Review policy

Human-review sampling rate: 20%

This case study models a marketing team generating short ad copy variations with an LLM at scale, roughly 1000 pieces per week. A small fraction of generated pieces contain a genuine brand-risk issue: a factual error, an off-brand claim, a tone mismatch. Human editors sample a percentage of output for review before publish.

Weekly outcome (1000 generated pieces)

Risk issues caught pre-publish—

Risk issues published unreviewed—

Editor hours spent reviewing—

Generated pieces (red = true brand-risk issue, ring = reviewed)