Lenny Rachitsky · 3 min read · AI & Product
🎙️ This week on How I AI: How Intercom 2x’d their engineering velocity with Claude Code
Read the original on lennysnewsletter.com →
Intercom’s Brian Scanlan describes how they doubled engineering velocity in nine months by adopting Claude Code, leveraging custom skills, telemetry, and a permission-based culture. The key is treating engineering as a product—instrumenting it, enforcing guardrails, and building foundational practices so AI can scale safely and effectively.
Key Takeaways
- Treat your engineering organization like a product: instrument everything and measure what matters with telemetry and dashboards.
- You can't scale AI adoption without visibility and strong fundamentals (mature CI/CD, test coverage, and a high-trust culture).
- Build custom skills with hooks to enforce quality at creation time (guardrails and golden-path enforcement).
- Activation energy for experimentation is cultural, not technical; leadership must give permission and assume accountability for AI-enabled work.
- Shift toward an agent-first paradigm where automation handles routine work, reducing bottlenecks and enabling faster delivery.
Excerpt
How Intercom 2x’d their engineering velocity in 9 months with Claude Code | Brian Scanlan Listen now on YouTube • Spotify • Apple Podcasts Brought to you by: Celigo —Intelligent automation built for AI Cursor —The best way to code with AI Brian Scanlan , Senior Principal Engineer at Intercom, breaks down how the company doubled engineering throughput in just nine months by going all-in on Claude Code. He shares how custom skills, deep telemetry, and a culture of permission turned AI into a true force multiplier. Biggest takeaways: Treat your engineering org like a product, and instrument everything. Intercom…