#case-study

The Year We Taught a Machine to Tutor

A team built an ambitious AI tutor, watched it slowly degrade under its own safeguards, paused the pilot, and rebuilt it twice — the stubbornly simple rebuild won, decisively.

Staged rollouts for agent deployments

Why "ship to 1% of traffic" doesn't map cleanly onto agents, and a four-stage rollout — shadow, sandbox, gated, general — that does.