Ready, Set, Insure - AI Voice Agent for Insurance Companies

🚀 Philadelphia CodeFest 2025

👨‍💻 Team Members

• Shams Abrar (sa3868@drexel.edu)
• Elan Rubin (edr53@drexel.edu)
• Logan Voravong (lsv27@drexel.edu)
• Chibuike Nwume (cdn49@drexel.edu)
• Jainam Kashyap (jk3862@drexel.edu)

🎤 Elevator Pitch

Ready, Set, Insure is an advanced AI Voice Agent designed to streamline customer interactions for insurance companies. Leveraging state-of-the-art neural network models, it enables seamless human-like conversations over the phone, handling tasks such as answering inquiries, verifying policy details, and processing claims efficiently.

📝 Overview

Our system provides an intelligent voice assistant that improves insurance customer service through automated, natural conversations. It enhances efficiency, reduces wait times, and ensures seamless interactions between customers and insurers.

📝 Description

Traditional voice AI systems in the insurance industry struggle with natural conversation flow, particularly turn-taking—understanding when a customer has finished speaking without relying solely on silence detection. Human speech is complex, with natural pauses, interruptions, and overlapping dialogue, making real-time processing challenging. Inefficient AI-driven interactions often lead to frustrating customer experiences, increased wait times, and a greater dependency on human agents.

Ready, Set, Insure is designed to modernize customer engagement in the insurance sector by integrating advanced AI models that ensure seamless, real-time conversations. Our system accurately transcribes, processes, and generates human-like responses while intelligently managing turn-taking for a more natural dialogue. By efficiently handling policy inquiries, claim processing, and general customer assistance, our AI agent minimizes wait times, reduces the need for human intervention, and enhances overall customer satisfaction.

🔁 The Voice AI Loop

1. 📞 Phone Call Connection – The system connects an incoming call.
2. 🎙️ Speech-to-Text Transcription – Converts spoken words into text using Deepgram Nova 3.
3. 🧠 Processing with LLM – GPT-4.0 Cluster interprets and formulates responses.
4. 🔊 Text-to-Speech Synthesis – Eleven Labs Turbo generates natural voice output.
5. ☎️ Telephony Output – Twilio transmits the AI-generated speech to the caller.

🏗️ Tech Stack

• 🖥️ Backend: Flask (Python), Vapi API, Ngrok
• 🌐 Frontend: Next.js (TypeScript)
• 🤖 AI Models:
  • 📝 Transcription: Deepgram Nova 3
  • 🗣️ Voice Generation: Eleven Labs Turbo
  • 🤯 Conversational AI: GPT-4.0 Cluster
• 📞 Telephony: Twilio for call handling

🚀 Features

✅ Real-time speech-to-text conversion
✅ AI-powered natural conversation flow
✅ Seamless text-to-speech synthesis
✅ Enhanced turn-taking capabilities
✅ Secure and scalable architecture

🔮 Future Enhancements

• 🎯 Integrate real-time sentiment analysis
  Enhance the AI agent's ability to understand customer emotions and respond accordingly.

• 🌍 Expand language support for multilingual interactions
  Implement multilingual capabilities to serve a broader audience.

• 🧠 Implement dynamic intent recognition for improved customer service
  Utilize advanced NLP techniques to enhance AI-driven query resolution.

📂 Repository & Demo

• 📌 Repo URL: https://github.com/Elan-Rubin/codefest_2025
• 🎥 Demo URL: [Insert Live Demo Link]