What Google-Extended is and isn't
Google-Extended is a user-agent token Google uses to determine whether a site's content can be used to train Gemini and other Google generative AI products. Blocking Google-Extended in robots.txt does NOT affect classical Google Search — Googlebot is a separate user-agent. The opt-out applies only to AI training. AI Overviews uses the classical Google index, so blocking Google-Extended doesn't necessarily exclude a Shopify store from AI Overviews.
Google-Extended vs Googlebot
Two different bots, two different roles. Googlebot crawls for the classical Google Search index and powers AI Overviews. Google-Extended crawls for AI training. A Shopify merchant can allow Googlebot (mandatory for any SEO traffic) and block Google-Extended (training opt-out) at the same time — the two decisions are independent.
The opt-out trade-off
Block Google-Extended → your store is excluded from Gemini's training corpus and other Google generative AI training. You may still surface in Gemini if Gemini fetches your page live during a conversation (similar to ChatGPT-User behavior). You retain full classical Google Search visibility because Googlebot is unaffected. The trade-off: future Gemini may not 'know' your brand as well as competitors who didn't block training, but current Gemini visibility from live-fetch and Catalog isn't lost.
Google-Extended audit checklist
Four checks. The Google-Extended decision is deliberate; if the file doesn't reference it, the default is allow.