
How to Create 100 Ad Images in 30 Minutes (And Let AI Handle Everything Around Them)
Creating images for my ads took forever. And I say that as someone with a degree in graphic design.
I can design fast and make things look good... and it still took way too long.
For a recent launch I needed about a hundred ad images. The old way? That's not a task. That's a week. 😅
And ads aren't the only time I need a pile of graphics. It's the same story anytime I need a batch of them:
- Event promos. A promo image for every speaker at an event.
- Affiliate graphics. A folder of ready-to-share images for my affiliates.
- Launch graphics. Email and social images for every launch.
Deciding what to put on the images was never the hard part. The hours of design work were. And everything around them: naming files, catching typos, writing copy variations, and building it all out one image at a time.
The images were never the slow part. Everything around them was.
Why I built this
I was getting ready to run ads for a launch, and I knew the platform would want a lot of creative.
Variety is the whole game with ads right now. The more different images and hooks you give the platform, the better it gets at finding your people.
So I flipped the whole thing. Now AI runs every step around the images. It plans, it checks, it names, it writes the copy, and it designs the batch. I make the creative calls and approve the work.
Making graphics stopped being a project I dreaded and turned into a 30-minute task I can rerun for any launch, event, or affiliate push.
It's a lot cheaper to prevent a typo than to find one.
That's the part most people skip. They jump straight into an image tool, generate a pile of graphics, and then spend an hour squinting at every one of them looking for the year that's wrong or the word that's misspelled.
I do it the other way around. The rules get written before a single image exists. What the images can never say. The wrong-year traps, because AI image tools love putting last year's date on things. All decided up front.
And then once the images exist, AI checks every single file against those rules. Not me.
What actually changes when AI runs your graphics production
❌ Before
- ❌ One image and one caption at a timea whole afternoon for a handful of ads
- ❌ Hours of design work for every batcheven when you know exactly what you want
- ❌ Messy file names and typos slipping throughno way to find the broken ones
✅ After
- ✅ AI writes the prompts and the rulesabout a hundred images in one sitting
- ✅ Every file named, checked, and flaggedtypos caught before they run
- ✅ One process you can rerunfor any launch, event, or affiliate push
Where to start with your own batch of graphics
- One real batch to makePick the next thing you actually need graphics for: an ad campaign, an event, or your affiliates. A real project beats a hypothetical every time.
- An AI image tool you already pay forI used the one I already pay for, on the regular twenty dollar a month plan. Nothing fancy.
- A few graphics that have worked for you beforeYour past winners are the best brief you'll ever write.
- One folder for everythingEvery image lands in the same place so AI can check the whole batch at once.
Build your image production line
Copy this into your AI tool to plan the rules and prompts before you generate anything. Swap "ad images" for event promos or affiliate graphics if that's your batch.
I'm creating ad images for [offer or event]. First, write my image prompt template with 10 rules the images must follow: words that can never appear, wrong-year traps, logo and brand traps, and anything that would make the ad feel generic.
Then give me [10] distinct image prompt variations across different visual styles, a file naming convention that sorts by style family, and a proofreading checklist.
After I generate and save the images, I'll share the folder so you can check each one against the rules, rename every file to the convention, and tag any problem file with FIX in the name so I can find them by search.
The takeaway
Write the rules first, then have AI check every finished image against them and put one searchable word, like FIX, in the name of anything broken. A batch of graphics isn't a design project anymore. It's a production line.
The full breakdown is inside The Shortcut
The prompt above builds your image set. Members get the full process, including how I turned mine into a running ad campaign.
- The copy prompt that writes five versions of every ad, each with a different hook and the same offer
- The exact campaign structure I use, and why I stopped using interest targeting
- The tracking link format I put on every single ad so I know what to scale
- How I built a skill from my past winning ads so AI suggests new ones I'd never think of
- The tool that pushed my whole campaign out without building ads one at a time
- My audio walkthrough of the full strategy, plus an AI assistant that customizes it to your offer, event, or affiliate program
Let's be real about what AI graphics can and can't do
I'll be the first to say this doesn't replace a professional designer. As someone who spent tens of thousands of dollars to get a degree in design, I don't hand AI anything that needs true, creative thinking and strategy.
What it replaces is the repetitive part: the hundredth version of the same idea, resized, checked, and named. I still make every creative call. AI just handles the volume.
And AI can make a hundred images fast. It can't tell you which one will sell.
That's the platform's job. And it only works if you give it enough variety and enough data to learn from. A hundred images with no tracking is just a hundred guesses.
And AI will still make mistakes. Wrong dates, misspelled words, a logo that looks almost right. That's exactly why the rules list and the file check exist. You're not trusting it blindly. You're building the checks in.
Last thing. Right now, as I'm writing this, Claude has a new connector for Meta that can build your campaigns for you. I haven't tested it yet, so I can't recommend it. If you want to try it, go for it. 👀
Common questions
Can AI create ad images that are good enough to run?
Yes, as long as you give it rules first. AI image tools are fast, but they make predictable mistakes like wrong dates, misspelled words, and off-brand logos. Writing a rules list before you generate, then having AI check every finished file against it, is what makes the images safe to run.
How many ad images should I make for one campaign?
More than you think. Ad platforms learn by testing variety, so giving them many different images and hooks helps them find your audience faster. Batching your images in one sitting with AI makes high volume realistic instead of a week of work.
Do I need a paid AI tool to create ad images?
You need access to an AI image tool, but not an expensive one. I made mine on a standard twenty dollar a month plan, using the regular chat.
How do I find typos in AI-generated ad images?
Have AI proofread every image for you and tag any file with a problem using one searchable word in the file name, like FIX. Then you can pull up every broken image with a single search instead of scrolling through every thumbnail.
Should I use interest targeting for my ads?
Many small businesses now get better results going broad and letting their creative do the targeting. When you give the platform enough variety in images and hooks, it finds the right audience on its own. Whatever you choose, put tracking on every ad so you can see what actually brought in sales.
Recommended Resources 🛠️
The tools behind this shortcut.
- EverAds · pushed out the whole campaign
- Simple AI Pages · built the landing page the ads point to
- ChatGPT · generated the images
- Claude · ran the prompts, checks, and copy
Some of these are affiliate links. If you sign up through them I may earn a commission, at no extra cost to you.


