Billionhands

Human-curated rankings, enhanced by AI.

foundations designer

· product manager

Mar. 2025

CONTEXT

People share opinions on what's best all the time, but nowhere consolidates them into something trustworthy. Existing platforms are either too niche, algorithmically opaque, or locked to a single category. Billionhands was built to turn community taste into something structured, searchable, and shareable.

At Billiohands, I led the design team in crafting the app from scratch. The goal was to create a dynamic platform where users could easily create lists, rank their favorite things, and share opinions on topics they care about.

RESEARCH

Before designing anything, we ran 20+ user interviews to understand how people discover and trust rankings. Three patterns emerged: users distrust rankings with no visible methodology, they want to participate — not just consume — and they share opinions in context (chat, DMs) rather than on dedicated platforms. That shaped every design decision.

24 users interviewed previously

2 2 competing mental models identified

THE APPROACH

We built around a simple loop: create a list → rank with AI assistance → invite community votes → share. The AI layer doesn't replace judgment — it suggests candidates based on category, which users then reorder and trim. This removed the blank-page friction that kills engagement on creation-heavy apps.

4.1x

more engagement on human-curated lists vs. AI-only rankings

MVP OUTCOMES

The MVP shipped with list creation, community voting, and AI-assisted ranking. Early cohort data validated the human-first model.

Early feedback revealed a trust gap with AI-generated rankings: users engaged with them but hesitated to share or endorse them as their own opinion.

ITERATE

SO WE ITERATE…

Adoption of human-only lists grew fast, and those lists consistently outperformed AI-assisted ones on shares and comments. The AI was useful for reducing the blank-page problem, not for replacing the human voice.

47%

of lists shared externally within 24h of creation

TAKEAWAY

The original assumption was that AI would do the heavy lifting and humans would refine. The real model was the opposite: humans lead, AI assists on demand. Building the toggle between modes — and making it obvious which type of list you're looking at — became one of the more important design decisions in the product

The original assumption was that AI would do the heavy lifting and humans would refine. The real model was the opposite: humans lead, AI assists on demand.

Building the toggle between modes — and making it obvious which type of list you're looking at — became one of the more important design decisions in the product

Cloudfox

Building a fintech platform from scratch that powered R$47M in transactions.

Cloudfox

Building a fintech platform from scratch that powered R$47M in transactions.