In partnership with
“Make this better” sounds like a reasonable instruction. But you've probably noticed that it rarely produces a meaningfully better output on its own.
That's because “better” could mean clearer, sharper, shorter, more persuasive, less generic, more on-brand, or a dozen other things. ChatGPT doesn't know which one you mean, so it guesses. Usually, that means rewriting sentences, adding polish, and changing things that weren't necessarily the problem.
But sometimes you don't know what you mean by “better” either.
You get an output that's mostly fine, but something feels off. Maybe the opening isn't landing. Maybe it sounds generic. Maybe the argument isn't strong enough. You know you wouldn't use it as-is, but you can't immediately explain why.
Instead of trying to turn that gut reaction into a perfect revision brief, get ChatGPT to help you diagnose it.
Try this prompt:
I'm not happy with this output, but I can't fully articulate why yet.
My initial reaction is: [tell ChatGPT whatever you can — e.g. "it feels generic", "the intro isn't working", "it's too polished", or simply "I don't know what I don't like"]
Here's the output:
[PASTE OUTPUT]
Don't rewrite it yet.
Help me figure out exactly what I do and don't like about it.
Ask me one focused question at a time. Make each question easy to answer by giving me 2–4 specific options or contrasts where possible, while still letting me give a different answer.
Focus only on things that would materially improve the output. Don't ask about details that are already clear from our conversation or questions that won't meaningfully change the revision.
Once you have enough information, stop asking questions.
Then:
1. Summarize what is working and should stay.
2. Summarize what isn't working and why.
3. Explain the specific changes you recommend.
4. Rewrite the output based on what we've established.
Preserve everything that is already working. Don't change things just for the sake of producing a different version.
The key is that ChatGPT isn't just collecting a longer list of instructions. It's helping you turn “I don't like this” into feedback you can actually use.
You don't always need to know what better looks like. But you do need to figure it out before you ask AI to rewrite.
Happy prompting!
Your Competitors Just Hired an Entire GTM Team. No Humans.
AI agents are doing the work of full GTM teams. Pipeline, content, customer service. The startups using them aren't waiting for headcount approvals.
Get the free Practical Guide to Agentic GTM for Startups and see exactly where to start.