Relationship One ran a two-week Forge program with an agentic coworker on restricted Eloqua API creds, drafts only. The read-only QA pass caught planted potholes plus a zero-member segment, leftover merge syntax, and a form that never wrote to the list the canvas reads.
Why it matters: Steal the operating model: restricted creds, drafts only, and a scripted pre-launch review that is allowed to fail the send.
Meltwater has YouTube as the most-cited source across eight assistants in August, and Tinuiti's Profound cut shows YouTube's share in Google AI Mode more than quadrupled from January to April. OtterlyAI: 94% of those citations are long-form, 32% sit in the 10-20 minute band; Scrunch saw citation rate fade 43% within a month.
Why it matters: Brief the next YouTube buy for chapters and a 10-20 minute how-to, and treat a citation as a 30-day asset, not a permanent win.
Comscore's panel: ChatGPT prompt share fell 70% to 50% from January to June while Gemini went 17% to 30% and Claude 2% to 11%. Tripadvisor is 61% of travel sourcing and only 21% of citations; iPullRank's Michael King says citations are the lagging half of retrieval.
Why it matters: Score visibility per assistant, and split source vs cite before you call a GEO win.
Merchant Center emails say eligible products are auto-included in AI Mode and Gemini via UCP; opt-out lives in Shopify under Sales channels > Agentic. Menachem Ani posted the client email; nic_mcdonough saw it live.
Why it matters: Open every Shopify plus Merchant Center account today and decide whether Google keeps the cart before Q4 volume hits.