01Design Systems · Platform Strategy · AI-Enabled Delivery
Built a token-based enterprise design system and design-to-engineering operating model used across 12+ squads and six domains, reducing design-to-build time by 80%, rework by 25% and avoiding more than $9M in cost.
- Outcome
- 80% faster design-to-build cycles
- Outcome
- 25% reduction in rework
- Outcome
- $9M+ cost avoidance, finance-confirmed
Challenge
Fragmented patterns, duplicated components and inconsistent implementation were slowing delivery and increasing risk across digital products.
Approach
As executive leader for NAB's design-systems capability, I set Elevate's experience strategy, secured investment, established the operating model and governed adoption across product domains. A specialist design and engineering team developed the token architecture, component library and supporting tooling under this direction. Connected Figma, Storybook, design tokens, Figma MCP, Cursor and Copilot to support faster design-to-engineering workflows, while retaining human review and existing governance controls.
Impact
- 80% faster design-to-build cycles
- 25% reduction in rework
- $9M+ cost avoidance, finance-confirmed
- 12+ squads across six domains
- 60% reduction in critical accessibility defects
Key decisions
- Chose design tokens as the shared source of truth rather than creating another static component library.
- Introduced contribution governance without centralising every product decision.
- Embedded accessibility into release controls rather than relying on post-build audits.
- Connected design and code tooling while retaining human review and engineering accountability.
AI-assisted design-to-engineering
We treated AI-assisted design-to-engineering as a deliberate operating practice, not a tooling experiment. Using Figma's MCP integration, Cursor and GitHub Copilot, the design and engineering teams generated component scaffolding, token updates and documentation from design files, which shortened the design-to-code handoff across the 12+ squads. Human review and existing governance controls remained in place to preserve quality and accountability.

