Retrieval-augmented generationRAG
Fetching relevant documents first, then writing an answer from them. It is how most AI search works.
In plain English
RAG has two steps: retrieve the best passages for a question, then generate an answer using them. If your page is not retrieved, it cannot be used, which is why being findable and quotable both matter.
In practice
For “do I need a permit for a deck in Austin?”, the AI retrieves the city’s permit page and a contractor’s guide, then writes its answer from both.
Related terms
How AI reads the webPro Grounding When an AI checks live sources before answering, instead of relying only on what it learned in training. In practice, and related terms → Content & structurePro Passage A short section of a page that AI tools can pull out and use on its own. In practice, and related terms → How AI reads the webPro Embedding A list of numbers that captures what a passage means, so AI can match it to questions worded differently. In practice, and related terms →