Automation gets talked about as a spectacle. In practice, the automation that matters is almost invisible: a report that assembles itself, a notification that fires at the right moment, data that moves between systems without anyone re-typing it.
Treated as infrastructure rather than novelty, automation stops being a demo and starts being something the business quietly runs on.
Automate the friction, not the org chart
The best candidates for automation are the repetitive, rule-based steps between a trigger and a result — the parts of work that are necessary but add no judgment. Removing that friction gives people back the time to do the parts that do require judgment.
AI extends the range of what can be automated, but the principle is the same: apply it where it removes a real, understood cost — not where it looks impressive.
Reliability is the whole point
An automation that works most of the time creates a new kind of work: checking whether it worked. Anything that runs unattended has to be observable, recoverable and trustworthy, or it isn't saving anyone anything.
