The launch lands wrong, so you go back and reread the brief, looking for the ambiguity that caused it. There isn’t any. Every requirement was met. The work is still off.

The military named this problem long before software did: no plan survives contact with reality. The fix, which Chip and Dan Heath later carried into organizational life, is a single line above every plan. Commander’s intent: what must be true when this is over, whatever happens to the steps below. Instructions tell people what to do. Intent tells them what to do when the instructions stop applying.

Most handoffs carry only instructions. When reality shifts mid-build, the person holding the work can guess or come back and ask. The guess becomes a miss. The question becomes a small mark against their judgment. Either way, your conclusion is the same: next time, more detail. That closes the loop. Each added requirement buries the purpose deeper, people start reading the document for compliance instead of for the point, and the judgment you wanted has less room to show up. The briefs I’m shown after a handoff has gone wrong are rarely vague. They’re precise about everything except what the work was for.

Before your next handoff, write one sentence above the requirements: if everything below turns out wrong, this must still be true. Then test it. Ask the owner what they’d do if the core assumption breaks. If the answer is “come find you,” keep writing.

If the plan broke tomorrow, would your team know what you want, or only what you said?