I Gave ChatGPT, Claude, and Gemini the Same Cold Email Brief — The Results Were Embarrassing for Two of Them
Most marketers assume all AI tools write "pretty much the same." I used to think that too — until I ran a side-by-side test that changed how I use AI for outreach completely. I gave ChatGPT, Claude, and Gemini the exact same cold email brief: same target audience, same product, same goal. What came back exposed a gap that's costing people real reply rates. By the end of this article, you'll know exactly which tool to use for cold email, why the other two fall short in specific, fixable ways, and how to prompt each one to get the best possible result.
The Brief I Used — and What Each AI Did With It
The brief was simple and realistic. Here it is word for word:
"Write a cold email to a marketing director at a mid-sized SaaS company. The goal is to book a 15-minute call. We sell an AI-powered analytics tool that reduces reporting time by 40%. Tone: professional but conversational. Length: under 150 words. Include a specific subject line."
That's the kind of brief a real marketer would write. Specific enough to test the AI's judgment. Open enough to see what each tool prioritizes.
ChatGPT (GPT-4o) came back with a clean, punchy email. Strong subject line: "Cut your reporting time by 40% — worth 15 minutes?" The body opened with a pain point, delivered the value prop fast, and ended with a low-friction CTA. It felt like it was written by someone who's actually sent cold emails before.
Claude (claude-sonnet) wrote something genuinely impressive — arguably the most human-sounding of the three. It avoided buzzwords, used a softer opener, and felt like it came from a real person. But it ran 172 words. It missed the brief's only hard constraint.
Gemini played it safe. The subject line was forgettable ("Improve Your Reporting Efficiency"), the opener was generic ("I hope this email finds you well" — yes, really), and the overall tone read like a press release. It technically answered the brief. It just wouldn't get replies.
The Hidden Difference: How Each Tool Understands "Persuasion"
This is the layer most comparison articles skip entirely. It's not about word count or tone — it's about whether the AI understands why a cold email works.
Persuasion in cold email is about one thing: making the reader feel like you already understand their problem. The best cold emails don't sell a product. They sell a moment of recognition — "this person gets what I deal with every day."
ChatGPT consistently nails this because it's been trained on enormous amounts of sales and marketing content. It intuitively front-loads the reader's pain before mentioning the product. In the test email, it wrote: "If your team still spends hours pulling reports that are outdated by the time they're done — I built something for that." That's not a feature pitch. That's a mirror.
Claude understands persuasion at a sophisticated level too — sometimes too sophisticated. It tends toward nuance and empathy, which is powerful in longer-form content. But cold email is a sprint, not a novel. Claude's instinct is to build rapport gradually, which works beautifully in a follow-up sequence but costs you the opening email.
Gemini treats cold email like an information transfer problem. It asks itself: "What does the reader need to know?" That's the wrong question. The right question is: "What does the reader need to feel?" Until Gemini solves for emotion before information, it'll keep producing emails that are technically correct and practically useless.
How to Get the Best Cold Email From Each Tool Today
You don't have to just pick one and hope for the best. Here's how to use each tool in a way that plays to its strengths.
Step 1: Use ChatGPT for your first draft. Start with this prompt: "Write a cold email to [job title] at a [company type]. Goal: [specific CTA]. We solve [specific problem] by [specific mechanism]. Tone: direct and conversational. Max 120 words. Write 3 subject line options." GPT-4o will give you something 80% of the way there on the first try.
Step 2: Run your draft through Claude to humanize it. Paste the ChatGPT draft into Claude and use this prompt: "Make this cold email sound less AI-generated. Keep it under 130 words. The tone should feel like it was written by a sharp, thoughtful person — not a marketing team. Don't add fluff. Cut anything that sounds like a brochure." Claude's editing instinct is sharper than its cold-drafting instinct.
Step 3: Test your subject lines with Gemini. Gemini is actually solid at generating variations and doing structured comparisons. Use it like this: "Here are 3 subject lines for a cold email. Rank them by likely open rate and explain why in one sentence each." Then paste your options. Gemini's analytical framing works well here because you're asking for logic, not emotion.
Step 4: Finalize and personalize manually. Add one specific detail about the company or person you're emailing — a recent product launch, a LinkedIn post, a funding round. No AI tool does this well for you. That detail is your secret weapon, and it takes 60 seconds to add.
The Part Most People Get Wrong
Most people treat AI cold email output as a finished product. They copy, paste, and send. That's wrong, and it's why their reply rates are stuck at 1-2%.
AI tools write for the average reader. Cold email converts when it speaks to one specific person. Those two things are fundamentally in conflict. No matter how good your prompt is, the AI doesn't know that your prospect just missed their Q3 targets or that their company is hiring three new salespeople. You do.
The other mistake is using a single AI tool for everything. The test proved that each tool has a different superpower. ChatGPT drafts, Claude refines, Gemini analyzes. Using just one is like hiring a writer who's also supposed to be your editor and your data analyst. The work suffers.
Finally, most people write prompts that are too vague. "Write me a cold email" is not a brief — it's a guess. The more specific your input, the better every AI's output becomes. Specificity is the skill that separates people who get mediocre AI results from people who get great ones.
Key Takeaways
- ChatGPT for first drafts: It instinctively leads with pain points and writes punchy, persuasion-forward copy that's ready to work with immediately.
- Claude for refinement: Its strength is making writing sound genuinely human — use it to edit, not to draft cold emails from scratch.
- Gemini for analysis: It's better at structured comparison and ranking than at emotional, persuasive writing — use it to evaluate, not create.
- Specificity beats prompts: The quality of your AI output is directly proportional to how specific your brief is — vague in, generic out.
- AI drafts, you personalize: No AI tool can add the one specific detail that makes a cold email feel personal — that 60-second step is yours to own.
What to Do Right Now
Open ChatGPT and paste this prompt: "Write a cold email to a [specific job title] at a [specific industry] company. Goal: book a 15-minute call. We help [target customer] solve [specific problem] by [mechanism]. Tone: direct and conversational. Max 120 words. Give me 3 subject line options." Fill in your actual details, get your draft, then take it to Claude for a humanization pass. The whole process takes under 10 minutes — and the output will be sharper than anything you'd write starting from a blank page.