What is RAG (retrieval-augmented generation)?

How RAG works

1. The user asks a question.
2. The system searches an index, often a web search index, for relevant passages.
3. The best passages are given to the language model as context.
4. The model writes an answer based on that context and may cite the sources.

Why it matters for visibility

When an assistant such as ChatGPT search, Perplexity or Google’s AI Overviews cites a page, it is usually because that page was retrieved first. Being retrievable therefore comes before being quoted: your content must be accessible to crawlers, clearly structured and relevant to the question.

What makes a page easy to retrieve and cite

• It can be crawled: not blocked in robots.txt, rendered without heavy scripts, and present in your sitemap.
• It answers a specific question directly, in a self-contained passage.
• It uses clear headings, short paragraphs and structured data.
• It carries trustworthy signals: named authors, sources, consistent business details and independent mentions.

RAG and SEO are linked

Because many assistants retrieve from search indexes, classic SEO still feeds AI visibility. Strong rankings, clean technical foundations and content built from real questions improve the odds of being retrieved and cited. That overlap is why we treat GEO as an extension of SEO rather than a separate discipline.