Design interactions around visible objects.
Build interactions on objects that users can see and understand, not hidden or abstract commands.
Contents 3. Interaction Mechanics & Agency 3. Direct Manipulation
How action is executed.
“Can users of your system act directly on visible objects with safe, reversible results?”
“Can users act directly on what they see, and undo it if needed?”
“Do actions feel immediate and reversible?”
“Can users see something, act on it, and instantly see the result?”
“Does it feel like direct manipulation rather than issuing commands?”
“Can users explore without fear of permanent mistakes?”
“Is undo easy and always available?”
“Does the system respond right away?”
“Is the impact of an action clear before it is finalized?”
“Does AI support user control rather than override it?”
“Can automated actions be paused or stopped?”
Build interactions on objects that users can see and understand, not hidden or abstract commands.
Reduce mental translation required to achieve outcomes.
Ensure users understand what will be affected prior to execution.
Maintain tight feedback loops during user manipulation.
Preserve the perception of control through timely system response.
Avoid demanding fine motor accuracy for low-risk or common actions.
Ensure users do not unknowingly switch behavioral contexts.
Provide reliable undo or recovery mechanisms.
Design systems so undo replaces excessive confirmation dialogs.