How do you know when your use cases are complete?
Systems analysis would be a great job if we never had to deliver anything. On a project of any size, you can keep modelling forever, but (seen from outside) continued modelling has clearly diminishing returns. Stakeholders don't want to pay for unnecessary modelling -- but then again, they may consider any modelling unnecessary. So, how can you know when you have enough the use cases -- so that you can stop? Conversely, how do you know when you don't have enough use cases -- so that you don't end up revisiting obvious things that you've missed? Answers to these questions will need to convince other teams members, stakeholders, and so on, so how should you present you use case models to make these arguments?
Therefore: Draw a use case diagram to show how actors and use cases are related.
A use case diagram shows stick figures for the actors in with the system, and ovals for each use case. We write the descriptions of the users beside them, and the names of the use cases inside the ovals. The users involved in particular use cases are connected to those cases by lines. Draw a box (the ``system box'') outline around all the use cases, to make clear the boundaries of the system to be designed.
A use case diagram is quite simple, but can serve a subtle purpose. By depicting the users and the set of use cases, the diagram can be a useful focus for activities to check the use case model. The box around the use cases makes the ``black box" nature of the system clear, and the lines between users (outside the system) and use cases (within the system) highlight the ways in which users interact with the system.