Inspiration
We all know the pain: it's Friday night, you feel like a king, and your bank account is full of confidence. But then Monday comes, and you're too scared to open your banking app because you spent half your money on extra fries and late-night decisions!
People genuinely dread checking their balance after a wild weekend. We wanted to solve this "Friday Night Effect" by creating a smart AI financial coach that helps users manage their spending before the damage is done.
What it does
FinancialFriday is an AI-powered personal financial coach that analyses your spending habits and transaction history to help you make smarter financial decisions.
Here's how it works:
Personalised AI Coaching: Chat with your AI financial coach, set savings goals, and evaluate purchases from five different perspectives: Financial, Research, Devil's Advocate, Lifestyle, and Future.
Smart Alerts Before You Spend: By analysing historical spending patterns, FinancialFriday identifies when you're most likely to overspend, especially on Friday evenings. It gives you friendly reminders before those spending habits repeat.
Gamified Checkpoints & Streaks: Saving money becomes a game! Complete financial checkpoints, maintain your streaks, and build healthier spending habits through achievable challenges.
Future Forecasting: Explore possible financial outcomes using Monte Carlo simulations. See how small changes in your spending habits could influence your savings over time.
Spending Insights: Understand where your money goes with interactive charts, spending heatmaps, category breakdowns, and personalised budgets.
How we built it
We built FinancialFriday as a fully native iOS application using Swift, SwiftUI, SwiftData, and Swift Charts.
Apple Intelligence: We integrated Apple's on-device Foundation Models framework to power our AI financial coach. The model evaluates purchasing decisions using five specialised perspectives through context engineering.
Financial Analysis: We built a transaction analysis engine to calculate income, expenses, cash flow, and spending patterns from historical transaction data.
Predictive Insights: We use historical spending patterns and statistical calculations to identify potential overspending periods and generate proactive financial alerts.
Financial Forecasting: We implemented Monte Carlo simulations to explore possible future financial outcomes under different spending and saving scenarios.
Gamification: We developed checkpoints, streaks, and progress tracking to encourage better financial habits.
Privacy: Financial information and AI processing remain on-device, without requiring an external AI API or backend.
For our prototype, we use synthetic transaction data to demonstrate realistic financial behaviour without exposing anyone's personal banking information.
Challenges we ran into
One of our biggest challenges was making financial advice feel personal rather than generic.
We needed our AI coach to understand financial context, evaluate purchases from different perspectives, and provide meaningful recommendations without inventing financial information.
Another challenge was turning historical transaction data into useful predictions. Identifying recurring spending patterns while distinguishing genuine trends from random spending events required careful consideration.
We also had to make complex financial forecasts understandable on a small iPhone screen without overwhelming users with numbers.
Most importantly, we wanted FinancialFriday to feel like a supportive friend, not a judgemental accountant.
Accomplishments that we're proud of
- Built a fully native iOS financial coaching app using Apple's frameworks.
- Integrated on-device Apple Intelligence to evaluate financial decisions from five different perspectives.
- Turned transaction history into actionable insights, including spending trends, alerts, budgets, and financial forecasts.
- Introduced gamified financial checkpoints and streaks to make saving more engaging.
- Created a privacy-focused prototype that works without sending financial data to an external AI service.
Our biggest achievement is turning a simple idea — "What if your banking app warned you before you overspent?" — into a working iOS prototype.
What we learned
We learned that effective financial coaching isn't just about showing people numbers. It's about helping them understand their habits and make better decisions at the right moment.
We also discovered how context engineering can transform a single on-device AI model into a financial advisor with multiple specialised perspectives, without requiring multiple separately trained models.
Finally, we learned that financial planning becomes much more approachable when combined with friendly interactions, clear visualisations, and gamification.
What's next for FinancialFriday
- Real Banking Integration: Connect securely with banking APIs to analyse real transactions and provide personalised insights.
- Real-Time Spending Alerts: Deliver timely notifications when users approach their predicted high-spending periods.
- Smarter Financial Coaching: Improve personalisation as the app learns more about individual financial habits.
- AI-Powered Bill Negotiation: Explore agents that can help users identify opportunities to reduce recurring expenses.
- Social Savings Challenges: Introduce optional challenges where friends can motivate each other to reach their savings goals.
Our vision is simple: Stop telling people where their money went. Start helping them decide where it should go.
