GEO GLOSSARY
The process by which a language model anchors its answer in retrieved external sources — search results, crawled pages, a knowledge base — rather than relying only on what it memorized during training.
A model’s training data has a cutoff and skews toward whatever was widely written about at the time. Grounding lets a model discuss brands, prices and details that postdate training or were always too niche to be well represented in it — but only if that content is publicly crawlable.
Being crawlable is necessary but not sufficient: the model still has to retrieve the specific page for a specific question and decide it is worth citing. See retrieval-augmented generation for the underlying mechanism.
See how these terms are used in practice in MagRank AI’s methodology.
