The largest group of cold emails sent by AI companies pitches automation. It is also the weakest, replying at 3.5%. Emails that frame the same work as digitalization, modernising how a process runs rather than removing the people from it, reply at 4.3%.
That is not a copywriting trick. Automation tells a buyer that something they own is about to be taken away and that headcount questions are coming. Digitalization tells them the process gets better and they still run it. Same project, different thing to say yes to.
Emails offering a purpose-built agent for one specific workflow reply at 3.9%. They beat generic automation because they name a real queue or process, and they lose to digitalization framing because an agent still reads as a replacement.
The pattern across all three is the same. The more the email describes a category of technology, the worse it does. The more it describes one process the reader recognises, the better.
Cut the word automation from your subject line and your opening sentence. Replace the category description with the single process you would fix first, named the way the reader would name it. If you cannot name that process before you write, you are not ready to email that account.
Hi [First name],
Most operations teams your size rebuild the same few reports by hand every month, then lose the first week of the quarter reconciling them.
We turn that into one pipeline that updates itself, usually in about four weeks, and the team carries on working the way they already do.
Worth a look at what yours would take?
Hi [First name],
Your support queue looks like it needs a first-pass triage before anyone useful sees it.
We build agents that do that first pass and hand off anything that needs a person, so the queue arrives sorted rather than raw.
Want to see one running on a queue like yours?
Hi [First name],
We help companies automate manual processes using AI, from data entry through reporting to customer support.
Our clients typically save twenty hours a week and cut operating costs significantly.
Are you open to a quick call to explore how AI automation could help your business?
Join us to elevate your outreach!
Because it describes a category rather than a problem. Every AI company opens that way, so the reader has seen the same sentence many times that week. It also raises the headcount question in the first line, which is the fastest way to get an email forwarded to nobody.
Not in a first email. The buyer deciding whether to reply is almost never the person who cares which model you use, and naming one invites a technical objection before you have a conversation. Save it for the call.
Technical enough to prove you understand their process, not technical enough to describe your architecture. One accurate sentence about how their workflow currently runs beats a paragraph about your stack.
Only if the number is theirs. A generic claim about saving twenty hours a week reads as marketing, since the reader knows you have not measured their process. Asking what the current process costs them works better than telling them.
