Kezia Lim

Proof of Work

Flagship Case Study

BECOME

An AI journal that helps people see who they are becoming while they are still living it.

TODO: BECOME hero shot / product banner

The Product

BECOME sign-in screen

Sign in — a private space from the first screen

BECOME onboarding with founder's message

Onboarding — the founder's own story, then straight into writing

BECOME journal entry with inline AI reflection

Journal Entry — write freely, BECOME reflects inline

BECOME Pattern Dashboard in dark mode

Pattern Dashboard — after 5 entries

BECOME Strengths & Blind Spots in dark mode

Strengths & Blind Spots — after 10 entries, framed gently

BECOME Inner Compass summary

Inner Compass Summary — after 4 weeks, the signature feature

Timeline

📝

PRD Written

🌙

Built Nights & Weekends

🪞

Founder-Led Personal Use

🚀

Live

3 weeks from PRD

The Challenge

Most people are changing — growing, learning, struggling, becoming — but they can't see it while it's happening. Journals exist, but they're passive; they capture what happened, not what it means. The person this needed to reach: The Conscious Seeker, someone in their late 20s to 40s already doing some version of self-reflection, but missing a single place to make sense of it over time.

Design Decisions

The AI, not a coach

No advice, no direction. It asks the question that unlocks something in you, and trusts you to find your own answer.

Quiet by default

No clutter, no dashboards on empty states. Data visualisations only when there's data worth showing.

Mobile-first

Designed for the phone in your hand at the end of a long day.

What was left out

No social features, no habit tracking, no onboarding quiz, no therapist integrations. BECOME is not therapy, and doesn't pretend to be.

Architecture

📓

Journal Entry

Next.js

AI Reflection

Claude API

🗄️

Data & Auth

Supabase

📊

Analytics

PostHog

Results & Roadmap

Near-term goal: 100 active users with weekly retention above 40%. Beyond that: a weekly reflection digest, guided prompts for specific life chapters, exporting your journey as a document, and eventually sharing entries with a trusted coach, therapist, or partner.

Lessons Learned

The single biggest risk was always that the AI's reflections would feel generic instead of genuinely insightful — which is why founder-led personal testing happened before anyone else touched the product, including two days of after-hours debugging on the Inner Compass feature that ended with reading what it had written and crying, because it was accurate. Fourteen years of building other people's products taught me these things take months. This one didn't — proof, mostly to myself, that I could still move like that for something of my own.

Other Projects

Each entry follows: Challenge → Approach → Impact

Appointment AI / Booking Assistant

TODO: challenge, approach, and impact details

Government AI Workflow Projects

TODO: high-level description of scope and outcomes

Chatbot Builds

TODO: which chatbots, for whom, and what they did

Hackathon Projects

TODO: which hackathons, what was built, results

Future AI Products / Experiments

TODO: what's currently in progress