Lusis Pro
I turned a manager-assisted intake into a self-service flow for families of children with autism: managers now start admissions with 80% of the child's information, up from 20%.
Industry
Health & ABA Therapy
Scope of work
Product Design · Design System · AI Workflow
Duration
2 months
Challenge
Families of children with autism in the U.S. couldn't complete the pre-clearance intake on their own. More than 50% abandoned the form halfway, so every case depended on a manager filling it in with the parents during the first admission meeting.
That made intake slow and costly: managers arrived at the first meeting with only 20% of the child's information, and much of their time went into walking parents through confusing forms instead of focusing on care. For families, it added administrative stress at an already sensitive moment.
Goal
My goal was to make the intake fully self-service without losing the reassurance families needed, and to give the clinical team complete, reliable information before the first meeting.
I framed the work around three measurable outcomes: families completing the form on their own, fewer incomplete or incorrect submissions, and less manager time spent on data entry. Every design decision was evaluated against them.
Solution
I redesigned the intake as a 10-step self-service flow (down from 12 screens), built end to end with an AI-first workflow in Subframe and Claude. Plain-language guidance, clear action states and visible progress replaced jargon and guesswork, so caregivers always know what to submit and where they are.
Results: managers now arrive at the first admission meeting with 80% of the child's information (up from 20%), spend 40% less time helping families complete the form, and incomplete or incorrect submissions dropped by at least 30%. The design system and flow were connected to development through Subframe's MCP, keeping the handoff pixel-perfect.
