A company stretched to its limit usually hesitates between two expenses: hiring someone, or investing in tools. Both can be the right call. Both can be costly mistakes. What separates them is neither budget nor trends: it's the nature of the work to be absorbed.

Step 1: describe the work, not the job

Before making any decision, write down what is actually done over one week: the tasks, how often they occur, how long they take, and the decisions they involve. The exercise is tedious, and it is decisive. It almost always reveals that a substantial share of the time goes into moving information around: looking things up, retyping, passing things on, chasing people.

Without this description, every decision rests on a hunch. And a hunch is expensive: a failed hire costs you several months, an ill-suited tool costs you in abandonment.

Step 2: eliminate before you automate

A good share of repetitive tasks shouldn't exist anymore. Reports nobody reads, approvals that have never blocked anything, double data entry inherited from a tool no one uses. Automating them sets a pointless habit in stone.

Automating a pointless task makes it permanent.

Step 3: decide based on the nature of the task

  • Repetitive task, clear rules, high volume → automation. A machine doesn't get tired and doesn't copy the wrong line at 6 p.m.
  • Task requiring judgment, variable context, relationship at stake → human. No system knows how to negotiate a deadline or win back an unhappy customer.
  • Mixed task → the human decides, the machine prepares. This is the most common case, and the most profitable: sorting and preparation are automated, the final call stays human.

The “we'll deal with it later” trap

Putting off the decision has a silent cost: the team compensates. It holds on through extra hours and stopgap fixes, until the day someone leaves and you discover the process only existed in their head. Losing unwritten knowledge is one of the most common and most expensive accidents in a growing company.

Building systems is therefore not only about saving time: it makes the company less fragile. A process that is described and equipped with the right tools survives a departure. A habit doesn't.

Describe one week of real work

Describe one week of real work, cross out whatever can disappear, and sort the rest into “clear rules” and “judgment required.” By the end of the exercise, you won't be choosing between hiring and automating anymore: you'll know what to automate, and exactly what you need someone for.

The right question is not “do we need a human or a machine?” but “where exactly does human judgment create value?”.