AI Cold Email

Also known as: AI-powered cold email, personalized cold email

AI cold email is cold outbound where the email body is written per-lead by a language model that has been given lead-specific research as input — recent posts, job postings, GitHub commits, podcast appearances. The output is not a template with merge tags; it is a distinct email per recipient that references something concrete and connects to the sender's offer.

The category did not exist three years ago because LLMs could not write business-grade copy at scale. Today the question is not whether the model can write the email — it can — but whether the research pipeline gives the model enough signal to write something worth sending. Bad research produces bad emails even with a great model.

AI cold email is not magic; it is replacing template-plus-merge-tag with research-plus-writing per lead. The reply-rate lift on tight ICPs is 3-5x. On loose ICPs the lift is smaller because the model has nothing concrete to anchor on.