Cognitive load
The amount of thinking an interaction demands, especially remembering facts, decoding controls, and choosing among competing options.
See it
What it is
Cognitive load is the mental effort an interaction asks for. Some comes from the task itself, but the interface adds more whenever people must remember an earlier value, decode an unfamiliar icon, compare scattered options, or decide among choices that do not matter. Working memory is limited, so every needless fact competes with the actual goal. John Sweller named the idea in 1988 while watching students solve maths problems, and his split is the one worth stealing: the load the problem carries, and the load your presentation of it adds on top.
Reduce it with recognition over recall, sensible defaults, chunked forms, plain labels, visible context, and progressive disclosure. Keep related information together and let the interface remember selections, totals, and prior steps instead of assigning that job to the user.
The gotcha is confusing fewer elements with less thinking. Hiding essential controls behind ambiguous icons or nesting a short form across six screens can make a design look calm while increasing memory and navigation work. Measure the decisions and facts people must juggle, not the number of pixels on the screen.
Ask AI for it
Reduce cognitive load in this form without hiding essential context. Group related fields with fieldset and legend, keep labels visible above inputs, prefill stable values with HTML autocomplete tokens, and show a persistent order summary beside every step. Replace memory questions with visible choices, use progressive disclosure through the details element only for optional advanced settings, and keep entered values when navigating back. Remove choices that do not change the outcome, then report the number of decisions and remembered facts before and after the rewrite.