Geographic Entitlements
ConstructConnect | Research | OOUX | AI-Assisted Research
Claude AI
The Challenge
When a legacy platform became unavailable, the organization needed to rebuild geographic entitlement functionality for their construction management system—but with limited documentation of how the original system worked. I had very little knowledge of how geographies were set up, how they were used, or what business rules governed them.
This project was managed concurrently with the enterprise Design System overhaul, requiring a high degree of synchronization between specialized feature logic and global system foundations.
Understanding Before Building:
Working with OOUX : I chose the OOUX methodology because the challenge required understanding complex business logic and data relationships with limited existing documentation. OOUX's object-mapping approach was ideal for systematically uncovering how geographic entitlements interact with other system components.
Research & Discovery: By investigating current workflows through technical documentation I was able to fill in the gaps that I might have missed in the object mapping. I utilized Claude to help me analyze database schemas and existing business logic to understand the baseline system already being used.
Workflows and Prototype: Both of these steps helped me identify what should migrate versus what could improve. Result: New workflows informed by existing system understanding.
Project Context:
- Platform migration requiring geographic functionality
- 137 Market Zones, 2,147 Counties, 12,683 records to understand
- Timeline: 5 weeks from discovery to wireframes
- Solo designer on cross-functional team
AI Assisted Methodology
Used strategic AI orchestration to accelerate research and synthesis:
- Research Tools: Claude Desktop, Atlassian Rovo, and Glean for documentation analysis and pattern identification
- Analysis & Synthesis: Claude AI for iterative OOUX/ORCA data modeling (17 iterations), user flow development (13 flows across 7+ revisions), and requirements synthesis
- Prototype Generation: Using VS Code with CoPilot (using Claude Sonnet agent) I was able to create a clickable prototype that we could refine. This included realistic data from the Product Manger.
- Distribution: GitHub Pages for frictionless testing (no downloads, device-agnostic, instant updates)
Deliverables
- 6-object OOUX/ORCA data model documenting business logic and relationships
- 13 validated user flows covering all identified scenarios
- Clickable prototype deployed to GitHub Pages for stakeholder testing
The Impact
- 12,683 Records Analyzed
- 17 OOUX/ORCA Model Iterations
- 13 User Flows Created
- 3 Internal Users Tested
- Gaps Found Before Customer Testing ✓
- Ready for Customer Validation
Key Outcome
By researching first, we prevented costly development rework by understanding domain complexity upfront and demonstrated ability to rapidly learn unfamiliar domains through systematic research
