Spendy

Inspiration

We drew inspiration from real-world banking problems, introduced by RBC and the RBC Nomi assistant. Although many banking applications enable users to monitor their expenses, we realized that most financial applications are passive – they display information but never really advise on what to do. We sought to reimagine personal finance as an interactive process where AI actively assists in making better financial choices rather than just logging transactions.

What it does

Spendy is an AI-driven banking assistant that converts passive expense logging into active financial advice. The application analyzes users’ spending habits in real-time and offers them customized information, proactive suggestions, and behavioral cues to assist users in developing better financial practices.

Rather than just displaying where the money was spent, Spendy presents estimated cost in future events and helps users to prepare financially. Also, Spendy integrates various data ingestion methods through receipt and transcript scanning and calendar based integration allowing us to create a full Machine learning analysis pipeline for analyzation of user spending habits and spending trajectory analysis. These ingestion pipelines will feed into our NER and Mathematical models for categorization, analyzation, and summarization of user spending habits and extrapolated by Gemini to help users improve.

How we built it

Frontend: react native (expo)
Language: typescript
Backend: superbase
Deployment: Expo Go
API: Gemini API

Challenges we ran into

One of the most difficult aspects was to take unprocessed financial information and turn it into actionable and credible advice. Financial advice needs to feel true and tailored to the user, which was a delicate process of logic and AI dialogue optimization.

Another aspect was to strike a balance between automation and user agency, making sure that the advice was useful but not annoying. Also, working with multiple components in a hackathon setting with a tight deadline required excellent teamwork and rapid development.

Accomplishments that we're proud of

Built a functional prototype that showcases real-time financial advice
Developed an AI dialogue system that extends beyond basic expense management
Built a user-friendly interface that makes complex financial insights accessible
Worked well together in a tight deadline setting to deliver a unified product vision

What we learned

From this project, we learned how AI can transcend automation and become a decision-enabling tool in our daily lives. We learned how to integrate AI with user-centric design, design scalable application architecture, and communicate complex financial concepts in a clear and simple manner.

We also learned the value of rapid prototyping, effective team communication, and focusing on essential product features in a time-constrained development setting.

What's next for Spendy

Personalize advice with more in-depth behavioral analysis
Implement secure banking APIs to handle real transaction data
Develop a long-term financial planning and forecasting feature
Make the AI advice more transparent so users understand the reasoning behind recommendations