AI Creative System

ROLE

Creative direction and automation: brand documentation in Obsidian, Claude Code + Remotion + After Effects for paid media reels, Codex + Gemini Image for carousels and product editorial, and art direction in Gemini Flow and Higgsfield for generative video.

YEAR

2026

DISCIPLINE

IA GENERATIVA

CONTEXT

AI creative system applied to real pieces: brand documentation, visual exploration, automated motion, generative video, and multi-format adaptation. This is not a tool test; it is a production method where AI helps reach an editable base faster and human judgment decides what works, what needs adjustment, and what should be discarded.

THE CHALLENGE

The challenge was not using AI for novelty, but integrating it without breaking each brand identity. Every flow keeps a human decision point: reading guidelines before generating, selecting cuts, directing acting in video, retouching lighting, and deciding when stock works better than AI backgrounds so the result feels more natural.

PROCESS

The system is divided into five flows, the same ones documented on the AI page: brand documentation, motion, generative image, generative video, and static design. For motion, the brief goes into Claude Code and becomes a Remotion composition with timing and cut structure; I review, select, and finish the deliverable in After Effects. For static pieces, Codex and Gemini help explore visual components from the brief. For generative video, Gemini Flow and Higgsfield are used with art direction from the initial prompt to acting corrections.

RESULT

A documented workflow that reduces repetitive work and speeds up the first editable base without removing creative review. It includes reels produced with the /reel skill, generative video pieces with acting direction, editorial posts created with generative imagery, and a World Cup 2026 flyer for KOP adapted into 5 formats with a Photoshop script. To see the full system, the AI page shows the working logic behind these cases.