AI is approving our pull requests: Here's how we made it safe - The Intercom Blog
intercom.com · 1,771 words · saved by 1 readers
We're producing more code than ever at Intercom. Here's how we're safely using AI for PR approval.
At Intercom, shipping is our heartbeat. We push code to production hundreds of times a day. Engineers, engineering managers, designers, and PMs all contribute to this, safely. The average time from merging code to it running in production is 12 minutes. We’ve long held a belief that might sound counterintuitive: speed is not the enemy of safety. It’s a prerequisite for it. Accumulating code creates risk. Shipping small batches minimizes it. The faster you ship, the smaller each change is, and the easier it is to catch problems, and roll back when something goes wrong as the context is…
saved by
related reading
- crawshaw - 2026-05-07crawshaw.io
- AI is removing the middle class of software engineeringblog.florianherrengt.com
- Building a software factory for AI SDKvercel.com
- When AI builds itself \ Anthropicanthropic.com
- I built an AI code review agent in a few hours, here's what I learnedsourcebot.dev
- Auto-review of agent actions without synchronous human oversightalignment.openai.com
- Your job is to deliver code you have proven to worksimonwillison.net
- Code was our medium for thoughtwattenberger.com
- A Practical Guide to Becoming an AI-Native Engineerblog.bytebytego.com
- Now – Updates from the Linear teamlinear.app
- AI code reviews for complex codebases | cubic | Try free nowcubic.dev
- Using AI to write better code more slowly | Read the Tea Leavesnolanlawson.com