A Guest Post by Team Loan Rangers (Aditya Vijay, Aryan Nijhawan, Nishant Kumar, Prakhar Singh, Tushar Sharma) | Second Runner-up, True Balance AI Hackathon 2026
How can we make applying for a loan as easy as texting a friend? That single question drove our hackathon project.
While our core True Balance app provides a comprehensive financial dashboard, we wanted to create an additional, ultra-lightweight entry point for customers. We envisioned a complete loan journey, from KYC to document uploads to final disbursal, built entirely within a familiar WhatsApp chat window. Here is how we built it in a single weekend.
For many first-time digital borrowers, jumping straight into a full-featured financial application can feel like a big step. Chat, however, is deeply familiar to India’s 650 million WhatsApp users.
Bringing financial services to this platform creates a highly accessible new channel that perfectly complements our existing app. By streamlining the initial onboarding experience, we believe conversational lending can reduce new customer acquisition costs (CAC) by nearly 50%, aligning with our goals for user growth and faster disbursals.
How we built it
To build this within a strict hackathon timeline, we split the work differently. Instead of dividing by front-end and back-end, we divided ownership by the customer journey.
Team members took absolute responsibility for critical stages like KYC, bank statement uploads, and final disbursal. One team member acted as the integration anchor to connect these independent workflows into a single conversational thread, while another managed business analysis and coordination with partners like Meta and Fyno.
Our goal was to extend the existing True Balance lending platform into a new interface, not replace it.
A Python backend manages the core logic, combining an OpenAI language model with LangChain and LangGraph for intelligent conversations. Fyno handles WhatsApp communication, and the chatbot integrates directly with our existing internal loan APIs. Once a user replies to an interactive promotional message, the conversational engine takes over to guide them step-by-step in multiple local languages.
Making the Chat Box Finance-ready
Building a basic chatbot is easy, but making it reliable for finance is incredibly difficult. Chat interfaces allow users to type literally anything. They might answer indirectly, switch languages mid-sentence, or share unnecessary context.
To solve this, we built an AI-powered extraction layer. This system intelligently isolates only the exact data required for the current step, such as an ID number, ensuring the application progresses smoothly without getting derailed. A separate validation layer then strictly verifies those inputs. This balances the flexibility of natural conversation with the rigid reliability expected from a financial system.
To ensure system stability at scale, we also implemented a configurable rate-limiting mechanism. If a user repeatedly sends irrelevant messages, AI processing temporarily pauses. This safeguards capacity and keeps the platform fast for genuine borrowers.
What’s Next?
Winning third place was great, but watching an ambitious idea evolve into a seamless WhatsApp loan flow in just days was the real reward.
The hackathon is over, but this project continues. Our team is working closely with Legal, Compliance, and Security to evaluate this solution for production. Once approved, our goal is to put conversational lending into the hands of real customers, offering them yet another seamless way to access financial services.
Our vision remains the same as it was on day one: make accessing financial services as simple as starting a conversation.
