← back
Shipping an Enterprise Voice AI Agent in 100 Days - Peter Bar, Intercom Fin
Original: Shipping an Enterprise Voice AI Agent in 100 Days - Peter Bar, Intercom Fin
Takeaway
Voice agents ship fast by reusing chat-side RAG and starting with the out-of-hours voicemail-replacement wedge, but conversation design must be re-thought for sub-second latency.
Summary
- Intercom shipped Fin Voice in 100 days as a frontline phone-support teammate; chat-side Fin already handles 56% avg resolution (70-80% for some customers) across 5k+ customers.
- Voice rationale: >80% of support teams use phone, 1/3 of global support is voice, human calls cost $7-12 while AI is 5x cheaper, plus 24/7 multilingual coverage.
- Architecture combines STT→LLM→TTS chain with OpenAI Realtime API at the core, plus reused chat-side RAG and existing telephony; out-of-hours deploy used as low-risk wedge.
- Conversation design adapts to voice: target ~1s latency for simple queries, different turn-taking and backchanneling than chat—voice isn't 'chat with sound'.
voicecustomer-supportintercom
Original description
What does it take to go from blank page to live enterprise voice agent in 100 days? That’s the challenge we took on with Fin Voice at Intercom. Enterprise customer service demands high-quality, reliable voice interactions - but delivering that fast means wrestling with tough problems like latency, hallucinations, voice quality, and answer accuracy. We rapidly evaluated and integrated a full voice stack - including transcription, language model, text-to-speech, retrieval-augmented generation, and telephony - while designing tools that fit seamlessly into existing human support workflows. In this session, I’ll share key lessons from our accelerated development of Fin Voice. We'll explore the technical and operational hurdles we faced, the trade-offs we made, and how we built deployment and handover tools that work for customer service teams. You'll leave with insights into building AI-driven voice products that are both powerful and practical. About Peter Bar I’m Peter Bar, a Product Lead with over 10 years of experience in the tech industry. At Intercom, I’m responsible for Voice AI initiatives and led the development and launch of Fin Voice, our AI voice agent. My background spans both B2B and consumer tech, blending technical depth with strategic product leadership. Before Intercom, I drove growth and customer experience efforts at Deliveroo (food delivery) and worked on music discovery products at Shazam. I hold a Master’s degree in Computer Science from Imperial College London. Recorded at the AI Engineer World's Fair in San Francisco. Stay up to date on our upcoming events and content by joining our newsletter here: https://www.ai.engineer/newsletter