Lost In Translation: How Language Changes Which Brand AI Recommends

Search Engine Journal’s study “Lost In Translation: How Language Changes Which Brand AI Recommends” reveals that AI‑driven recommendation engines surface different brands depending on the language of the query, based on responses from 1,650 participants. For website owners, this means that a site’s visibility in AI‑powered search results can vary dramatically across languages. A brand that ranks highly in English‑language AI recommendations may be invisible to Spanish‑speaking users if its content, metadata, or structured data aren’t optimized for that locale. Consequently, AI traffic can be fragmented, leading to missed conversion opportunities and skewed analytics that under‑report international interest. **Actionable tips:** 1. **Tag language‑specific bots in your llms.txt** – add entries like `User‑Agent: *; Allow: /; Disallow: /es/*` to control which crawlers can access each language version, then monitor the logs for AI bot hits per locale. 2. **Implement granular AI bot tracking** – use a dedicated analytics segment that captures the `User‑Agent` strings of major AI assistants (e.g., Google Gemini, Microsoft Copilot) and filters by `Accept‑Language` header to see which languages drive the most AI traffic. 3. **Create and submit hreflang tags** – ensure every language version is correctly annotated so AI models can match the right content to the right audience, improving both organic and AI‑generated recommendations.