Connect bank account
Chat view
Smart Dashboard View

lums.ai - Your Smart Personal Financial Assistant

Inspiration

As international young professionals from Colombia and France living in Toronto, we experienced firsthand the challenges of managing finances in a foreign country. Between navigating different banking systems, understanding investment options, and making smart financial decisions, we realized there was a gap in personalized financial guidance that truly understands your unique situation.

Our personal experience reflects a much broader crisis: according to Northwestern Mutual's 2025 Planning & Progress Study, about 70% of Americans say that financial uncertainty has made them feel depressed or anxious—an 8 percentage point increase from 2023¹. This growing financial anxiety affects millions of people who struggle to make sense of their financial lives without proper guidance and tools.

The idea sparked during a casual conversation, when we realized that while AI assistants could help with almost everything, none truly understood our personal financial data or could provide tailored advice based on our actual spending patterns, goals, and circumstances. We envisioned a smart financial companion that would live in your pocket and provide instant, personalized insights whenever you needed them.

What it does

The first step for a good financial state is to be informed and confident about our personal finances. lums.ai is a mobile-first smart personal financial assistant that transforms how people interact with their financial data. Our app provides:

Intelligent Financial Analysis: Connect your bank accounts through Plaid integration and get instant insights into your spending patterns, saving opportunities, and financial health
Personalized Advice: Ask any question about your finances and receive tailored recommendations based on your actual data, not generic advice
Smart Budgeting: Automatic categorization and budget suggestions that adapt to your lifestyle and goals. This addresses a critical gap: while 68% of people don't know how to create a budget effectively² and 43% of Americans only "somewhat follow" a budget inconsistently³, research shows that people who do budget feel more in control and less stressed about money⁴
Real-time Insights: Instant answers to questions like "Can I afford this purchase?" or "How am I doing with my savings goal this month?"
Secure & Private: All your financial data stays protected with enterprise-grade security

Please note: lums.ai does not provide financial investment advice. We focus on helping you understand your current financial situation and make informed decisions about your spending and budgeting.

Think of it as having a personal financial advisor who knows your complete financial picture and is available 24/7 in your pocket.

How we built it

Our diverse team of four brought together complementary skills that allowed us to tackle every aspect of the project simultaneously:

Architecture

We designed a scalable, three-tier architecture:

Frontend: Mobile app built using Bolt.new's AI-powered development platform, focusing on clean UX design and intuitive user experience
Main API: Centralized backend handling authentication, Supabase database interactions, and Plaid.com API integrations
Smart Core API: Standalone AI-powered engine using Llama3 on Google Cloud Vertex AI for smart search, intelligent financial analysis and recommendations

Team Approach

Full-stack development for robust backend and seamless frontend integration
UX/UI design ensuring the app is intuitive and user friendly
Marketing and branding to create a compelling user experience
Sales strategy to understand user needs and market positioning

Technology Stack

We leveraged cutting-edge tools including Supabase for our database, Google Cloud Vertex AI for AI capabilities, Plaid for secure bank connections, and Bolt.new for rapid frontend development.

Challenges we ran into

Technical Complexity: Integrating multiple APIs while maintaining security and performance was challenging. Handling real financial data required implementing robust security measures and ensuring seamless data flow between our three-tier architecture.
AI Fine-tuning: Making the LLM provide accurate, helpful financial advice required extensive prompt engineering and testing to ensure responses were both intelligent and responsible.
Team Coordination: With team members bringing different expertise areas, coordinating development across frontend, backend, AI, and marketing required excellent communication and project management.
Time Constraints: Developing a production-ready app with AI capabilities, user-centric UI, and robust backend in hackathon timeframe pushed us to be extremely efficient and focused.

Accomplishments that we're proud of

Mobile-First Design: Created an intuitive, aesthetic app that makes complex financial data accessible and actionable
Scalable Architecture: Our three-tier system can handle growth from hundreds to millions of users
Real AI Integration: Successfully implemented LLM to provide genuinely helpful financial advice, not just generic responses
Team Synergy: Four friends from different countries and backgrounds came together to create something greater than the sum of our parts
Hybrid Approach Implementation: Successfully implemented a hybrid architecture to ensure seamless support across both desktop and mobile platforms, enhancing accessibility and delivering a consistent user experience regardless of device.

What we learned

Technical Growth: We mastered integrating multiple complex APIs, learned advanced AI prompt engineering, and gained deep experience with cloud-native architecture on Google Cloud Platform.
Team Dynamics: Working intensively as a multicultural team taught us the power of diverse perspectives and the importance of clear communication in rapid development cycles.
Financial Technology: Deep-diving into fintech taught us about regulatory requirements, security best practices, and the real challenges people face managing their finances.
User-Centric Design: Building a financial app reinforced how critical it is to prioritize user experience – financial data can be overwhelming, so making it simple and actionable was key.
AI Responsibility: We learned the importance of responsible AI implementation, especially when providing financial advice that could impact people's real lives.

What's next

lums.ai is more than a hackathon project – it's our vision for the future of personal finance management:

Immediate Plans

Launch alpha version with select users to gather feedback and refine our AI models
Expand bank integrations to support more financial institutions globally
Implement advanced features like complex pattern recognition, security alerts, and debt optimization

Short-term Goals

Web platform launch for desktop access while maintaining mobile-first approach
Multi-language support for our international user base
Enhanced AI capabilities with more sophisticated financial planning tools

Long-term Vision

Become the go-to financial assistant
Expand to include advanced financial planning, loan optimization, and tax planning
Build partnerships with financial institutions to provide even more comprehensive insights
Scale globally while maintaining the personalized, friendly experience that sets us apart

Our goal is simple: make everyone feel confident and informed about their financial decisions, with their personal AI advisor always in their pocket.