Grounding

Grounding is how an AI engine keeps its answers honest. Instead of generating from memory alone — the path to hallucination — a grounded system retrieves real sources at query time and builds the answer on top of them. Retrieval-Augmented Generation (RAG) is the most common grounding method.

Grounding is why citations exist. When ChatGPT, Perplexity, or Google AI Overviews footnote a source, that source is the grounding material — the evidence the engine chose to stand on.

For brands, grounding is the whole game. Your content is either eligible to be grounding material — extractable, authoritative, structurally clean — or it isn't. Becoming the source an engine grounds its answer in is the operational definition of citation share.