The hardest design problem was the eligibility flow, a branching conversation that had to route employees correctly based on leave type, employment status, hours worked, tenure, state of residence, and reason for leave, while still feeling like a supportive conversation rather than a bureaucratic form.

I mapped the full decision logic in a structured document that served as the single source of truth for both design and engineering, capturing every question, possible answer, conditional path, and resulting action.

To support reliable decision making at scale, I first modeled the experience as a structured operational system rather than a linear flow. Each user input triggered eligibility checks, policy rules, and state transitions that determined the next available actions.

This approach allowed engineering to implement the system as a deterministic decision model while ensuring users experienced a clear, guided interaction rather than navigating raw policy logic. Over time as complexity increased, we explored linear models while delivering to the client in a narrative, digestible way.