Deployments should measure collisions, object drops, inference latency, data collection cost and downtime exposure alongside benchmark success.