
Design a brand-locked AI prompt system
You'll have a written set of AI system prompts and component libraries that force your design tool to output work matching your brand, not generic defaults.
What you'll make
Brand-Locked AI Design Prompt System
- Brand Constraints: the 5-7 non-negotiable visual rules
- Component Prompts: reusable blocks for layouts, colors, typography
- Override Triggers: when to reject AI output and what to fix
- Tool-Specific Adaptations: how to modify prompts for Midjourney vs Figma AI vs others
Write each prompt with its use case and paste examples of good vs bad output it produces
The 4 parts
Each part ends with a check you can see.
Audit Your Brand Breaking Points
Test AI design tools on your current brand and document exactly where they go wrong
Done when
You have 8-10 examples of AI design output with specific annotations of what's wrong and what should be different
Build Your Constraint Architecture
Convert your brand violations into a system of explicit rules that AI tools can follow
Done when
You have 5-7 concrete brand constraints written as AI-readable rules with before/after examples
Build Reusable Prompt Components
Create modular prompt blocks you can mix and match for different design contexts
Done when
You have 4-6 component prompts tested for different design needs (headers, layouts, color schemes, typography) that you can combine
Finalize Your Brand-Locked System
Compile your tested constraints and components into a complete prompt system with tool adaptations
Done when
You have a documented prompt system with examples that works across at least 2 different AI design tools
Capstone: you put together your Brand-Locked AI Design Prompt System.
Where this helps
An elective: a single lesson that sits outside rtwo's job certificates and doesn't count toward one.
About this lesson
What you'll be able to do
- Build a set of AI prompts that consistently output brand-specific designs, not generic variations
- Identify where your brand breaks down in AI-generated output and create override rules
- Have a working prompt library you can copy-paste into any design AI tool
- Know when to trust AI design output vs when your brand judgment must override it
Who it's for
- Designers working solo or on small teams who use AI design tools but get frustrated with off-brand output
- Brand managers who need to control AI-generated visual assets
- Anyone who's tried Midjourney or similar tools and gotten beautiful but wrong results for their brand
What it won't cover
- Building custom AI tools from scratch like Stripe did
- Advanced prompt engineering theory
- Brand strategy fundamentals — assumes you know what your brand should look like
- Production deployment or team sharing systems
Skills you'll practise
- Brand constraint documentation
- AI output evaluation against brand standards
- Iterative prompt refinement
- Component-based prompt architecture