Grounding
Grounding gives an AI model trusted source material and instructs it to base its answer on that evidence. It reduces unsupported claims but does not eliminate model error.
For a docs chatbot, grounding means the model answers from retrieved documentation passages rather than from its training data. The retrieved text is placed in the model's context with instructions to answer only from it.
Grounded systems should link to their source material and decline questions that the available evidence does not support. An assistant that never declines is more likely to answer beyond the retrieved evidence.
Grounding quality depends on retrieval quality. If chunking or indexing misses the relevant section, the model either refuses or answers from a partially relevant passage. Most grounding work is really retrieval work.
Frequently asked questions
How can I tell if a chatbot is grounded?
Ask it something your docs do not cover. Check whether it declines unsupported questions, and verify that citations support the claims in its answers. A single refusal or a source link does not prove that all answers are grounded.
Is grounding the same as RAG?
No. Grounding means basing an answer on supporting sources. RAG is one way to do that by retrieving source content before generating the answer.