How does GEO work in the USA — and in US Spanish?

The scale is measured, and it is no longer a side channel: 49% of US adults use AI chatbots, 60% encounter AI summaries in search — and when an AI summary appears, clicks on regular results drop by roughly half (8% of visits versus 15% without one), with 26% of those sessions ending in no click at all. For English commercial prompts — "best X in the US", "top X for small business" — assistants lean on a familiar supply chain: industry listicles, G2 and Clutch reviews, major trade media. Research on generative engines and our own testing point the same way: models prefer sources they can corroborate, and in the US English corpus there is always something to corroborate against.

The consequence is strategic: a US English program is mostly off-site work. Your own pages make you eligible; placement in the lists and review platforms assistants already cite is what gets you named. That takes months, and any agency promising otherwise is selling against the mechanics.

68 million US residents are Hispanic — a fifth of the country. 68% of Hispanic households speak Spanish at home, and Hispanic adults use AI chatbots at least as often as white adults (49% versus 46%, Pew). The purchase questions follow: "mejor abogado de inmigración en Houston", "mejor software de nómina para pequeños negocios". Yet the Spanish-language corpus about US businesses is dramatically thinner than the English one: fewer review pages, fewer listicles, fewer localized sites worth citing.

Thin corpora are where structure wins fastest — the same effect we describe for Hebrew and for Ticino Italian. A US brand with native-quality Spanish pages, a Spanish-declared entity and a handful of consistent Spanish-language mentions can become the source assistants cite for an entire category, while its English competitors fight over one listicle slot.

The English battery is intent-heavy and local: "best X in the US", state and metro variants (Texas, Miami, Bay Area), "for small business" and price qualifiers. The Spanish battery mirrors it with its own phrasing — "mejor X en Estados Unidos", "cerca de mí" patterns — plus bilingual constructions where an English brand name sits inside a Spanish sentence.

The two batteries must be sampled separately and compared: the gap between English and Spanish share of voice is the size of your bilingual opportunity, stated as a number. For most US categories that gap is enormous and nobody is measuring it.

US corroboration runs through review infrastructure more than anywhere else, and this is measured: in an analysis of 30 million sources cited by AI search, G2 and Yelp rank sixth and seventh among all cited domains, with Trustpilot and Clutch inside the top hundred. A claimed, consistent, honestly-reviewed presence on the two or three platforms that matter for your category is entity work of the first order. But note where the heaviest citation traffic actually goes — community and media surfaces like Reddit, YouTube, Wikipedia and Forbes — so review-platform presence is necessary, not sufficient.

For the Spanish track, the entity must be declared bilingually: the same organization, the same facts, alternate descriptions in both languages, hreflang between the versions. Assistants notice when the Spanish description of a brand diverges from the English one — our measurement treats that language-consistency check as a first-class metric.

English and Spanish are both native-coverage languages on our own site — part of the nine-language foundation documented on our site and tested in our public case study. A US engagement typically splits: off-site corroboration and review-platform work for the English track, and content plus entity construction for the Spanish track, where results arrive faster because the field is emptier.

Measurement is bilingual by default: fixed prompt batteries in both languages, share of voice per assistant, and the English–Spanish consistency check. The first deliverable is a baseline that shows, in numbers, which answers you already appear in, who is cited instead of you — and how large the unclaimed Spanish territory in your category actually is.