02H - AI enabled design system
02H encodes our designer, developer, and UX writer into agent skills so AI produces government-compliant, accessible UI by default
Booth 8F-13Team 10 — Design system gurus
Problem Statement
Users
02H serves the entire Whole of Government — from developers and engineers who build and maintain digital services, to product managers and policy officers with little to no frontend experience who need to prototype or deliver digital products quickly.
Use Case
Government teams regularly need to build digital services that are:
-
Accessible
-
Consistent across Whole of Government
-
DSS-compliant
However, using SGDS correctly requires knowledge of:
-
Design guidelines
-
Component usage
-
Accessibility requirements
-
UX writing standards
Pain Point
- Using a design system consistently requires familiarity with:
- Documentation
- Components
- Usage guidelines
- Even experienced developers struggle with:
- Choosing the right components
- Ensuring consistent implementation
- Verifying compliance
- Non-technical users face a steep learning curve and often cannot contribute to frontend work directly
- Teams may unintentionally introduce inconsistencies due to varying levels of technical expertise
Consequence
- In a survey of 15 project teams:
- 10 teams identified SGDS implementation effort as a significant pain point
- This results in:
- Fragmented user experiences across Whole of Government
- Slower delivery timelines
- Risk of falling short of DSS compliance
- Greater dependency on teams with strong frontend expertise
User Research and Market Analysis
User Research
The team conducted:
-
Surveys
-
Interviews
-
Observational sessions with government officers
Key findings:
-
UX inconsistency persists across government digital products
-
Teams struggle to:
- Identify the correct SGDS components
- Apply components consistently
- Meet compliance requirements
-
10 out of 15 survey respondents flagged SGDS implementation effort as a meaningful barrier
De-risking
Primary technical risk:
-
AI hallucination leading to incorrect or non-compliant UI code
Mitigation approach:
-
Integrate SGDS knowledge through structured Agent Skills
-
Use guided prompting workflows
-
Steer generated outputs towards compliant and consistent implementations
-
Reduce reliance on model assumptions
Stakeholder Buy-in
- Design Practice is a key stakeholder and strong supporter, with business owner coming from the Design Practice
- Internal GovTech teams have expressed interest in the approach
- The hackathon team has also engaged SGDS to share to the participants how to use SGDS agent skills, showing existing organisational investment in solving this problem
Solutioning
Problem-Solution Fit
02H makes SGDS accessible through natural language.
Instead of:
-
Reading documentation
-
Identifying components manually
-
Writing compliant frontend code from scratch
Users can:
-
Describe the UI they want
-
Receive SGDS-compliant frontend code automatically
Functionality
Core User Flows
- Prompt-to-UI
- Users describe the UI they want
- 02H generates SGDS-compliant frontend code
- Legacy UI Conversion
- Users upload an existing UI
- 02H converts it into SGDS-compliant implementation
- Figma-to-Code
- Users upload a Figma screen
- 02H converts it into SGDS-compliant frontend code
All flows are powered by AI agents with SGDS Agent Skills.
User Experience
For SWEs
02H:
- Reduces time spent navigating documentation
- Removes compliance verification overhead
- Allows teams to focus on higher-value engineering work
For product & UXD
02H:
- Lowers the barrier to frontend development
- Enables rapid prototyping without coding knowledge
- Provides a consistent prompting experience across user types
User Validation
The team has built demos showcasing 02H's capabilities on existing government websites. Formal user feedback is being gathered as the next step, and findings will be used to refine the prompting layer and improve output quality.
Impact & Next Steps
Success Metrics
North Star Metric:
-
Time saved
Measured through:
-
Hours saved per sprint for development teams
-
Time to first working prototype
These two lenses capture value for both technical and non-technical users.
Expected Impact
Short Term
- Faster SGDS-compliant prototyping and development
- Reduced dependency on specialised frontend expertise
Medium Term
- Greater consistency across citizen-facing digital services
- Improved baseline DSS compliance across Whole of Government
Long Term
- Demonstrates a scalable framework for operationalising:
- Design systems
- Standards
- Government guidelines
- Potential expansion beyond SGDS into other government domains and workflows