You're staring at a blank email draft. You need to send 50 of these this week. And you know that generic, templated garbage doesn't work - but writing 50 unique, personalized cold emails from scratch is going to take 10 hours you don't have.

So you turn to AI. ChatGPT, Claude, whatever. And then you get back something that reads like a robot wrote it - generic, salesy, and practically guaranteed to land in spam or the trash.

Here's the thing: AI is actually good at cold email. But only if you know how to use it. Most people use it wrong. They ask for a "cold email that gets replies" and get back template garbage. Instead, you need to use AI as a tool to speed up specific parts of your writing process - not to replace your thinking.

The Framework: AI as a Research and Drafting Tool, Not a Writer

The best way to use AI for cold email is to break the writing into stages. You handle the strategic part - what to say and why. AI handles the drafting - making it sound natural and getting words on the page faster.

Here's the actual workflow:

The key is that you're doing the hardest part - finding the angle and the research. AI is just helping you turn that into prose faster.

Step 1: Research First, Then AI

Before you touch AI, spend 3-5 minutes on the prospect. Look at their website, recent LinkedIn posts, or news about their company. Find one specific thing - a service they just added, a pain point you see in their industry, a recent hire.

This is non-negotiable. If you don't have a real angle, AI will give you generic filler, and it will fail.

Let's say you're reaching out to a marketing agency that just hired a new Chief Revenue Officer. That's your angle. Don't go into AI asking "write me a cold email to this marketing agency." Go in with: "This marketing agency just hired a CRO. I help agencies close more clients through cold email. Draft an email about how their new CRO might want to add cold email to their revenue stack."

Step 2: Use AI to Draft Around Your Angle

Now you give AI what it needs. Be specific about:

Here's an actual prompt that works:

I help service businesses get 5-10 new high-ticket clients per month using cold email. I'm reaching out to [Prospect Name], the founder of [Company], which does [what they do]. They're currently [specific situation - maybe they're hiring aggressively / just launched a new service / have been quiet on social for 6 months]. Write me a short 3-4 sentence cold email that mentions this specific situation and asks for a 15-minute call to talk about whether cold email could work for them. Keep it conversational, not salesy. Maximum 75 words.

Notice what this prompt does: it gives AI the constraints (short, conversational, specific), the context (what you do, who they are, why now), and the outcome (a specific ask). AI responds much better with boundaries.

When you run this, you'll get something like:

Hey [Name] - saw you recently brought on a new CMO. Most agencies we talk to are great at service delivery but leave revenue on the table. We help agencies like yours add cold email to their offer and close 5-10 new clients per month. Curious if it's on your radar? Open to a quick 15-min call?

This is a usable starting point. It's not perfect, but it's miles better than starting from blank.

Step 3: Edit Ruthlessly - Remove 40% of the Words

AI tends to pad. It adds extra words, hedging language, and unnecessary explanation. Your job is to cut it in half.

Take that draft and remove:

The raw AI draft above is already pretty good, but you'd still tighten it:

Saw you brought on a new CMO. Most agencies leave revenue on the table with their existing offers. We help agencies add cold email and close 5-10 new clients monthly. Worth a quick call to explore?

That's tighter, more direct, and reads more like a person wrote it.

Step 4: Add Your Personality Back In

After editing, read it out loud. Does it sound like you? If not, adjust a few words to match how you actually talk. Change "worth a quick call" to "worth 15 minutes" if that's how you normally speak. Move punctuation around. Swap one word for a synonym that feels more natural.

The final email should be 85% AI-drafted and edited, 15% your voice. Not 100% you (that takes forever), but not 100% AI either (that's generic).

Common Mistakes People Make With AI Cold Email

Asking AI to write without research: "Write me a cold email to a fitness studio owner." This gives you generic trash. AI needs specifics.

Keeping everything AI writes: If you don't edit, it sounds like every other AI-written email. Edit aggressively.

Over-personalizing: Some people ask AI to write different subject lines and openings for every prospect. This takes longer than just writing it yourself. Use AI to draft the body, then change the opening based on what you know about them.

Not setting word count limits: Tell AI "maximum 50 words" or "4 sentences max." Without constraints, it rambles.

The Reality Check

Using AI this way - as a drafting tool, not a writer - cuts your cold email writing time in half. You still need to do the research and the editing, but the middle part is faster.

If you're sending 100+ cold emails per week, you'll feel the time savings. If you're sending 10, it might not matter much.

And here's what AI won't do: it won't figure out who to email, it won't find the research angle, it won't manage your sending infrastructure, it won't handle replies. Writing cold emails that actually get replies still requires strategy and testing - AI just makes the drafting part faster.

Most service businesses see their real bottleneck isn't writing speed - it's everything else. The list-building, the email verification, the infrastructure setup, the follow-ups, understanding what's actually working. That's why booking 20 meetings a month with cold email is less about having faster AI drafts and more about having a complete, managed system.

Related Guides