Defaulting every AI capability to a chat interface ignores the physical and cognitive realities of users. This piece argues for intentional modality selection — matching input and output methods to user intent, environment, and cognitive load. It introduces two practical frameworks: a Task Audit (covering input constraints, output constraints, social constraints, and cognitive load) and an Input/Output Alignment Matrix that maps user intent to optimal modality combinations. A real-world case study of field technicians on high-voltage electrical grids illustrates how voice input and audio output replaced touch-based interfaces, reducing diagnostic time by 20%. A downloadable Modality Task Audit Field Template is provided to help teams document physical barriers before writing code.
