Inspiration

Leapro was born out of frustration with traditional banking support—long wait times, endless menus, and rigid service channels. In today’s fast-paced world, banking should be as easy as having a conversation. A Salesforce survey found that 63% of service professionals believe generative AI can improve response times. Gartner predicts that by 2025, 80% of customer service teams will use generative AI to boost productivity and enhance customer experience. Banks need to adapt, and Leapro is here to lead that transformation.

What it does

Leapro transforms banking experience into a seamless, real-time conversation.

Phase one:

  • Face detection and loading basic customer information
  • Customers can inquire about bank's information, like products, services, fees, ATM locations, ..etc Phase two:
  • Customers can inquire about balances, view statements, or check pending transactions. Phase three:
  • We don’t stop at inquiries. Leapro plans to enables self-service actions like paying fines, settling fees, or transferring funds—securely and effortlessly. Your banking copilot is here to help, anytime, anywhere.

No waiting, no hassle. From Phone, desktop or Kiosk

How we built it

Leapro is powered by a robust, scalable tech stack:

  • Frontend: React.js with Tailwind CSS for an intuitive user interface, and it interact with the back end through websockets, transferring voice between the client and server.
  • Backend: Java with Spring framework for business logic and AI orchestration, and it interact with the backend using APIs.
  • AI Integration:
    • GitHub Copilot with VS Code & IntelliJ
    • Azure OpenAI (GPT-4o & GPT-4o Realtime)
    • Azure AI Search
    • Azure Spring Apps
    • Embedding models (text-embedding-3-large)
    • Face recognition (Face-api.js)
    • Java Spring AI We designed a balanced architecture that blends interactivity with AI-driven intelligence, ensuring a smooth and responsive experience.

Features & Functionality

  • AI Assistant Configuration: Bank staff can fine-tune responses and upload documents for RAG-based conversations
  • Customer Data Management: Upload customer details (name, image, age) manually or integrate with CRM
  • Multi-Platform Access: Customers interact with the AI assistant from any device
  • Face Recognition: Enables seamless and secure user identification Real-Time Conversations: Instant, hassle-free interactions with banking services

Challenges we ran into

  • AI Integration: Learning by doing was a challenge, but it expanded our expertise
  • User Experience: Removing friction from interactions was a top priority—face recognition was complex to implement but greatly enhanced the experience Leapro is more than just an AI assistant—it’s the future of customer-first banking.

Accomplishments that we're proud of

We're proud to have brought advanced AI into a challenging industry like banking. We've built a tool that can truly transform the banking experience and improve the lives of millions.

What we learned

How to refine prompt engineering for a smarter AI assistant. Keep thinking big and pushing boundaries.

What's next for Leapro

Enhancing the AI assistant’s capabilities by integrating with more tools and APIs. Automating the assistant to make it more proactive and intelligent.

Built With

Share this project:

Updates