Homepage v6 foundation went in, the engagement loop is now two failures deep
Ten new homepage sections went live Tuesday. The reply loop has been silent for 53 days and the scraper that feeds it has been silent for 12. The AI-search gap got cheaper to fix.
This week the homepage got bigger and the engagement loop got quieter. Both of those have to be true at the same time, and both are.
What shipped
The homepage v6 foundation went in on Tuesday, May 26. Ten new sections in one push: a new Hero with a perspective-tilted browser frame, an extension empty-state, an extension-data panel, a VeloIntro, a VeloBrain feature grid, a ForWhom block, the FoundingOffer with pricing and the guarantee, a ScorecardCTA, a Pendulum bridge, and a sticky floating CTA on mobile. The whole page is now a light editorial flow instead of the block-stack it was.
Header CTAs now route to /scorecard instead of opening the waitlist modal. Fewer agents are dropping out at the modal step; we will know in a week whether the dedicated page picks up the slack on completion rate.
Behind the scenes the Hero got a second pass on Tuesday afternoon — Browser.png, grain, depth shadows. The "hero + flip before brain scene" is in progress and not live yet, so no claim there.
What broke
The engagement reply loop is now silent for the 53rd day. New this week: the morning scraper that feeds the loop has also gone silent. Zero scrape files for the last 12 days. The earlier framing was "we have cards, we are not posting them." The new framing is "we have no cards." Two failures, not one.
In AI search we are at 0.0 GEO score for the seventh week running. DealFinder by Homesage.ai still owns the Chrome-extension answer slot for the fourth week. No outreach to the three listicle authors has gone out.
What we learned
The cheapest fix in the GEO loop is one that is not a content fix at all: the brand-disambiguation problem (Google auto-corrects "Unvelo" to "Invelo") moves with branded Google Ads on the exact-match keyword. We have not started that yet but it is now the top of the priority list.
The "AI-powered listing intelligence" answer slot in Perplexity and ChatGPT is currently unoccupied. HouseCanary owns valuations, SmartZip owns predictive sellers, ListingAI owns descriptions. The intelligence overlay slot is the one we are built for, and nobody is in it. That is where the next authority post is going.