Device - Macbook Air

Challenge / Problem Statement

The client faced significant challenges in enhancing user experience:

  • Users struggled to understand complex insurance policy terms.
  • Manual customer support was costly, slow, and unscalable.
  • Policy documents were lengthy, unstructured, and hard to navigate.
  • Lack of quick, contextual answers led to poor user satisfaction and increased support costs.

Objectives

  • Simplify insurance policy understanding through conversational AI.
  • Provide users with real-time, accurate answers to policy-related queries.
  • Reduce dependency on manual support teams.

Enhance decision-making during medical emergencies and claim submissions.

Process & Implementation

  • Ingested and pre-processed insurance policy PDFs.
  • Chunked policy documents for optimized LLM performance.
  • Built FastAPI backend for query handling.
  • Applied Gemini Pro + RAG for contextualized responses.
  • Integrated chatbot with Digi Sparsh’s app ecosystem.

Our Solution

We developed a policy-focused conversational chatbot powered by Google Gemini Pro and integrated with FastAPI. The solution:

  • Parsed insurance policy PDFs into structured chunks for LLM comprehension.
  • Applied prompt engineering and retrieval-augmented generation (RAG) for contextual, precise answers.
  • Built a FastAPI backend to process user queries in real time.
  • Integrated the chatbot into the Digi Sparsh app, enabling seamless access across mobile and web platforms.

Tools & Tech Used

  • Backend & APIs: Python, FastAPI
  • LLM Engine: Google Gemini Pro
  • Document Parsing: PyMuPDF, LangChain
  • Authentication: JWT-based secure access
  • Integration: Digi Sparsh mobile and web app

Results & Impact

  • Conversational chatbot capable of answering complex policy-related questions
  • Reduced user dependency on manual customer support
  • Improved user satisfaction during claim and policy queries
  • Responses delivered within 2–3 seconds on average
  • Seamless integration into the Digi Sparsh app ecosystem

Key Takeaways

  • Conversational AI bridges the gap between complex policy language and user understanding.
  • LLMs reduce support costs while improving scalability and efficiency.
  • MoreYeahs enables healthcare and insurance companies to deliver AI-powered, user-friendly solutions.

Project Duration & Team

  • Duration: 6 weeks
  • Team: 1 LLM/AI Developer, 1 FastAPI Backend Developer, 1 Prompt Engineer
Share