The quietest automations are the most valuable

There is a pattern I keep running into when I audit ops teams. The automations that get the most internal applause (the Slack bots, the demo dashboards, the "AI agent" prototypes) are almost never the ones doing real load-bearing work six months later.

The work that survives is almost always invisible. A nightly script that reconciles two ledgers. A webhook that updates a single field on a single record. A small Python service that sits between a CRM and a billing platform and does nothing more interesting than rename three columns.

Three traits these quiet automations share: they have exactly one owner; they fail loudly when they fail; and the team can describe what they do in a single sentence. When any of those three fall away, the automation becomes a liability, usually within a quarter.

My current rule of thumb: if an automation needs a runbook longer than a Post-it note, it is probably trying to do too much.