HealthTech · Kickoff to testable build

Conversational onboarding for a consumer health app

A health app where the intake interview is an AI conversation rather than a twelve-page form, with scoring built on connected platform data.

Conversational onboarding, prompts assembled server-side
AI intakeConversational onboarding, prompts assembled server-side
Health platform connection as an explicit step
Consent-firstHealth platform connection as an explicit step
API contract migration completed mid-project
Schema-firstAPI contract migration completed mid-project

The problem

Consumer health apps lose most users in onboarding: the data needed to personalise anything is exactly the data nobody wants to type. The product also had to read from external health platforms under explicit consent, and derive a score users would trust from data of uneven quality — including users who simply have not got enough history yet.

What I did

Onboarding was split: a short structured form for the demographic facts, and a conversational intake for goals and preferences, with prompt assembly kept server-side so the model’s behaviour can change without shipping an app. Platform connection was built as a consent step rather than a permission dialog. Insufficient data is handled as a first-class state everywhere instead of a silent empty screen, and the API was migrated to a schema-first contract mid-project so mobile and backend stopped guessing at each other.

How it ran

  1. 01

    Research and framing

    Competitor review and user stories before build, to decide which of the intake questions actually earn their place.

  2. 02

    Access and consent

    Social and email sign-in with multi-factor, then health-platform connection handled as an informed consent step.

  3. 03

    Intake

    Structured demographic form plus conversational goal-setting, with baseline goals generated rather than asked for.

  4. 04

    Scoring and insight

    Activity scoring surfaced through a sector dashboard, with explanation on demand and an explicit state for insufficient data.

  5. 05

    Contract and quality

    Schema-first API migration, automated end-to-end mobile tests, and a managed test-case suite rather than manual scripts.

Something similar on your plate?

Thirty minutes, no deck. I will tell you whether it is worth doing at all.