RootCause

Situation

RootCause launched as a simple tool: enter a symptom, get a GNM-based explanation linking it to an emotional conflict. The idea resonated. Downloads grew organically, reviews showed real emotional impact, and the app reached 40+ countries without paid acquisition. But the product stopped at the insight. Users learned what might be behind their symptom, then had nowhere to go. V2 added features like journaling, body signal lookup, symptom pattern analysis, and a health report, but the same gap remained. The app told you what was wrong and stopped there.

Task

Take ownership of the product and rebuild it from the ground up. V3 wasn’t an iteration. It was a new app. The goal: take users from identifying a symptom all the way through to a structured healing plan with daily support. Every screen, every flow, every AI prompt needed rethinking. And the GNM framework had to be medically accurate throughout, not just plausible-sounding.

Action

Product and UX

Reviewed and restructured the full V3 spec from scratch.
Designed the core three-step pathway (Identify > Resolution > Healing Plan) that gives every user a clear trajectory from symptom to action.
Designed four distinct daily check-in card types on a rotating schedule, replacing the static health percentage from V2 with dynamic, contextual progress tracking.
Created the paywall and monetization structure across three tiers (Free / Monthly / Yearly).
Ran 12+ rounds of persona-based UX testing simulating real user journeys and produced fix briefs for every bug, friction point, and content error found.

AI and Content

Designed and wrote the full prompt architecture for Andy, the in-app AI companion that walks users through conflict identification, resolution, and healing plan creation.
Defined the companion’s character, conversation boundaries, and GNM accuracy rules across multiple prompt iterations, all running through Google Gemini in structured JSON output.
Ran a systematic audit of every piece of GNM content against the framework’s actual rules and documented three hard content rules the AI must follow.

Systems and Documentation

Specified a push notification engine with four notification types and a full scheduling system adapted to user behavior.
Consolidated every system, rule, and specification into a 14,000+ word engine document serving as the single source of truth for the team.
Created comprehensive handoff documentation for new engineers joining the project.

Launch

Wrote copy and built the community page for public beta recruitment, driving testers into WhatsApp and Reddit communities.
Set up GA4 event tracking across the community page and all core app features.

Result

Nearly2,000 Users with organic reach across 40+ countries No paid acquisition.

25% Conversion rate from free to paid users

4.8 Star rating across app stores

Product transformed from a one-step symptom lookup into a full guided healing journey with daily engagement.

V3 built on a documented, testable, and GNM-accurate foundation where every prompt, check-in card, and notification follows defined rules.