Decision-Centric Enterprise
A ground-up redesign of AI-assisted supply chain setup. From blank config form to a structured, explainability-first onboarding flow.
THE PROBLEM
The old setup had a 40% completion rate. Users hit a blank configuration form asking for strategic context most companies had never documented. Free-text inputs degraded AI recommendation quality, and completed profiles went stale with no way to keep them current.
RESEARCH
Five stakeholder interviews across executive, operational, and technical roles. Common thread: the AI needed to be explainable and contextually calibrated, and operational users needed less manual work, not more. That shaped everything downstream.
DESIGN
Replaced the blank form with a three-phase AI-drafted profile flow. The AI generates a draft from existing supply chain data, walks users through 8 to 12 targeted questions, and maintains a living amendment log. Wireframed five screens across two questionnaire formats with explainability panels and AI-inferred answer badges.
VALIDATION
Ran a CSCO persona session against the wireframe. Key finding: the confirm screen put a strategic liability statement on the wrong person. The Implementation Specialist can't ratify company strategy, only the CSCO can. Surfaced the async approval gap before engineering touched a line of code.
THE RESULT
Target completion rate 85%+ versus a 40% baseline. Full setup completable in under 4 hours. Phase 2 is gated on completion so there's no bypassing setup without consequence. AI recommendations are now calibrated to company-specific rules instead of generic defaults.