The Four-Question Automation Filter
This page explains the Four-Question Automation Filter and its relevance to business operations.
D06 - Technology Coherence
Definition
A four-question filter test applied before automating any task:
- Does it eliminate more than two hours per week?
- Does it require intelligence rather than judgment?
- Is the output usable with less than 20% editing?
- Do failures fail predictably rather than silently?
Most automation failures happen because the wrong task was automated — one that requires judgment, one whose output cannot be trusted without heavy editing, or one where failures surface only after they have caused real damage.
The Four-Question Automation Filter exists to catch all three failure modes before the automation is built.
Indicators
- An automation was built that still requires significant editing or oversight to be usable.
- An automated output was acted on before being verified, and the output was wrong.
- No defined criteria exist for what tasks should be automated versus better left to human execution.
Quick Win
- Pick one task you are considering automating.
- Answer these four, yes or no:
- Does it eliminate 2+ hours/week?
- Is it intelligence, not judgment?
- Is the output usable with <20% editing?
- Does it fail predictably?
- If all four answers are yes, proceed. If any answer is no, do not automate this task yet.
Comments ()