Grounding AI Outputs

Use when an LLM's output feeds a decision of record (compliance, audit, security mapping, requirements, status), when synthesizing specs from notes, or when a model might invent details the source never contained.