The first automation paradox: why it takes longer than doing it manually

Your first automation will likely take longer to build than the manual work it replaces.

This surprises many people. They calculate the minutes, multiply by annual repetitions and say: here is the business case. Then building it takes three times that.

It does not mean the business case was wrong. It means the first automation is not just about saving time.

Why the first one takes longer

A process that feels simple is not always simple once you start writing code.

Edge cases surface. Exceptions nobody mentioned. That one user or customer who does things differently from everyone else. An integration where the API does not behave the way the documentation suggests.

There is also the learning: a new tool, a new API, new error handling patterns. All of this takes time.

A McKinsey figure that says more about expectations than results

McKinsey's November 2025 research found that only 39% of companies report actual business-level impact from their automation.

The number sounds bleak. But it says more about expectations than results.

Most failures do not happen because the automation did not work technically. They happen because the wrong problem was chosen. A process was automated that was not properly understood. Or something was automated where the manual effort was actually faster than building the automation.

What the first automation really is

The first automation is an investment in understanding just as much as it is an investment in saving time.

Building it forces you to understand the process better than you ever did before. Edge cases get resolved. Undocumented exceptions surface. Ambiguities need to be settled.

These lessons make every subsequent automation faster and better.

How to choose the right first automation

The first automation is worth choosing carefully. Repetitive: it happens at least weekly, preferably more often. Mechanical: it requires no judgement or creativity, just the same steps each time. Error-prone: humans make mistakes doing it, automation does not. Bounded: a clear start and end, not too many dependencies.

If you expect pure ROI from the first attempt, you will be disappointed. If you expect the next one to go much faster and better, you will get that reliably.

Want to map out the right starting point for automation? Get in touch.