AI Voice Agent for Virtual Assistant Sales

A bilingual (French/English) AI voice agent built with Retell AI that handles inbound sales calls, qualifies leads, books appointments, and provides 24/7 customer support for DB Studio.

🎯 Project Overview

Built for DB Studio, a company providing virtual voice assistants to professionals. This agent automates the entire inbound sales process - from initial inquiry to appointment booking.

✨ Key Features

• Bilingual Support: Seamlessly handles French and English conversations
• Lead Qualification: Gathers business needs and qualifies prospects
• Appointment Booking: Checks calendar availability and schedules consultations
• Smart Routing: Answers FAQs, transfers complex calls, and escalates when needed
• SMS Notifications: Sends appointment confirmations via Twilio
• Call Summaries: Generates conversation summaries for follow-up

🛠️ Tech Stack

• Voice AI: Retell AI
• Telephony: Twilio (Voice & SMS)
• Calendar: Google Calendar
• Languages: French & English

📋 Agent Capabilities

1. Call Routing

• Greets callers and identifies intent
• Routes to appropriate conversation flow

2. FAQ Handling

• Answers questions about services
• Explains pricing (3 tiers: €149, €249, €449)
• Covers features and policies

3. Appointment Booking

• Qualifies business needs
• Checks calendar availability
• Collects contact information
• Books and confirms appointments

4. Human Escalation

• Transfers frustrated callers
• Handles complex requests
• Sends SMS for callbacks

🗂️ Repository Structure

/prompts
  ├── identity.md
  ├── intro.md
  ├── faq.md
  ├── appointment_booking.md
  └── call_transfer.md
/knowledge-base
  └── db-studio-kb.md
/integrations
  ├── twilio-sms.md
  └── calendar-setup.md
/assets
  └── demo-video.md
README.md

🚀 Results

• Handles 100% of inbound calls without human intervention
• Instant response time (vs previous manual handling)
• 24/7 availability for international clients
• Seamless bilingual experience

💡 Implementation Highlights

• Prompt Engineering: Crafted conversation flows for natural interactions
• Context Management: Maintains conversation state across different intents
• Error Handling: Gracefully recovers from misunderstandings
• Bilingual Logic: Detects and adapts to caller's language preference

📞 Sample Conversation Flow

1. Caller reaches DB Studio
2. Agent greets in appropriate language
3. Identifies need (pricing question)
4. Provides accurate information from knowledge base
5. Offers to book consultation
6. Collects details and checks availability
7. Confirms appointment and sends SMS
8. Offers further assistance or ends call

🎥 Demo

See the agent in action: Demo Video

🔐 Privacy & Security

• No sensitive data stored in repository
• API keys managed via environment variables
• GDPR-compliant data handling

📝 License

This is a portfolio project. Contact for commercial use.

👤 Author

Taiye Promise Kotiku

• GitHub: Taiye's GitHub
• LinkedIn: Taiye's LinkedIn
• Portfolio: Taiye Website