How do you know that you have captured all the user stories you need?

Your storyboards are important tools when you design, develop, and update your skills. You use storyboards from original conception to publishing, and from certification to maintaining your published skill. When you record utterances, you try to anticipate what the user might say. The ultimate goal is to capture all possible interactions that your skill must handle. Sometimes, this goal requires more work, testing, and documentation than what you originally planned. As a rule, strive to reach a point where you can't think of any more reasonable user utterances to Alexa that the skill can’t handle. The iterative process is over only when you have exhausted all the possible options for a user.

To reach to the point of exhausting all possible options, you must consider whether you have thought of all the error cases.

For example, what happens if the user states the row first, and then the column? Or, if they say, “X” instead of "cross?" What might go wrong if the user says, “mid row” instead of “middle row?" It’s critical that you use multiple sample utterances and synonyms in your interaction model.

Consider if there is any chance that the user would say something like, “Place a cross on top, left," while they are currently assigned to the circle mark? Would it work to combine the scope of two intents? Such considerations are fundamental in Alexa skill development, and they don’t always have a single correct answer.

You'll learn more about how to design good voice experiences later in the labs.