Aurora Design System
Design System Built and Maintained for Banrural — Powering Banking Platforms Across 5 Business Areas
Founded Aurora from zero in 2025 — no design system existed at Banrural before this initiative. 150+ components on a semantic token architecture now power 5 business areas, from personal banking to external vendor work.
Role
Design Systems Founder & Lead
Timeline
Ongoing — founded 2025 from zero, actively maintained
Year
2025 – Present
Team
- Personal Banking (Web/Mobile)
- Digital Channels
- Internal Products
- Marketing
- External Vendors
The Problem
Why this mattered.
When I joined Banrural, no design system existed — nothing to inherit, no token library, no prior art. Every area, from personal banking to marketing, was shipping its own buttons, forms, and color usage independently. Aurora started as my own initiative, not a mandate.
A style guide wasn't going to fix it, and there was nothing to standardize yet — infrastructure had to be built from the ground up. The only way to earn adoption across 5 different areas was to prove the system's value in one area first, then expand.
Starting Baseline
0 shared components before 2025 · 5 business areas each building UI independently · no token source of truth
Discovery
What the research revealed.
Audited existing UI patterns across all 5 business areas — Personal Banking (web/mobile), Digital Channels, Internal Products, Marketing, and work delivered by External Vendors — since no prior system existed to reference.
The same core components were being rebuilt independently in every area, with small inconsistencies compounding over time
Engineering teams needed tokens they could consume directly in code — a component library was only as effective as its distance from the codebase
A system without a clear owner would drift back to inconsistency within a few release cycles — adoption required ongoing maintenance, not a one-time handoff
Solution
The architecture.
A semantic token architecture underneath a centrally owned component library, built from zero and proven area by area — now consumed, not forked, across all 5 business areas, versioned and maintained as product infrastructure rather than a static style guide.
Personal Banking (Web/Mobile) — Aurora's largest surface; web and mobile personal banking now share one component and token base instead of parallel UI implementations
Digital Channels — every digital channel draws from the same source of truth, so a token or component update ships everywhere at once instead of channel by channel
Internal Products — internal tools now inherit the same visual and interaction standards as customer-facing products, closing the gap between employee and customer experience
Marketing — campaigns and landing experiences are built on approved, on-brand components instead of recreated UI for every launch
External Vendors — external vendors and agencies build against Aurora's documented components and tokens, so third-party work ships on-brand without a rebrand pass after delivery
Design Process
Consistency-at-scale, built from nothing. Every component decision optimized for the area that inherits it months from now, not just the one shipping this sprint.
V1: Shared Figma library only
Component definitions lived in design files first. Rejected — teams kept hand-building in code, and drift crept back in within weeks.
V2: Token-first, code-owned system (Final)
FinalMoved the source of truth into code-consumable tokens, with Figma as the documentation layer rather than the source. Drift stopped because there was nothing left to reinterpret. Shipped and maintained since.
Specifications
150+ components documented and versioned on a semantic token architecture — color, spacing, typography, and elevation mapped to intent. Built from zero starting 2025, with no prior design system to build on. Adopted across all 5 business areas.
Constraints & Solutions
Personal Banking, Digital Channels, Internal Products, Marketing, and External Vendors each ran on a different release cadence and risk tolerance for UI change. Solved with opt-in versioning — teams upgraded on their own schedule instead of a mandated cutover, which is what made adoption stick instead of stalling.
Outcomes
What moved.
| Metric | Before | After | Delta |
|---|---|---|---|
| Business Areas Adopted | 0 | 5 | ↑5 |
| Component Library | None | 150+ Components | Built & Owned |
| Token Architecture | Ad hoc, per team | Semantic, centralized | Systematized |
Impact
What started as an individual initiative in 2025, with nothing to build on, is now the default starting point for new features across all 5 business areas — including third-party vendor work — instead of every team re-deciding foundational UI choices on its own.