FinanceFlow

Most personal finance apps just show you what already happened. You get a chart, a category breakdown, and a reminder that you overspent on food last month. But by the time you see it, it's too late to do anything about it.

People who actually struggle with managing money don't need prettier graphs. They need something that pays attention for them, spots problems early, and tells them what to do before things go wrong.

The biggest product decision I made was to build something that works in the background, not just when a user opens the app. I called it the Budget Guardian. Instead of waiting for users to ask questions, it runs every hour on its own, looks at the user's full financial picture, and decides whether to send an alert or a suggestion.

The simpler option was to use basic rules — like sending a notification when someone hits 80% of their budget. I chose not to, because rules like that miss too much. A user might be spending twice as fast as usual in a category that hasn't hit its limit yet. Or they might have three bills due in the first week of the month that make it look like they're overspending when they're actually fine.

The Budget Guardian picks up on those kinds of patterns, explains its reasoning, and logs every decision so it can be reviewed and improved over time.

When a user adds a transaction or connects their bank account, the AI automatically figures out what category it belongs to. Before that result gets saved, it goes through a validation step to make sure the AI's answer actually makes sense. Receipt scanning works the same way: if you upload a photo of a receipt and the line items don't add up to the total, the system rejects it rather than saving something wrong.

On top of that, the Budget Guardian pulls together the last 90 days of spending, spots any unusual patterns, and feeds everything into what I called the Big 4 Decision Engine. That component gives users a clear view of where they stand financially over the next 30, 60, and 90 days, written in plain language — not just numbers on a screen.

FinanceFlow is fully live and covers everything a user needs: bank sync, AI-powered transaction sorting, proactive spending alerts, receipt scanning, 3-month forecasts, and a personalized financial summary. I also built in a system to test different versions of the AI's responses over time, so the product keeps getting more useful based on real user feedback.

This project shows I can take a messy, real-world problem and turn it into a product that actually looks out for the people using it.