COMPRIA

Inspiration

We built COMPRIA after realizing that many small businesses don’t fail because they lack sales, but because their cash doesn’t arrive when they need it. We wanted to turn banking from something that only shows what already happened into something that helps businesses prepare for what comes next.

What it does

COMPRIA analyzes a company’s financial history and forecasts its future cash flow. It can detect possible liquidity problems, simulate situations like a customer paying late, and compare actions such as delaying eligible payments, accelerating collections, or using available credit. An AI copilot also helps users understand their data, forecasts, and recommendations.

How we built it

We built the backend in Python and created our own cash-flow forecasting and optimization engine. We use historical patterns, time-series models, and Monte Carlo simulations to model different possible futures. We also integrated Gemini for document analysis and explanations, SQL Server for storing company data and historical information, and a web interface that brings everything together.

Challenges we ran into

The hardest part was connecting everything into one system. Financial data can be messy, forecasts are uncertain, and recommendations need to reflect actions a business can actually take. We also had to make sure the AI explains the analysis rather than inventing financial calculations, so we separated the AI from our Python financial engine.

Accomplishments that we're proud of

We’re proud that COMPRIA goes beyond simply displaying a forecast. It can analyze financial information, model thousands of possible cash-flow outcomes, identify risk, compare possible interventions, and explain the results through one interface. We also built the system so the business remains in control of every simulated decision.

What we learned

We learned that predicting cash flow is only the beginning. A useful financial tool also needs to communicate uncertainty, understand the company’s available options, and explain why a particular action may help. We also learned that AI is much more useful when combined with reliable mathematical tools instead of being asked to do everything itself.

What's next for COMPRIA

Next, we want to connect COMPRIA to real banking and accounting data, improve our models with larger historical datasets, and make recommendations even more personalized to each business. Our long-term goal is for COMPRIA to become a financial copilot that helps small businesses identify cash-flow problems early enough to actually do something about them.