User Stories — Part 2: Writing Stories & Acceptance Criteria

maang.io AI-Native Builder Series

What is this?

Picture ordering a pizza over the phone and saying only "make it good." When it arrives, can you complain? Not really — you never said what "good" meant. Now picture ordering it like this: "When the pizza arrives, the cheese should be fully melted, there should be exactly eight slices, and it must be here within 30 minutes — or it's free." Now there's no argument. Either the cheese is melted or it isn't. Either it's eight slices or it's seven. You turned a vague wish into something checkable.

That checklist on the pizza order is exactly what acceptance criteria are for a user story. A story says what a user wants. The acceptance criteria say precisely how you'll know it was delivered correctly — the specific, testable conditions that have to be true before you can call the story "done." Without them, "done" is just a feeling, and feelings can't be verified. With them, you (and your AI cook) have a target you can actually hit and check.

The one-line idea: acceptance criteria turn a story's fuzzy "done" into a list of specific, testable conditions — so "is it finished?" becomes a yes/no you can actually answer.

The melted-cheese-in-30-minutes image is the spirit of this whole chapter: make "done" so concrete that anyone — including an AI — can check it. Let's earn it.