Most AI outputs look impressive. But they still don’t feel like a real ad. That gap isMost AI outputs look impressive. But they still don’t feel like a real ad. That gap is
The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Most AI outputs look impressive.
But they still don’t feel like a real ad.
That gap is exactly what I set out to close with AUREL🚀
—
I didn’t approach this like prompting.
I approached it like building a creative pipeline.
Everything starts with structure.
I used the Compositor Node as the core—treating every visual like a layered campaign build:
• model as the primary subject
• product placed with intent
• controlled lighting and shadow passes
• grain and texture for depth
Nothing is random.
Every layer is composed to feel like it went through an actual art direction process.
That’s where the “ad feel” comes from.
—
Then I pushed the product beyond flat placement.
Inside the Creative Code Node, I brought in Three.js to generate a 3D version of the product.
This changed everything.
Now the product:
• has real depth
• reacts to light correctly
• sits naturally inside the frame
It’s not pasted in—it’s constructed into the scene.
—
From there, I designed the system to branch into four distinct campaign directions:
Experimented a bit today with Krea and image generation for a case study I’m putting together around AI EarPods connected to OpenAI.
The focus has been on creating fashion-forward product imagery and art directing a world that feels specific to the identity, rather than just generating nice-looking AI images.
The trickiest part has been product consistency. Especially getting the EarPods to actually sit snug in the ear. If you’ve worked through this process, you probably know the struggle 😅
Simply telling AI to “make it fit more snug or in the ear” doesn’t always work. It loves to reinterpret the product every time.
Still experimenting, but getting closer. If anyone has found a good workflow for keeping products consistent across AI-generated shoots, I’d love to hear it!
What happens when you take a hero frame from an AI-generated video and push it back into an image edit model?
This weekend I took a still from a recent Seedance 2.5 piece I made —depicting a solo mountaineer in a harsh alpine environment — and ran it through Midjourney's new v8.2 edit model to see how far I could iterate on a single frame without regenerating from scratch.
Tested wardrobe swaps, environment shifts, and style transfers.
Video walkthrough of the process coming in a separate post.
On the Bullseye World.
New visuals for EXOT Imaginary, my personal space to explore new ways of creating through AI, visual storytelling and experimentation.
Hope you like it! a.exot