Progressive disclosure

Progressive disclosure is a way of handling complexity by splitting features into layers. The first layer holds what most people need most of the time. Everything else sits one step away, behind an "Advanced" link, an expandable panel or a secondary screen. Jakob Nielsen describes it as deferring advanced or rarely used features to a secondary screen, making applications easier to learn and less error-prone.

Why it matters

Showing every option at once makes people slower and more likely to make mistakes. Progressive disclosure helps two groups at once. According to Nielsen, it helps novices prioritize their attention and avoid mistakes with settings they do not need yet, and it saves experienced users time because they do not have to scan past a long list of features they rarely use. The Interaction Design Foundation puts it simply: it declutters the UI to prevent confusion and cognitive overload. That makes it one of the most practical ways to reduce cognitive load without removing capability.

How to apply it

Nielsen's two-step recipe is: initially, show users only a few of the most important options, then offer a larger set of specialized options upon request.

Do:

• Use real usage data or task analysis to decide what goes in the first layer. Nielsen stresses that you must disclose everything users frequently need up front.
• Label the way forward clearly ("More filters", "Advanced settings") so people know what they will find.
• Keep it shallow. Nielsen notes that designs beyond 2 disclosure levels typically have low usability because users get lost.
• For a meeting-notes summariser, show the summary first with "Show action items" and "Show full transcript" one click away.

Don't:

• Hide something people need often just to make the first screen look clean.
• Bury critical warnings, costs or consent choices in the second layer.
• Mix progressive disclosure up with staged disclosure. Staged disclosure (like a wizard) walks people through a linear sequence of steps. Progressive disclosure gives optional access to secondary features.

Common mistakes

• Getting the split wrong. If the primary layer misses a frequent task, users dig through menus every time and the pattern backfires.
• Unclear triggers. A tiny chevron or an icon with no label is easy to miss, so the advanced layer might as well not exist.
• Too many levels. Nested "More" menus inside "Advanced" panels lose people.
• Forgetting state. If someone opens the advanced panel every time, remember that choice.

In AI products

AI tools tend to accumulate settings: model choice, temperature, tone, length, memory, data sharing, connected tools. Most people never need most of them. A good pattern is a simple default (one prompt box, one sensible model) with advanced controls tucked behind a clearly labeled panel for people who want them.

Progressive disclosure also applies to output. A long generated answer can lead with a short summary and let people expand detail, sources or the model's reasoning on request, a pattern covered under AI reasoning disclosure. The exception is anything that affects trust or consent: what data is used for training, what an agent is about to do, and what it costs should stay in the first layer.