Grounding
Tying an AI answer to real, checkable sources instead of the model's memory alone.
Grounded answers can be trusted and cited. Ungrounded ones are just confident guesses.
Frequently asked questions
What does it mean for an AI answer to be grounded?
Grounding means the answer is tied to real, checkable sources rather than resting on the model's memory alone. A grounded answer can be trusted and traced back; an ungrounded one is just a confident guess.
How is grounding different from RAG?
Grounding is the goal, an answer anchored to real sources, and RAG is a common technique for achieving it by retrieving documents and feeding them to the model. In short, RAG is one way to make answers grounded.
Why does grounding matter for my business?
Because grounded answers can be cited and verified, which is what you need before acting on AI output for customers, contracts, or money. Without grounding you are trusting the model's word, which can sound right and still be wrong.
New to all this? Start with what an AI agent really is, browse the full glossary, or explore the learning hub.
← Back to the glossary