Effectively Learning from past incidents is crucial to improving MTTR. Despite implementing blameless postmortems, runbooks, collaborative incident responses, and on-call handoff meetings, organizations struggle to effectively share and leverage collective knowledge.

In this talk, I will explore why these traditional methods fall short and how misaligned incentives (e.g., no one’s promoted for writing runbooks) contribute to locking critical expertise away in the minds of individual experts.

I think we can do much better than the current status quo. To that end, I’ve built Savvy and its open-source CLI, which uses modern LLMs to meet developers where they are and automate the manual processes of knowledge sharing.