Inspiration
Have you ever wanted to buy something, checked your bank account, seen that you have enough money, but still wondered whether it's actually the right decision?
Maybe you can afford it today, but what about next week's bills? Would waiting until Black Friday save you money? Or could putting that money into a short-term deposit be a better choice?
These small everyday decisions can add up over time and affect our financial stability. We wanted to build something that helps people understand the consequences of their choices before making them.
That's how Nat AI Assistant (FinTime) was born.
Our goal is simple: help people build financial resilience, one decision at a time.
What it does
Nat AI Assistant is an AI-powered financial coach that helps customers understand, plan, and take control of their financial future.
Instead of only showing users how much money they have or where they've spent it, our platform helps them explore what could happen next.
Our key features include:
- Financial Time Travel: Forecast future balances, upcoming expenses, and potential financial difficulties before they happen.
- Smart Purchase Decisions: Ask whether a purchase is affordable and see how it could affect future finances.
- Buy Now or Wait: Compare purchasing today with waiting for potential discounts, such as Black Friday.
- Short-Term Deposits: Explore whether setting money aside for a short period could be worthwhile, including potential returns and trade-offs.
- AI Financial Coach: Ask questions naturally and receive personalised explanations grounded in financial data.
- Spending Insights: Understand spending habits and identify opportunities to save.
- Financial Accessibility: Choose between friendly, direct, and simple English explanations, with optional voice support.
Most importantly, our assistant doesn't make decisions for customers. It gives them the information and confidence to make their own.
How we built it
We designed Nat AI Assistant using a modular architecture that separates financial calculations from AI-generated explanations.
Our technology stack includes:
- React and Vite for an interactive, responsive frontend.
- Python and FastAPI for backend services and communication.
- SQLite for managing financial information and transactions.
- A deterministic forecasting engine to calculate future balances, simulate purchases, and evaluate financial scenarios.
- An LLM-powered agent with tool calling to interpret financial data and provide personalised explanations.
- ElevenLabs for voice-based accessibility.
- Synthetic financial datasets to demonstrate different customer circumstances without exposing real banking information.
One of our most important design decisions was separating the AI from the financial calculations.
Rather than allowing the language model to invent figures, our forecasting engine performs the calculations, while the AI explains the results using those figures.
This makes the system more transparent, testable, and suitable for sensitive financial conversations.
Challenges we ran into
One of our biggest challenges was balancing the flexibility of conversational AI with the accuracy required for financial information.
Financial decisions depend on reliable numbers, and even small calculation errors can lead to misleading conclusions. We addressed this by separating deterministic financial calculations from AI explanations and designing checks to trace the figures used in responses.
Another challenge was integrating several independent components within a limited hackathon timeframe. Our team worked in parallel on data, forecasting, AI, frontend development, and integration, which required clear interfaces and continuous testing.
We also focused on accessibility. Financial terminology can be confusing, so we designed different communication styles, including simple English, to make explanations easier to understand.
Finally, we considered reliability. External AI services and network connections can fail, so we designed fallback mechanisms to keep essential financial explanations available.
Accomplishments that we're proud of
We're especially proud that we built more than a traditional budgeting dashboard or generic AI chatbot.
We created a financial decision-making experience that helps customers explore different possible futures before committing to a choice.
Our proudest accomplishments include:
- Bringing the concept of financial time travel into everyday spending decisions.
- Connecting purchase simulations, future forecasts, and savings scenarios in one experience.
- Designing an AI assistant that explains financial calculations rather than making them up.
- Making financial information more accessible through conversational language and different response styles.
- Building a modular system that can be extended with additional financial tools and services.
- Keeping the focus on real customers and the everyday decisions that affect their financial wellbeing.
What makes us proudest is the potential impact: helping someone avoid an unexpected shortfall, make a more informed purchase, or feel more confident managing their money.
What we learned
This hackathon taught us that building responsible financial AI is about much more than connecting an LLM to a banking application.
We learned the importance of transparency, explainability, and making sure financial predictions can be traced back to reliable calculations.
We also realised that financial inclusion isn't only about access to financial services. It's about helping people understand those services and feel confident using them.
From a technical perspective, we gained experience in coordinating parallel development, integrating AI tool calling with deterministic backend systems, and designing around real-world failures.
Most importantly, we learned that the most valuable technology doesn't necessarily make decisions for people. It empowers them to make better-informed decisions themselves.
What's next for Nat AI Assistant
We believe Nat AI Assistant has the potential to become a proactive financial companion for everyday banking customers.
Our next steps would include:
- Secure Open Banking integration to provide insights based on real customer transactions, with appropriate consent and safeguards.
- Smarter financial forecasting that accounts for recurring bills, irregular income, seasonal spending, and unexpected expenses.
- Proactive financial alerts that identify possible shortfalls before they happen.
- More financial scenarios, including saving for major purchases, building emergency funds, and comparing different savings options.
- Greater accessibility, including multilingual conversations and improved voice interaction.
- Personalised financial education that helps customers understand the reasoning behind financial decisions.
- Customer testing and validation to measure whether the assistant genuinely improves financial understanding, confidence, and decision-making.
Our long-term vision is to make proactive financial support accessible to everyone, regardless of their income, background, or level of financial knowledge.
Because financial resilience isn't built through one big decision. It's built through the small decisions we make every single day.
Nat AI Assistant — See Tomorrow. Take Control Today.
