How to Use AI In Your Emails Without Losing Your Voice

James Milsom -

In early 2024, I got a cold email that didn’t feel cold. It referenced a podcast I’d been on, shared a sharp idea, and felt genuinely human. I replied.

Today, more people are trying to write emails like that; founders looking for feedback, marketers promoting content, sales teams starting conversations. But something’s shifted. The emails are often polished but flat. Lifeless. Like this:

“I hope this email finds you well. I wanted to reach out because I believe there may be synergy between our organizations.”

Not wrong, but no pulse. And with AI everywhere, it’s a chance to stand out.

This post isn’t about avoiding AI. It’s about how to use it without losing your voice.

Summary

Treat cold email like a targeted ad: Think of it like a Google display ad - clear targeting, a relevant message, and a measurable outcome.

Start with your real voice: Write how you naturally explain your product to someone. Use that as your tone reference before you start prompting.

Understand your audience’s job to be done: Take the time to understand who you’re emailing, what they care about, and why you might be able to help.

Prompt AI with context, not just commands: Don’t ask for “a cold email.” Ask for one a message that sounds like you, written for a specific type of person, with a clear reason for sending it.

Edit AI output like it matters: Cut the generic, add specificity, and make sure the message still sounds like you, not a copywriter bot.

Use cold email like a live test: Just as with ads, observe who opens, replies to, or ignores you. Refine based on real feedback, not assumptions.

Be consistent with your tone across platforms: Your email should feel like it came from the same person who wrote your LinkedIn posts or spoke on that podcast. Otherwise, trust breaks.

It’s okay to use AI - just don’t delegate all of it: AI can help you get started, but the thinking and editing still need to come from you.

Why this matters

If you’re sending an email to someone who doesn’t know you, trust is everything. It's built in the first few seconds. Before they read your offer, before they click a link, before they even decide to scroll, they’re asking themselves:

  • Does this sound like a real person?
  • Do I believe this person knows what I care about?
  • Do I want to hear more?

When you rely on generic AI, the answers to those questions are all "no."

But here’s the good news: according to Hunter’s State of Cold Email 2025 report, two-thirds of decision makers didn’t mind if AI was used to help write a cold email, as long as the email still felt human.

That’s the distinction. It’s not whether you use AI, but whether your email still sounds like it came from you. Here's how you do that.

A cold email is an ad in a different format

I like to think of cold email the same way we think about Google Ads. A good ad does three things:

1) It targets the right people based on intent

2) speaks directly to their goals or pain points, and 

3) shows quickly whether it worked - bounce, click, convert.

Cold email should follow the same rules. You know who you’re reaching, what they need, and where other solutions fall short. You write a message that lands in their inbox instead of a webpage, and you watch what happens next - opens, replies, or silence.

Where it breaks down is when AI gets involved without direction. No targeting. No tone. No specificity. Just words for everyone, instead of someone.

The issue isn’t using AI. It’s using it without care. When your email reads like generic copy, not something you wrote, the trust breaks.

Your voice already exists in your LinkedIn posts, your blog, maybe a podcast. If your email doesn’t match that, people notice. It feels off.

That’s why prompting and editing matter. AI can help you write faster - but only if you stay in control of who you’re writing to, and how you sound.

Robotic vs. real messaging example

Let’s look at what happens when tone gets lost.

Our email on the left is a generic ChatGPT output - lifeless, generic, and without a clear ask. On the right we've tweaked it reflect my tone of voice, and controlled our ask of the reader.

The "Real" email has a heartbeat. It references something specific. It’s short, human, and direct. It makes you want to read on, even if just out of curiosity.

4 steps to adding your voice into AI emails

Step 1: Get clear on your real voice

Before you write anything, or ask ChatGPT to write it for you, pause and ask:

“What do I actually sound like when I’m explaining what I do?”

You’re not writing a pitch deck. You’re writing like you would if you bumped into someone at a coffee shop and they asked what you’re working on.

A simple exercise I use:

Write three sentences as if you were texting a friend about the thing you’re emailing someone about. That’s your tone baseline. Honest. Unforced. No fluff.

Step 2: Give AI something real to work with

If you’re using a tool like ChatGPT, the difference between a decent output and a lifeless one comes down to what you ask it.

Here’s a prompt I’ve used:

Write a short, conversational email from a founder to another founder. I like to be warm in how I write, but not overly familiar.

The goal is to see if they’d be open to a quick chat about a product that helps them find the right person to email at a company.

They want this solution because they’re always looking for more users, potential partners, and even sourcing funding.

The key with my solution is that I have verified data and an AI writing assistant that will keep their emails out of spam - which is becoming a massive issue in 2025.

Keep it under 100 words.

Sound like something someone would actually send.

And here’s what came back:

That’s usable. Not finished, but a good start.

Step 3: Edit like it matters

Don’t send the first version. You shouldn’t do that when you write something yourself, and the same goes here.

When I edit AI-generated emails, I always ask:

  • Does this feel like something I’d actually say?
  • Can I make it more specific?
  • Am I repeating phrases that don’t add value?
  • Are words randomly written in bold copy?
  • Are em dashes (—) overused?

Small edits go a long way. Replace “synergies” with real reasons. Cut filler. Add context. Read it out loud and see if it still feels natural.

Step 4: Build a muscle, not a script

The more you write, the more you start to hear the difference between your real voice and a generic one. Save the emails that feel like you. Start a folder. Over time, you’ll have a reference bank you can go back to whenever you’re stuck.

You’ll also start seeing what makes people reply. What opening lines land. What formats work best for your audience. That’s not something AI can figure out for you. That’s yours to learn.

Common mistakes to avoid with GenAI cold email copy

  1. Copying AI outputs without editing: AI is a first draft. Nothing more.
  2. Using vague prompts: The more you give, the more you get. Be specific.
  3. Over-formal tone: Most cold emails fail because they feel too professional. Write like a person.
  4. No clear reason for the email: Why now? Why you? Why them? That should be obvious within the first few lines.

Final thought: it's all about the reply

Even with the best data, replies only come when your message lands.

What makes it land isn’t just timing or targeting. It’s whether someone reads your email and thinks:

“This feels real. This feels like someone worth replying to.”

You don’t need perfect grammar, or a clever subject line, or even a polished deck. You just need to sound like you.

That cold email I got back in 2024? The one that didn’t feel cold at all? We ended up working together. That’s the power of an email that sounds like a person, not a machine.

If you found any of this interesting, make sure you give Hunter’s AI Writing Assistant a try. We’re building it precisely to help you create great emails, personal to each of your leads, using your tone of voice - but at scale.