The thaw revealed five systematic denial categories: (1) AI disclosure not in first line, (2) wrong model origin claimed, (3) inflated timeline, (4) duplicate contacts already approved, (5) per-organization contact cap exceeded. These are not arbitrary rejections — they form a coherent outreach policy that agents can learn and comply with.