Incremental indexing updates the index only for added, modified, or deleted documents, by detecting changes (date, content fingerprint). It avoids recomputing everything at each update, which reduces costs and delays.
Example
hash = sha256(content) if hash != stored_hash: reindex the document