AI glossary
Stop rule
A stop rule is a rule, agreed before a trial starts, that says when to stop if it isn’t working: for example, if errors rise above a set level, or time saved stays below a target after a set number of weeks. Setting it in advance stops a failing pilot running on hope.
Think of it as
Deciding your top price before the auction.
Why it matters at work
Once people have put effort into an AI project, it’s hard to call it off; a rule agreed beforehand makes that call easier and fairer. Module 10 builds one into every 90-day test.