ConversationRelay and the AI Agent Builder Dify: A Multimodal Example with Voice

Kousha Talebian and Rikki Singh from Twilio present a comprehensive guide on building a multimodal AI experience using the Agentic Workflow Builder Dify, Twilio Voice, and ConversationRelay. This setup allows users to create an airline concierge service that can respond to voice commands and send confirmations via SMS, enhancing customer interaction across multiple channels. The process involves setting up accounts with Dify and Twilio, configuring the AI model, and using tools like Node.js, Ngrok, and optionally Fly.io for hosting. The guide details steps for local development and cloud deployment, enabling a seamless user experience where the AI agent provides flight options, asks clarifying questions, and sends SMS confirmations, all while maintaining a conversational tone. The authors emphasize the flexibility and scalability of this system, which can be adapted to various applications beyond the demo example.