Knowledge work is mostly searching for information

It is tempting to start an automation project with the most visible pain: the report someone hates producing, the form everyone complains about. But the largest cost in knowledge work is often invisible, because it is spread across every task rather than concentrated in one.

The McKinsey Global Institute's analysis of knowledge work estimated that high-skill workers spend about 19% of their working week searching for and gathering information, and around 28% managing email. Nearly half the week, in other words, goes to finding and moving information rather than acting on it.

That distribution matters, because automation applied to the wrong step just makes a bad process faster. If half the week is lost to searching and re-entering information, building a slicker version of a single form barely moves the needle.

The higher-return move is usually to remove the search itself: to put the information people keep hunting for into one system where it can be queried, not reconstructed. That is a structural change, not a feature.

This is why the first phase of serious operations work is measurement, not construction. Follow the information: where it is created, where it is copied, and where people go looking for it. The map almost always shows the same thing: a handful of facts that everyone needs and no one owns.

Consolidating those facts into a governed system of record is unglamorous. It is also where the recovered time is, and it is the foundation every later automation depends on.