Murf AI

In this edition of the Murf AI Webinar Series, Vijay Rajagopal, Head of BFSI and Fintech GTM at AWS India, speaks with Sneha Roy, Co-founder and COO at Murf AI, to unpack why deploying voice agents inside a regulated financial institution is a fundamentally different challenge, and why most teams underestimate it. With two decades across Amazon Pay, PayPal, Zeta, and ICICI, Vijay has seen BFSI AI deployments up close: what moves from pilot to production, and what quietly stalls. His read: 75% of AI pilots in BFSI never make it to production, not because the technology fails, but because ownership, data foundations, and outcome expectations are never properly set. This session goes into what separates the teams that scale from the ones that stay stuck. Key Topics Covered: 1. Why trust is the core product in BFSI, and why one bad voice interaction compounds faster than a good one 2. The three customer archetypes AWS sees today: experimenters, scalers, and transformers 3. Why voice AI in BFSI is neither a build nor a buy problem but an integration problem 4. The real reasons 75% of BFSI AI pilots never reach production 5. India's multilingual opportunity: why vernacular voice could become the biggest financial inclusion lever of this decade Resources: Murf Agents: https://murf.ai/ai-voice-agent?utm_source=BFSIWebinar

In this tutorial, we wire up a fully voice-driven food ordering agent on Swiggy, locally, with real APIs and real money on the line. Here's what's powering it under the hood: Deepgram Nova - real-time speech to text over a raw WebSocket Open Claw - the agent brain that decides what to do Swiggy Skill (from Claw Hub) - hits the actual Swiggy API Murf Falcon TTS API: Streams audio back with 130 ms time-to-first-audio Every time you speak, six things happen behind the scenes. By the end of this video, you’ll know exactly what each part does, how the full voice pipeline works, and how to swap out any piece of the stack. What's covered: 1. Setting up Deepgram Nova via raw WebSocket 2. Configuring the Falcon TTS plugin in Open Claw 3. Installing and wiring the Swiggy skill from Claw Hub 4. The 4 core files: index.ts, brain.ts, voice.ts, and ear.ts 5. Latency polishing tricks 6. Fixes for bugs that you might face Resources: Github Repo Link: https://github.com/murf-ai/murf-cookbook/tree/main/examples/agents/food_ordering_agentcookbook/tree/main/examples/openclaw/food_ordering_agent Murf Falcon: https://murf.ai/falcon?utm_source=SwiggyOpenClaw Murf Plugin: https://clawhub.ai/plugins/openclaw-murf-tts Companion Blog: https://dev.to/sanchita_sunil/notes-from-the-openclaw-voice-tutorial-4ngg Walkthrough Blog: https://dev.to/sanchita_sunil/i-gave-openclaw-a-voice-and-it-ordered-me-dinner-40og Openclaw: https://openclaw.ai/ Please note: The Swiggy authentication step in this video has changed. mcporter auth swiggy-food no longer works. Run node scripts/swiggy-auth.mjs instead. See the repo README for current steps.