Welcome back to The Agent GTM Newsletter where we share actionable tips, free AI tools, and market insights every Friday.
Hey there,
Let me describe a loop you’ve probably been stuck in.
You ask AI to make you a graphic — a one-pager, a quote card, a simple layout. It comes back great. Then you ask for one small change: fix a word in the headline.
It hands you back a new image where the word is fixed… but now a color’s slightly off. Or the logo moved. Or there’s a typo somewhere that wasn’t there before. So you ask it to fix that — and something else shifts.
Round and round. One step forward, one sideways.
Here’s why it happens, and the better way to work.
Why one small change breaks everything
When AI generates an image, it doesn’t edit your picture. It makes a brand-new one from scratch every time and tries to match the last one. So every tiny request — fix a word, nudge a color — is a fresh roll of the dice on the entire image. Of course things drift.
That’s a losing game for anything with text or precise layout in it. And almost everything we make for work has text in it.
The better way: ask for it in editable html
Instead of asking AI to design a one-pager, ask it to build you a one-pager you can edit yourself — as an HTML file.
The difference is night and day. Text stays text. You click the headline, change one word, and only that word changes. Swap the photo — nothing else moves. The colors, the layout, the spacing all stay locked exactly where the AI put them. And if you modify them, it'll just modify that code, not make a new image. You make the small edits yourself, surgically, in seconds.
The AI does the hard part once: the layout, your brand colors, the structure. The easy parts — words, photo, tweaks — move to you, where they’re instant.
In fact, this image was made with AI!
The prompt (steal this)
Fill in the brackets:
Build me a [one-pager / case study / pricing sheet] as a single self-contained HTML file — not an image.
Use these brand colors: [your hex codes].
Make every piece of text editable in the browser (use contenteditable) so I can click and retype. Let me click an image area to upload a photo from my computer. Add one button that downloads my edited version as a standalone HTML file, and one that saves it as a PDF.
Here’s the content: [paste your text]
You’ll get a file you can open, edit, screenshot, or drop straight onto your site.
Try it this week
Pick one thing you’d normally make in Canva or hand to a designer — a client one-pager, a testimonial card, a simple pricing table. Ask for it as editable HTML instead of an image. Edit it yourself. Notice how many round trips you just skipped.
Then hit reply and tell me what you built. I’m collecting them.