We Fed the Same Terrible Ad Copy to ChatGPT, Claude, and Gemini — Here's What Happened
We took a real piece of weak, generic ad copy — the kind that sounds like it was written by someone who's never met a customer — and dropped it into all three AI giants at the same time. The results were shockingly different. One model played it safe. One got creative but drifted off-brand. And one rewrote it so sharply that we nearly sent it straight to a client. By the end of this article, you'll know exactly which AI to reach for when your copy is flat, what prompt to use, and why the difference between these tools isn't just style — it's strategy.
The Original Ad Copy Was Genuinely Bad (Here's What We Tested)
Here's the ad we used as our test subject — a fake but painfully realistic example for a productivity app:
"Our app helps you get more done. With powerful features and an easy-to-use interface, you'll be more productive than ever. Try it today!"
If you've been in marketing longer than a week, you've seen this copy a hundred times. It says everything and nothing. No specific benefit. No emotional hook. No reason to click.
We gave all three models the exact same prompt: "Rewrite this ad copy to be more compelling, specific, and emotion-driven. Keep it under 50 words. The audience is busy professionals aged 25–40 who feel overwhelmed at work."
The prompt is everything here. Being specific about audience, word count, and tone forces the AI to make real choices — and that's where the differences between ChatGPT, Claude, and Gemini start to show.
How ChatGPT, Claude, and Gemini Each Rewrote the Ad — The Actual Outputs
ChatGPT (GPT-4o) came back with this:
"You have 47 tabs open and 12 unread Slacks. Sound familiar? [App Name] turns your chaos into a clean, prioritized workflow — so you actually finish what matters. 7-day free trial. No credit card."
Strong. Specific. The "47 tabs" detail is borrowed from real life and immediately relatable. ChatGPT leaned into specificity and conversational tone, which is its sweet spot for short-form copy.
Claude (Claude 3.5 Sonnet) returned this:
"You're not bad at your job. You're just drowning in it. [App Name] gives overwhelmed professionals a clear system to reclaim their day — and actually leave work at work. Start free today."
This one hit differently. Claude went emotional and empathetic — almost therapeutic. It reframed the problem as not the user's fault, which is a classic copywriting technique called negative reframing. For brands that want to build trust before selling, this is the smarter play.
Gemini (Gemini 1.5 Pro) produced:
"Stop drowning in to-do lists. [App Name] uses smart prioritization to help busy professionals cut through the noise and focus on what actually drives results. Less chaos. More impact. Try free."
Competent, clean, professional — but it felt like it could be for any productivity app. Gemini played it safe. The copy is polished but generic, which is the exact problem we were trying to solve in the first place.
The Hidden Difference Nobody Talks About: Voice Consistency vs. Creative Risk
Most comparisons stop at "which output sounded better." That's the wrong question. The real question is: which AI can match your brand voice while still taking creative risks?
Here's the deeper insight. ChatGPT is aggressive about engagement. It defaults to hooks, punchy rhythm, and pop-culture adjacency. That's great for DTC brands, apps, and social ads. But if you're writing for a law firm or a healthcare company, that tone will get you fired.
Claude defaults to empathy and nuance. It thinks about the reader's emotional state before it thinks about the sell. That makes it exceptional for copy where trust is the conversion mechanism — think financial services, wellness brands, coaching programs. Claude also tends to write sentences that sound like a thoughtful human wrote them, not a content machine.
Gemini is the most "professional" out of the box, which is both its strength and its weakness. Use it when you need copy that fits into a larger brand system and won't raise eyebrows in a review meeting. Don't use it when you need the copy to stop someone mid-scroll.
The mental model that unlocks all three: think of ChatGPT as your bold creative director, Claude as your empathetic brand strategist, and Gemini as your reliable account manager. You want all three in the room — you just need to know who to ask first.
How to Use This Comparison to Rewrite Your Own Ad Copy Today
You don't need a marketing degree or a big agency retainer. Here's the exact workflow to run this yourself in under 60 seconds.
Step 1: Pull out your weakest piece of ad copy. It could be a Facebook ad, a Google headline, a product description, or an email subject line. If you don't have one handy, use the test copy from this article.
Step 2: Open ChatGPT, Claude, and Gemini in three separate browser tabs. All three have free tiers — there's no excuse not to test this right now.
Step 3: Paste this prompt into all three: "Rewrite this ad copy to be more compelling, specific, and emotionally resonant. Keep it under 50 words. My audience is [describe your audience in one sentence]. Here's the original: [paste your copy]."
Step 4: Read the three outputs out loud. Not silently — out loud. Bad copy sounds wrong when you hear it. You'll feel immediately which one fits your brand's voice.
Step 5: Take the winner and run it through one more round. Paste the best output back into the model that produced it and add: "Now give me three variations — one more emotional, one more direct, one with a stronger CTA." You'll have six solid options inside two minutes.
This workflow costs you nothing and takes less time than writing one version yourself.
The Part Most People Get Wrong
Most people use these tools like a vending machine — put in a weak prompt, expect a great result, then complain that "AI copy sounds fake." That's not an AI problem. That's a prompt problem.
The single biggest mistake is not specifying the audience. Saying "rewrite this ad" gives the AI no target. It doesn't know if your customer is a burned-out startup founder or a retired teacher trying to organize their calendar. The copy it produces will be average because your instruction was average.
The second mistake is treating the first output as final. Every professional copywriter writes multiple drafts. AI gives you a draft in two seconds — that's not a reason to skip the iteration, it's a reason to iterate faster. The first output is a starting point, not a deliverable.
And the third mistake: using only one model. You now know that ChatGPT, Claude, and Gemini have different creative instincts. Running the same prompt through all three costs you 90 seconds and triples your options. The best copy you'll ever get from AI is usually the one that surprised you — and you can't be surprised if you only ask one question.
Key Takeaways
- ChatGPT (GPT-4o): Best for punchy, scroll-stopping copy with specific hooks — ideal for social ads and DTC brands.
- Claude (Claude 3.5 Sonnet): Best for emotionally resonant, trust-building copy — ideal for service businesses, wellness, and coaching.
- Gemini (1.5 Pro): Best for polished, professional copy that fits brand guidelines — use it when safety and consistency matter more than edge.
- Audience specificity: Always include your target audience in the prompt — it's the single biggest lever for copy quality.
- Iteration beats generation: The best AI copy comes from a second or third prompt, not the first output — build iteration into your workflow.
What to Do Right Now
Open ChatGPT, Claude, and Gemini in three tabs right now — it takes 60 seconds. Paste this prompt into all three: "Rewrite this ad copy to be more compelling and specific. Audience: busy professionals aged 25–40. Keep it under 50 words. Here's the original: [your copy here]." Read the three results out loud, pick your winner, and you'll have better copy before your next meeting starts.