Leading AI at Work: Governance, Rollout and Practice / Unit 3 LAI 068 review, 2 to 3 minutes
A Feature Not Ready to Ship
The idea
A worked example, constructed for this course, shows the book's readiness rule stopping a feature that looked finished.
In practice
A software platform's team has built an AI feature that summarizes each workspace for its owner, and the release date is set. At the release review, the product manager, Lin, asks the book's question: which things does this feature touch? The engineers answer that it reads Tasks and Comments. Lin asks what it writes. After a pause, an engineer explains that the summary is saved as a new field on the Workspace, so that it loads quickly. Nobody had added that field to the Workspace's Pack, and the Workspace sits on the platform's billing matrix, because the plan tier is stored there too. Lin runs the safety checklist. The Pack is not current. The matrix has no row for the feature. Nobody has decided who reviews summaries, or what happens if a summary exposes a private comment to a workspace member who should not see it. The answer to several questions is no. The release is moved by two weeks. In that time, the team adds the summary field to the Pack, adds the feature to the matrix, decides that summaries respect each comment's visibility, and writes tests from that new card.
Where people slip
The feature worked; it was not ready. The book's rule, that a team that cannot say which things an AI feature touches is not ready to ship it, caught a write nobody had listed. Two weeks of delay cost far less than a private comment shown to the wrong person, which is the kind of failure the safety checklist exists to prevent.
Your place on the trail
Every lesson rests on earlier ones and opens later ones. Tap any to jump.
Unlocks
- Nothing yet depends on this.
Something wrong in this lesson? Flag it · Sources for Leading AI at Work: Governance, Rollout and Practice