We cross-check creators’ video metadata, transcripts and hashtags with audience signals - engagement patterns, view distributions, subscriber geography/language and recency - then flag anomalies (spikes, low reach, suspicious subscriber sources). Results surface as fraud scores and audience snapshots so you can verify reach and brand safety in minutes.