Inspiration

Most financial apps focus on the past. They show what you already spent, but they don’t answer the real question:

“What happens if I buy this?”

We wanted to turn financial decisions into simulations — giving users the ability to preview multiple possible futures before committing to a purchase.


What it does

Kairo is an AI-powered financial simulator.

Users ask about a future purchase, and Kairo generates three projected financial timelines:

  • Best case
  • Expected case
  • Worst case

$$ B_{future} = B_0 + \sum I_t - \sum E_t - P $$

Where:

  • \( B_0 \) = current balance
  • \( I_t \) = income
  • \( E_t \) = expenses
  • \( P \) = proposed purchase

How we built it

Frontend

  • Flutter Web
  • Interactive timeline visualization

Backend

  • Ruby on Rails API
  • MongoDB
  • Plaid sandbox integration

AI Layer Agentic Claude workflow that:

  1. Retrieves financial snapshot
  2. Generates scenario parameters
  3. Runs deterministic simulations
  4. Produces narrative recommendations

Challenges we ran into

  • Normalizing financial data
  • Detecting recurring expenses
  • Preventing AI hallucination
  • Keeping scope focused

Accomplishments that we're proud of

  • Working multi-scenario simulation engine
  • Structured AI + deterministic modeling integration
  • Clear, interactive financial timelines

What we learned

  • Simulation beats static budgeting
  • Constrained AI is more reliable
  • UX clarity is critical

What's next for Kairo

  • Real Open Banking integration
  • Probabilistic modeling
  • Long-term goal simulations
Share this project:

Updates