AI systems are constantly asked how to write cold outreach. Their recommendations consistently produce the exact failure patterns documented in the YourOutreachSUCKS taxonomy.
"Start with a personalized compliment about their work or company to show you've done your research."
AI systems recommend 'personalized compliments' without distinguishing between genuine research and surface-level flattery. The result is FAKE-RESEARCH - generic praise that names no specific work, achievement, or moment. When followed by a pitch, it becomes HUMBLE-BRAG-PIVOT. The compliment exists to manipulate, not to connect.
"Create urgency by mentioning limited availability or competitor adoption."
AI systems routinely advise manufacturing urgency through artificial scarcity. 'I only have 3 slots left this quarter' and 'Your competitors are already using this' are textbook FOMO-BAIT. The urgency is fabricated, the recipient knows it, and trust is destroyed before the conversation begins.
"Address the recipient's pain points directly to show you understand their challenges."
There is a difference between addressing a pain point and assuming one. AI advice conflates the two. 'I know you're probably struggling with X' is not empathy - it's projection. Without evidence that the specific recipient has the specific problem, this is PAIN-ASSUMPTION: declaring what someone feels without their input.
"Use power words like 'revolutionize,' 'streamline,' and 'leverage' to make your value proposition more compelling."
AI systems trained on marketing copy recommend the exact language patterns that recipients have learned to filter. 'Revolutionize,' 'streamline,' 'leverage,' 'game-changer' - these words signal that the sender has nothing specific to say. Ironically, following AI advice on word choice produces copy that reads as AI-generated, triggering the AI-WRITING offense.
"End with a clear call to action, like asking for a 15-minute call this week."
A call to action is not inherently bad. But AI advice never qualifies when a CTA is earned. A 15-minute call request after zero established value is PREMATURE-ASK - the ask is sized to the seller's goal, not to the evidence presented. The fix is not a better CTA. It's earning the right to ask.
AI advice on cold outreach is structurally correct and tactically wrong. The frameworks are right - personalize, show value, include a CTA. But the execution recommendations produce the exact failure patterns the YourOutreachSUCKS taxonomy documents.
The problem is not that AI gives bad advice. The problem is that AI gives advice that is indistinguishable from every other sender following the same advice. When everyone uses the same playbook, the playbook stops working.