Role: AI Creative Technologist

This AI-assisted UGC video project explores a node-based generative production workflow built in Figma Weave, combining AI-generated character, environment, product references, image generation, and image-to-video generation within a connected visual pipeline.

I developed a consistent AI character using multiple reference views and reused those references throughout the workflow to support character identity consistency across generated shots. I also created the AI environment separately and used it as a recurring visual reference to maintain environmental, lighting, and compositional continuity throughout the video.

The workflow combines reference-driven generation, structured prompting, character consistency, environment consistency, product preservation, iterative image generation, and image-to-video generation. Multiple character and environment reference images were connected through the node-based workflow and reused across shots to reduce visual drift and maintain continuity.

The project demonstrates how reusable visual references and modular node-based workflows can support scalable AI-assisted UGC and product-content production.

Tools / Workflow: Figma Weave • Generative AI • Prompt Engineering • Reference-Driven Generation • Character Consistency • Environment Consistency • Image Generation • Image-to-Video • Node-Based Workflows

Final UGC Video

AI Character Reference

AI Environment Reference

The AI-generated character and environment were established as reusable visual references and carried throughout the node-based workflow to support identity, environmental, and visual consistency across the generated video.

The Figma Weave graph shows the connected production pipeline used to develop the project. Character references, environment references, product assets, prompts, and generated outputs were organized as reusable nodes and routed into multiple generation stages. This allowed visual references to be reused across shots while supporting iterative prompting, controlled generation, and continuity throughout the sequence. Rather than generating each shot independently, the workflow was structured as a reusable system. Character, environment, and product references were maintained across generation branches to reduce visual drift and support a more consistent AI-assisted production pipeline.