OKRs are a simple way to turn a fuzzy ambition into something you can measure. Every OKR has two parts:
OObjective — what you want to achieve. A short, qualitative, slightly ambitious headline. e.g. "Make AI a board-level capability, not a pilot."
KRKey Results — how you'll know you got there. Three to five numbers that move from a start to a target. Each one scores 0 to 1 by how far it's travelled. e.g. "Business units with a live AI use case: 1 → 6."
%The score. Average the key results to score the objective; average the objectives for the quarter. You update the current number as you go and the scores move with it.
The bit most people get wrong: 0.7 is a win, not a miss. If you're scoring 1.0 across the board, your targets were too soft. OKRs are meant to stretch — aim high, and land around two-thirds.