Kabala

Inspiration

Our inspiration came directly from the Capital One coaches, who challenged us not just to have an idea, but to prove the problem with real data.

We dove into INEGI's economic data and were shocked by what we found. In Mexico, 96.1% of all businesses are "Micro" (0-10 employees).

But their reality is brutal:

They are blind: A huge portion of these businesses don't use any formal accounting system, managing their finances in a notebook.

They are alone: A staggering 86.6% of these micro-businesses have no access to credit.

They are dying: Because of this, the latest INEGI (EDN 2023) data shows that nearly 30% of businesses die within 4 years , and for the service/food sector, that number is even worse at over 32%.

We weren't just building an app; we were building a vital support system for the real Mexican economy.

What it does

Kabala is a free financial tool designed for the 96.1% of micro-business owners. It's a simple app that provides immediate peace of mind.

It does three things:

Predicts: It clears the financial "fog" for owners who are operating blind. It analyzes their real-time cash flow and predicts their future balance.

Alerts: It's a pro-active copilot. It doesn't show history; it shows the future. It sends a simple, clear alert.

Solves: This is the core. For the 86.6% of owners excluded from the financial system, Kabala partners with Capital One to offer the exact micro-loan or invoice-factoring they need, with one click, before the crisis hits.

How we built it

We built Kabala with a "high-velocity" stack, prioritizing a functional MVP and a clean user experience in 36 hours.

Framework: Next.js (App Router)

Frontend: React with TypeScript

Styling: Tailwind CSS for rapid, responsive UI development.

Backend (MVP): We made a strategic decision to simulate our backend using Next.js API Routes. This allowed us to build a fully functional demo, complete with a mock JSON API that returned the realistic data (based on our INEGI research) needed to power our predictive dashboard.

VCS & Deployment: GitHub for version control and Vercel for instant, continuous deployment.

Challenges we ran into

The "52%" Myth vs. The Truth: Our initial research pointed to a "52% failure rate," but the coaches pushed us to verify it. We had a crisis moment when we couldn't find the source. We pivoted, dug deep into the raw edn2023 CSV files, and found the real, verifiable data: 29.6%. This was a critical lesson in data integrity.

The 4 PM Backend: We realized on Saturday afternoon that building a full Python/Flask backend with a database was impossible. We had to pivot fast. We decided to leverage Next.js API Routes to build a mock backend in under 30 minutes, which saved the project.

The UI vs. The "WOW": Our first UI design was a typical dashboard: "total income," "total expenses." We realized this showed history, but our pitch sold prediction. We had to re-architect the UI to put the prediction and the alert card front-and-center.

Accomplishments that we're proud of

A Data-Driven Pitch: We aren't just guessing. We are most proud of building a project on a foundation of hard, official data from INEGI. We can prove our market is 96.1% Micro , our problem is that 86.6% lack credit , and the stakes are a 30%+ failure rate.

A Functional MVP: We didn't just make slides. We built a beautiful, functional frontend deployed on Vercel that successfully fetches and displays data from its own API endpoint.

Listening to Feedback: We took the coaches' advice to heart, found the data, and built a better project because of it.

What we learned

A Good Idea is Nothing Without Data: The hackathon truly began when we stopped talking about features and started digging into the "why" backed by INEGI stats.

The "WOW" Must Be the UI: Your core value proposition (for us, prediction) can't be a bullet point. It has to be the first and most important thing the user sees.

MVP Means "Minimum Viable Backend": A simulated Next.js API route is 1000x better than a half-finished "real" backend. It lets you test, demo, and deploy instantly.

What's next for Kabala

Build the Brain: Replace our mock API with a real Machine Learning prediction model. We'll use historical transaction data to create a true forecasting engine.

Real API Integration: Connect to Capital One's APIs (and Plaid/Open Banking) to link real bank accounts and offer actual pre-approved loans.

User Testing: Get Kabala into the hands of real, We need to test with the owners of the local taquerías, papelerías, and cafés to ensure our solution truly solves their pain.