Start with the question
Before choosing a camera, an engine or a format, name the thing you want to discover. Can this character carry a scene? Does this world feel inviting? Can someone understand this service in a few seconds? A useful first version answers one question well.
Make the smallest thing that reveals it
For a film, that might be a scene on paper and a sequence of reference frames. For an interactive world, it might be one space with one meaningful action. For an identity, it might be a single application that tests the whole visual language.
Look at what people actually do
A prototype is a conversation. Watch where people pause, what they notice and where they need help. Keep the parts that do their job. Change the parts that need explaining. Progress comes from making the next decision clearer.
Keep the work honest
A study is a study. A finished production has different responsibilities. Naming that difference makes it easier to invite feedback, collaborate fairly and choose the next step.