AI Creative Brief to Prompt Demo

An interactive prototype that turns a short creative brief into structured prompt components, generation constraints, tags, and review criteria — usable as a live demo for stakeholders.

Prompt quality varied widely between people. Briefs were freeform, prompts were inconsistent, and there was no link between what was requested, what was generated, and how it would be reviewed.

1. Freeform brief
   ↓
2. Ad-hoc prompt writing
   ↓
3. Inconsistent results
   ↓
4. No shared structure
   ↓
5. Manual review with no criteria

1. Structured brief input
   ↓
2. Component selection (theme / object / style / emotion)
   ↓
3. Constraint selection
   ↓
4. Prompt assembly
   ↓
5. Auto-suggested tags
   ↓
6. Suggested review criteria

The Prompt Builder demo on the Demos page is a working version of this flow.

• Designed the brief-to-prompt component model
• Built the prompt assembly logic
• Prototyped the interactive demo with vibe coding
• Connected prompts to tags and review criteria
• Validated output structure with stakeholders

1. 01Break prompts into named, reusable components
2. 02Keep constraints separate from creative intent
3. 03Generate tags from the same inputs as the prompt
4. 04Attach review criteria to every generated prompt
5. 05Make the demo runnable without a real AI API

• Coze
• Python logic
• Vibe Coding
• Prompt templates
• JSON data
• Figma

• Interactive Prompt Builder
• Structured prompt template
• Auto-generated tags
• Suggested review criteria
• Demo screens for stakeholders

• Turned freeform briefs into structured prompt components
• Made generation constraints explicit and reusable
• Linked prompts to tags and review criteria automatically
• Reduced prompt-writing variance across briefs
• Created a shareable, self-contained demo

Self-initiated demo — mock data, no company data.

The brief-to-prompt component model can be reused for any generation task that benefits from consistent inputs, including marketing assets, product imagery, and UI illustrations.

A good prompt system is less about clever wording and more about structure: consistent inputs produce consistent, reviewable, reusable outputs.