommendations on your dashboard got smarter about what you already have. The next-action engine now reads a snapshot of your real website before suggesting new pages, so it won't prompt you to publish a profile or FAQ page your site already covers. It also checks your online presence baseline — gaps like a missing LinkedIn company page or a brand website that AI platforms don't yet cite now surface as concrete action items, each routed to the right fix.
The online presence baseline itself expanded this release. It now checks G2, Trustpilot, and Capterra for review presence and looks for a Wikipedia page — returning real gaps ("not on G2", "no Wikipedia page") rather than silence. It also refreshes automatically after a visibility scan completes or after you save social handles, so it stays current without you needing to trigger it manually.
Agent-generated pages (FAQ, services, custom pages) now pull from your actual website content when building copy, not just your factsheet. If your site already covers a topic, the generated page is built from that material rather than written from scratch.
Visibility scan coverage now rotates across your full inventory of personas and prompts. Previously, accounts with more personas or prompts than a scan's query budget could handle would have the same subset permanently skipped — every scan. Rotation means the full inventory gets covered across successive scans. The SMS agent's reply voice also holds correctly on long conversations, dropping the informal shorthand and clean-ending every message.