By pointing out exactly what they consider to be not novel, e.g., the main result appears to follow as a special case of this result of XYZ. If brought up earlier, maybe the reviewers and authors could jointly figure this out. But at the last minute, the AC's should do it alone