How Stripe Built Kai
Stripe's internal AI assistant Kai reached 83% weekly employee adoption using LangChain's Deep Agents framework.
“the Deep Agents layer solves all non-Stripey problems, so his team could focus on Stripe's problems.”
Stripe built Kai, an internal AI assistant with 1,000+ skills and 500 tools, achieving 83% weekly employee adoption and 16x growth in a month by solving the context-window problem through skill-scoped tool loading. One engineer shipped the first working version in a week using LangChain's Deep Agents framework, which handles sandboxed code execution and S3-backed session memory. The story is a concrete enterprise case study showing how agentic context management unlocks large-scale internal AI deployment.