The problem
Every livestock auction needs its flyer: always the same structure, always different data. By hand that is about 35 minutes per piece. And generic AI produces things that do not look like the brand — which is why almost nobody uses it in production.
What we did
We built an app where the operator picks the client, enters the auction details, and gets the finished flyer. The key was not the model: we wrote a brand brief per client — exact colors, typefaces, tone, hierarchy, which logos belong — plus a set of reference artwork specific to each one, so the AI produces pieces recognizable as that brand. A breeders’ association does not communicate like an auction house, and the brief says so explicitly.
Every generated piece is logged to a spreadsheet for quality control. We also built a white-label version of the generator, ready for another agency to run with its own clients.
The outcome
From 35 minutes to under 5 per piece, with visual consistency held across three different brands.
Built with
- Next.js
- Image-editing AI
- Executable brand briefs
- Google Sheets
This case is part of Automated Operations
See the service
