AI Gives Different Answers in Different Languages
The same business query asked in different languages can produce completely different AI recommendations. Why Language Visibility matters.
Resources
AI visibility is a young field, and much of it is genuinely unsettled. This page collects what we are publishing, the primary sources we rely on, and the questions we do not yet have firm answers to.
Written to answer a specific question, including where the honest answer is that nobody knows for certain yet.
The same business query asked in different languages can produce completely different AI recommendations. Why Language Visibility matters.
Many websites are easy for people to understand but provide very little structured information for AI systems.
An invitation to work with three companies while improving AIVisibility.global.
Longer reading for the thinking behind the work rather than the tactics.
Author name
A short note on why this book is worth reading for anyone thinking about how machines read a business.
Author name
A short note on what this book covers and which part of AI visibility it speaks to.
Author name
A short note on the argument this book makes and why it belongs on this list.
Primary sources. Where these disagree with commentary about AI visibility — including ours — trust these.
Google's own documentation on how pages are crawled, indexed and represented in search.
The shared vocabulary for structured data that search engines and AI systems read.
The standards body behind HTML, accessibility and the wider architecture of the web.
Documentation and published policy for the models behind ChatGPT.
Documentation and published research for Claude, including how it uses the web.
Microsoft's technical documentation, covering Bing, Copilot and Azure AI.
The open web crawl archive that many language models are trained on.
The open questions we keep returning to. None of these are settled.
What makes a business look credible to a machine that has never met it.
How the language of a question changes which businesses an AI names.
When an AI answer credits a source, and what decides which source it picks.
The machine-readable facts a site publishes about itself, and how they are used.
Building sites that stay legible to software, not only to people.
A free check shows how AI systems currently describe your business, and which of these topics actually apply to you.