Context provenance is the record of where each element of an AI system's working context came from — the source system, version, author, classification and retrieval decision — carried alongside the content so that outputs can be attributed and trusted.
What is context provenance?
Provenance answers "where did this come from", where lineage answers "what path did it take to reach the answer". Without provenance, a model treats an authoritative policy document and a stale draft in a personal drive as equally credible, and a reviewer has no way to tell them apart after the fact. Carrying provenance through assembly makes citation possible, allows policy to prefer authoritative sources, and gives investigators a starting point when an output is challenged.