Skip to content
Workflow Lab

Explore interface experiments

Independent experiments in structured inputs, scoring interfaces and asset filtering. They use local simulated logic and example data; they do not reproduce the Magic Gifts process or its later workflow tool.

Choose a step to see its input and output.

BriefStep 01 / 08

Purpose

Capture the creative request in a structured, consistent format.

Input

Business goal, theme, audience, platform, deadline.

Output

A structured brief with clear scope and constraints.

Why it matters

A structured brief prevents ambiguity from flowing downstream into generation.

1 / 8

Two ways to organize the work

A conceptual comparison of ad-hoc and structured approaches. This diagram is not a measured production before-and-after result.

Ad-hoc

  • Open-ended requests
  • One-off prompt trials
  • Unrecorded review reasoning
  • Scattered examples

Structured

Make inputs, references and review decisions available to the next step.

  • Explicit creative inputs
  • Reusable reference patterns
  • Defined review criteria
  • Organized example library
Demos

Prompt Builder, Quality Scoring & Asset Filter

Three independent interface experiments. Local rules and example data drive the interactions; no live AI generation or image evaluation takes place.

Step 1

Prompt Builder

Select inputs and watch a structured prompt assemble in real time. The same inputs also suggest tags and review criteria. Press Simulate result to create example scores and an entry in the mock asset library below. No image is generated or evaluated.

Theme
Object
Style
Emotion
Constraint
Output
Structured prompt

Create a soft 3d themed gift concept for a valentine campaign, featuring a heart balloon with a intimacy mood. Use a centered composition with rounded three-dimensional forms, soft materials, and diffused lighting, no text.

Suggested tags
Soft 3DValentineIntimacyThemed Gift Concept
Suggested review criteria
  • Aesthetic fit to the target style
  • Style consistency across variations
  • Production usability
  • Business relevance to the campaign
Step 2

Quality Scoring Panel

Explore five example criteria through simulated scores. These values are an interface exercise, not measurements of image quality or the review method used in Magic Gifts. Showing example values. Try another set of options in the Prompt Builder.

Aesthetic Fit

0/20

Whether the generated result matches the target visual style and quality expectations.

Style Consistency

0/20

Whether the output stays consistent with the reference style across variations.

Production Usability

0/20

Whether the asset is usable in production without heavy manual rework.

Platform Safety

0/20

Whether the output avoids sensitive, misleading, copyrighted, or inappropriate elements.

Business Relevance

0/20

Whether the asset serves the business goal and the target scenario.

Example total

0/100

Simulated scores illustrate five review dimensions. Human judgment determines creative quality.

Adjust prompt ↑View asset library ↓

Step 3

Generated Asset Filter

Filter example assets by assigned style, risk, quality and use-case labels. These labels are mock data, not automated assessments of the images.

Style
Risk
Quality
Use Case

9 assets / 9 total

Cute

Heart Balloon — Soft 3D

88
Themed GiftHighSafe

Ready for gift catalog

Festival

Magic Box — Glossy

84
Themed GiftHighSafe

Use for festival event

Luxury

Lucky Charm — Premium

72
Campaign AssetMediumNeeds Review

Refine before campaign use

Cute

Cute Animal — Toy-like

86
StickerHighSafe

Ready for sticker pack

Game-like

Game Crystal — Game-like

69
IconMediumNeeds Review

Check icon legibility at small size

Flat

Confetti Burst — Flat

74
Campaign AssetMediumSafe

Good for background element

The production story

Follow one complete design decision

Read how reusable creative knowledge shaped batch generation and human selection in Magic Gifts. These Lab experiments are separate from that production work and its later tools.

Read the Magic Gifts case →