The in-app agent now covers the full blog post lifecycle — regenerating the AI text, creating a hero image, scheduling or unscheduling, moving back to draft, and updating the slug. It also picks up two settings tools (brand name and content language) and a scan status check. All of these match what the web UI and external MCP already do, so the agent can take a post from idea to scheduled publish without you switching context.
The Yelp Ads experience got a meaningful rework. Before you connect, the panel now leads with your AI-visibility score so you can see where you stand across AI assistants and what a sponsored placement would change. Once connected, a "Your Yelp standing" panel shows your rating, review count, and category alongside a live ranking check — your real position in the Yelp category search, whether you're showing as a sponsored ad, and which competitors are running ads. The Yelp beta also now qualifies any local business with a Google presence, not just businesses with an existing Yelp listing.
The Ads channel picker was rebuilt as a multi-channel view. The connect flow confirms the listing already resolved at signup rather than asking for an address again, the campaign budget is a directly editable field, and the ad preview renders as a real promoted-listing modal.
Team members can now delete the scan prompts they create and edit — topics, queries, and buyer personas — across the app, the in-app agent, and MCP. Previously a teammate could add a bad query but couldn't remove it without going to an owner or admin. Account-level actions like deleting the agent itself remain owner-only.
The /manage page no longer stalls on a blank screen or bounces to the homepage while your session loads. A live Yelp ranking lookup that was silently switched off due to a misconfigured environment variable now works correctly.