MoneyTalks

Our Team

Inspiration

In a world where payment processes are becoming increasingly frictionless, unconscious spending has quietly turned into a universal problem. No matter our age group, we all fall victim to impulsive purchases — sometimes within a single heartbeat. Moments of carelessness or oversight often lead to regret, and we wanted to create something that could interrupt that cycle. MoneyTalks was born out of a simple idea: What if your finances could talk back — and guide you — before you make a mistake?

What it does

MoneyTalks securely retrieves spending data from connected bank accounts and feeds this information into an AI system. Using a combination of structured prompts, user-provided context, and behavioural patterns, the AI generates personalised financial insights, warnings, and advice to help users make better money decisions.

How we built it

Backend: Express.js, Convex (serverless database) Frontend: React + Vite, TypeScript, Tailwind CSS Database: Convex (real-time database) AI: Claude (Anthropic), ElevenLabs (TTS)

Challenges we ran into

Prompt engineering for financial reasoning: Teaching the AI to give helpful, non-generic advice required multiple iterations and tuning. State management across real-time components: Syncing Convex updates with the UI while maintaining performance took careful architectural planning. Voice integration: Combining text and TTS responses in a smooth flow introduced latency and formatting challenges. Time constraints: Packaging backend, frontend, and AI into a functional experience within a short timeframe pushed our team to optimise aggressively.

Accomplishments that we're proud of

Successfully integrated voice AI, allowing MoneyTalks to literally “talk” to the user. Designed a clean, fast UI that feels modern, intuitive, and scalable. Created a meaningful tool that has the potential to positively impact everyday spending behaviour.

What we learned

How to architect a full pipeline involving backend, frontend, real-time database, and AI models. The importance of crafting high-quality prompts and guardrails to produce consistent AI behaviour.

What's next for MoneyTalks

Full bank API integration (e.g.,Finverse) for seamless onboarding. AI-generated budgeting plans tailored to spending patterns. Inclusivity features such as for the visually impaired