# How Intercom Cut $250K/Month by Ditching GPT for Qwen Page: https://stenobird.com/podcast/chain-of-thought-ai-agents/how-intercom-cut-250k-month-by-ditching-gpt-for-qwen Text version: https://stenobird.com/podcast/chain-of-thought-ai-agents/how-intercom-cut-250k-month-by-ditching-gpt-for-qwen.md Podcast: [Chain of Thought | AI Agents, Infrastructure & Engineering](https://stenobird.com/podcast/chain-of-thought-ai-agents) Published: 2026-02-26T10:00:00+00:00 Episode link: https://share.transistor.fm/s/0fd18337 Audio file: https://media.transistor.fm/0fd18337/b333c541.mp3 Processing state: not_requested JSON: https://stenobird.com/v1/public/podcasts/chain-of-thought-ai-agents/episodes/how-intercom-cut-250k-month-by-ditching-gpt-for-qwen Duration seconds: 3211 ## Resource Intercom was spending $250K/month on a single summarization task using GPT. Then they replaced it with a fine-tuned 14B parameter Qwen model and saved almost all of it. In this episode, Intercom's Chief AI Officer, Fergal Reid, walks through exactly how they made that call, where their approach has changed over time, and how all of their efforts built their Fin customer service agent. Fergal breaks down how Fin went from 30% to nearly 70% resolution rate and why most of those gains came from surrounding systems (custom re-rankers, retrieval models, query canonicalization), not the core frontier LLM. He explains why higher latency counterintuitively increases resolution rates, how they built a custom re-ranker that outperformed Cohere using ModernBERT, and why he believes vertically integrated AI products will win in the long term.If you're deciding between fine-tuning open-weight models and using frontier APIs in production, you won't find a more detailed decision process walkthrough.πŸ”— Connect with Fergal: Twitter/X: https://x.com/fergal_reidLinkedIn: https://www.linkedin.com/in/fergalreid/Fin: https://fin.ai/πŸ”— Connect with Conor:YouTube: https://www.youtube.com/@ConorBronsdonNewsletter: https://conorbronsdon.substack.com/Twitter/X: https://x.com/ConorBronsdonLinkedIn: https://www.linkedin.com/in/conorbronsdon/πŸ”— More episodes: https://chainofthought.showCHAPTERS0:00 Intro0:46 Why Intercom Completely Reversed Their Fine-Tuning Position8:00 The $250K/Month Summarization Task (Query Canonicalization)11:25 Training Infrastructure: H200s, LoRA to Full SFT, and GRPO14:09 Why Qwen Models Specifically Work for Production18:03 Goodhart's Law: When Benchmarks Lie19:47 A/B Testing AI in Production: Soft vs. Hard Resolutions25:09 The Latency Paradox: Why Slower Responses Get More… ## Actions - request_transcript: `POST https://stenobird.com/v1/public/podcasts/chain-of-thought-ai-agents/episodes/how-intercom-cut-250k-month-by-ditching-gpt-for-qwen/transcription-requests` β€” Idempotently request low-priority transcript generation for this episode. - read_markdown: `GET https://stenobird.com/podcast/chain-of-thought-ai-agents/how-intercom-cut-250k-month-by-ditching-gpt-for-qwen.md` β€” Read the agent-friendly Markdown representation of this episode resource. A page view does not enqueue transcription. Agents should invoke `request_transcript` explicitly when they need this episode processed. ## Transcript Full transcripts are not published on public pages unless there is a clear rights basis.