Project Goal

Create a visual proof of concept: a stylized dance routine rendered in the glowing, neon-infused aesthetic of Just Dance, with consistent character design, step-by-step choreography, and audience-ready video sequences.

Objectives

• Experiment with AI-driven choreography visualization.

• Design consistent male + female cowboy avatars styled for dance games.

• Use prompt engineering (JSON + image prompts) to maintain continuity across moves and scenes.

• Deliver portfolio-ready media showing how AI can accelerate motion design, music visualization, and branding concepts.

Tools & Workflow

1. Choreography Breakdown

• Researched the dance steps for Boots on the Ground and split the choreography into 4 modular parts (diagonal steps, forward/back with claps, rock & turn, walk & shake).

• Documented moves step-by-step for accurate translation into JSON prompts.

2. Prompt Engineering with ChatGPT

• Built structured JSON prompts for Veo/Flow that defined:

  • Master Character Block → consistent cowboy/cowgirl avatars (hat, western shirt, jeans, boots).

  • Stage Style → neon purple/teal/pink glowing dance floor.

  • Camera Rules → locked, full-body medium-wide framing.

• Ensured modularity: each segment could be generated individually, then stitched together.

3. Visual Development

• Created Just Dance–style sample visuals using image generation tools.

• Designed avatars with glowing outlines and bold, readable poses.

• Iterated female vs. male versions (female with folding fan prop, male with freehand dance moves).

4. Video Assembly in Google Labs Flow (Veo 3)

• Generated clips in 8–12s chunks for reliability.

• Used overlay text (“Step, Heel, Touch,” “Pop the Fan!”) for clarity.

• Combined clips into a full 40s prototype routine.

Final Deliverables

• Male Routine (Boots on the Ground) → 4 JSON-driven segments.

• Female Routine (Where Them Fans At?) → 2 JSON-driven segments with fan prop.

• Dual Dancer Routine → Cowboy + Cowgirl performing side by side.

• Sample Visuals → Just Dance–style screenshots for portfolio presentation.

Key Learnings

• Consistency is key: Defining a Master Character Block kept avatars stable across multiple prompts.

• Chunking choreography works best: Generating short clips reduced render errors and saved credits.

• AI can speed up prototyping: Within hours, a full dance game routine could be visualized for pitching or testing.

• Crossover opportunity: This workflow could extend to music videos, social content, and branded motion assets.

Outcome

Boots on the Ground demonstrates how AI workflows can reimagine music, dance, and entertainment prototyping. What might normally require motion capture, 3D modeling, and a game engine was instead visualized rapidly with AI, combining design direction, choreography planning, and generative tools.

For my portfolio, it showcases:

• Design versatility (branding + motion + entertainment design).

• AI prompt engineering (JSON, consistency blocks, modular pipelines).

• Creative innovation — using tools like ChatGPT + Google Labs Flow to push beyond static design into interactive experiences.