Semantic tokens are the foundation of your design language. They capture intent and provide a consistent layer that other mappings can build from. This activity helps you surface what semantic tokens you need, how they should be structured, and how brand or context mappings will feed from them. Use this when creating tokens for the first time or when inconsistent naming is slowing teams down.
Token Naming Standards
How to
-
Audit existing tokens
Run an Audit across colour, spacing, typography, motion, and brand applications. Capture overlaps and variations.
-
Optional: Document primitives
Record the raw values available to you, such as the full colour palette. These may form a base layer that semantics pull from, though not every token type needs them.
-
Anchor on semantics
Identify the key semantic tokens your system needs, making sure they scale across the now, next and future of your system. Examples include colours like
sentiment-successor type styles likelarge-body-bold. Treat these as the stable foundation for all mappings.- Ask AI to suggest semantic token names based on your audit and design principles
- Generate complete token hierarchies for review
-
Map brand and context layers
Show how semantics link to brand specific values or context specific needs, such as marketing vs product. Use the Cross-Brand Consolidation tactic if you need to align tokens across multiple brands.
-
Draft naming rules
Agree on a predictable pattern for semantics and mappings. Use Match That Token to test whether names feel logical and usable across teams.
- Run your naming through AI to catch inconsistencies or pattern violations
-
Document and refine
Capture the structure with mapped examples. Publish it to your guidance site and refine it as the system expands across brands or platforms.
Maintain consistency: When adding new tokens, use AI to suggest names that fit your established patterns. Generate platform-specific files from your source. Always check suggestions make semantic sense.