Everything on voice and chat AI agents — real-time speech processing, telephony and phone-system automation, conversational AI, voice UX design, and production deployment.
From the blog
Forrester 2026 found subscription businesses leave 32% of save-eligible accounts uncontacted. AI voice closes the gap — 40-55% recovery on renewal outreach vs 15-25% human-only. Here is the build.
Pipecat 0.0.7x ships a frame-based pipeline for real-time voice. Wire Daily WebRTC, Deepgram, GPT-4o, and Cartesia into a working agent — code + pitfalls.
Master WebRTC integration for real-time voice AI agents with peer connections, audio streaming, VAD, barge-in, and codec optimization.
Soniox v4 Async (Jan 29) and v4 Real-Time (Feb 5) deliver native-speaker accuracy across 60+ languages with code-switching. Inside the model and where it wins.
WebSocket gives you server-side control. WebRTC gives you sub-second latency. Here is the honest engineering tradeoff for production AI voice agents in 2026.
How Twilio Elastic SIP Trunking and OpenAI's Realtime SIP connector now bridge directly, what the call flow looks like, and the latency budget that actually works.
OpenAI's Realtime API speaks both WebRTC and WebSocket. Here is the production playbook CallSphere uses across 37 agents to decide which transport fits which call path.
Wire Twilio Media Streams to OpenAI Realtime in under 50 lines of Node.js. Real working code, mu-law to PCM16 transcoding, server VAD, barge-in, and production tips.
Average Shopify cart abandonment is 70.22% in 2026. Real-time AI chat recovers 20–35% versus 5–8% for email-only — here is the build pattern that works.
