# Grounding

> Tying a generated answer to a retrievable source, so the claim can be attributed rather than invented.

Source: https://anotherseoguru.com/en/glossary/grounding

Grounding is how an answer engine anchors what it says to something it retrieved. For a website, being useful as a grounding source is the whole of the available leverage: you cannot control how an answer is rendered, only whether your page supplies data clean enough to be quoted. That favours structured data matching the visible content, and answers stated plainly near the question.

## Example

A hotel page publishing price, coordinates and three captioned images gives an assistant something to attribute; a page of prose does not.

## How to apply it

Match structured data to what is visibly on the page, answer the question in the first two sentences under the heading, and render the text server-side.

> **Pro tip:** Grounding is retrieval, not training. Being in a model's training data and being retrievable today are different problems with different fixes.

## Related terms

- [AI Mode](https://anotherseoguru.com/en/glossary/ai-mode)
- [LLM SEO](https://anotherseoguru.com/en/glossary/llm-seo)