Multi-channel AI agent platform that routes customer conversations across Web Chat, WhatsApp, and Email from a single inbox. The backend runs on Node.js with Express, using WebSocket for real-time message delivery and typing indicators. Each channel has its own adapter: WhatsApp connects through the Business API webhook, email through IMAP polling with 30-second intervals, and web chat through a lightweight JS widget that drops into any page. The agent responds using retrieval-augmented generation — incoming messages are matched against a knowledge base of uploaded documents (FAQ, product catalogs, return policies) via vector embeddings, then formatted into natural conversational responses with source attribution. Built a conversation routing engine that handles automatic handoff to human agents when confidence drops below threshold, with full conversation history preserved across channel switches. Configurable personality, tone, and language settings per agent. Node.js + Express backend, React frontend with real-time state management via WebSocket.

This one is for teams that hand everyday work to AI agents. In one view you can see what got automated, how much time it saved, how each agent is doing, and you can ask the AI assistant anything. In the video, I ask "Who might churn this week?" and Lyra replies with three at risk accounts and check in emails ready to send. My goal was calm and minimal, with lots of white space, while the screen still feels alive.

What would you add or remove? I'd love your feedback.

asking it who might churn and getting the check in emails already drafted is the part that sells it