G-P Design System: Zero to 100% and Beyond

Building a scalable, accessible, internationalized design system for 50+ apps across 188 countries.

G-P needed a scalable, accessible, and internationalized design system to support a rapidly growing microfrontend ecosystem across 50+ applications in 188 countries. The existing environment lacked standards for accessibility, internationalisation, and governance, creating bottlenecks and inconsistencies.

Along with a UX design partner, we gathered the corporate support needed to drive the initiative forward. I then led the architecture, development, and scaling of the design system, prioritizing accessibility and internationalization from day one to ensure global adoption and compliance.

While certain details are confidential, the case study highlights the strategic approach, key results, and learnings from this transformative project. In particular, I can say unequivocally that certain design choices we made up front were instrumental in our success and the long-term health of the system, and that AI augmentation has been a total game changer for both contribution velocity and governance efficiency.

  • Architected a design system ecosystem for microfrontend delivery, isolation, communication, and customization.
  • Prioritized ASA and EAA compliance — accessibility as a first-class concern.
  • Implemented internationalization support for global operations.
  • Centralized styling and theming using pure (modern) CSS for maximum compatibility.
  • Built lightweight, modular components optimized for microfrontend integration.
  • Collaborated with designers on new branding while maintaining legacy support.
  • Scaled from solo to a team of 6 engineering contractors.
  • Established contribution processes, guardrails, and governance structures.

50+

Apps supported

Weekly

Consumer team contributions

AI-augmented

Governance & DLC

Early investment in accessibility and internationalization pays dividends in global adoption.

This was a critical factor in the success and widespread adoption of the system, and is often overlooked in design system initiatives.

While it may require more upfront effort, I believed that it was essential that the smallest components be the most compliant and that accessibility concerns form part of the core component interfaces. This was sometimes awkward and forced us to confront difficult questions early on but it meant that when the European Accessibility Act came into effect, we were already in a great position to support our customers in meeting the new requirements without requiring our consumers to change their implementations of these ubiquitous components.

Distributed governance and clear contribution processes are essential for scaling.

As the system grew, it became clear that a centralised ownership model would not be sustainable. We established a distributed governance model with clear contribution processes and guardrails, which empowered consumer teams to contribute while maintaining the integrity and consistency of the system.

This balance of autonomy and oversight was crucial for scaling effectively.

AI augmentation dramatically improves efficiency in both contribution and governance.

The integration of AI tools for code generation, review, and governance tasks significantly increased our velocity and efficiency. Thankfully we had architected the system in a way that allowed us to leverage these tools effectively as they became available, with strong detailed documentation, consistent patterns and conventions.

converting contribution guidelines and governance processes into prompts was very helpful; but the real enabler was the development of skills that could plug into the AI tools developers were already using, increasing the accuracy and quality of their contributions and reducing the friction of contribution to the point where we saw a significant increase in consumer contributions and engagement with the system. converting contribution guidelines and governance processes into prompts was very helpful.

On the governance side, AI tools helped us to automate routine maintenance tasks, enforce coding standards, and even identify potential issues before they became problems, allowing our small core team to manage a much larger and more active ecosystem of contributors and consumers than would have been possible otherwise.

Adam Harrington

Frontend Specialist

Software Engineer · Design Systems specialist · Frontends at Scale

Rather than a wall of CV text, query my experience directly. Ask about scale, leadership, architecture, or better yet: say what challenges you're facing and ask how I can help.

Press Enter to send