Before and After: Rebuilding HealthRoute's Onboarding with AI Personalization

AI · 5 min read

Before the redesign, HealthRoute's onboarding was a long, form-heavy experience that asked users to declare symptoms, preferences, and insurance details in one go. Many users abandoned midway; those who completed the flow often didn't return because the experience felt generic and dense. Product and design prioritized immediate clarity: how do we make the first five minutes feel valuable?

The 'after' introduced a lightweight triage using an on-device natural language prompt: users described their reason for visiting in their own words. An AI model suggested a personalized care pathway (televisit, specialist referral, or self-care content) and filled non-critical fields opportunistically in the background. The team limited AI output to suggestions and included clear affordances to edit any inferred information to satisfy regulatory and trust requirements.

Outcomes: a 30% lift in activation (first scheduled appointment or content consumed), a 22% drop in abandonment during onboarding, and increased subjective trust scores in follow-up surveys. The project shows how constrained AI — used for personalization and inference, not automation of decisions — can transform onboarding without overstepping user control or compliance constraints.