Nudges

What happens when a tool can do a million things and the user is only aware of a handful of them?

Nudges use progressive disclosure to demonstrate AI’s capabilities while providing actionable triggers that users can use to kickstart an action.

Nudges are especially prominent in generative tools, like Writer, Jasper, or Copy, as well as knowledge management systems like Coda and Notion. They provide ideas to use AI to create value on top of content that already exists (e.g. find action items), to take care of mundane tasks (generate a summary of your meeting notes), or to remix the content and improve it (make it shorter).

Different use cases

  1. As in-app clues: AI capabilities often extend beyond what an average user imagines. Nudges here act as scaffolding. GitHub Copilot shows inline completions with subtle hints, Adobe layers AI buttons into toolbars, Figma highlights suggestions in context, and Grammarly surfaces edits as lightweight banners. Nudges in this case build confidence and reduce anxiety about where to begin.
Sample nudges from GitHub Copilot, Adobe, Figma, and Grammarly
  1. As engagement flags: Not all features are beneficial from the start. In some cases, nudges can encourage deeper engagement to expand the AI’s context of the user’s preferences or connected information, allowing it to provide personalized services. For example, MyMind’s Serendipity only appears after 25 items are saved, ensuring it can credibly provide users a view into their history. Writer and Hypotenuse both list a buffet of use cases inside the app, but gate some behind more advanced investments. Midjourney profiles require the user to first rate 200 images before the personal profile tag can be generated and used.
Hypotenuse and Writer both include a long list of use cases in their app, some of which require integrations or advanced features
  1. As feature onboarding: Companies are constantly onboarding users, especially given the fast-moving nature of this technology. Nudges can help users identify new features in a way that gives them the most value by offering an array of options. As the AI learns the user’s preferences and defaults, it may shorten the list shown or allow users to customize their menus.
The blanket list of nudges provided by Notion and Grammarly is overwhelming.

Design considerations

  • Make nudges contextual. A good nudge is tied to user intent and content state. Showing “summarize” on an empty page is meaningless. Showing it after a 1,000-word doc builds immediate value.
  • Balance discovery with restraint. Too many nudges crowd the surface and reduce trust or value. Don’t simply show users a blanket list of options. Think through the full experience and prioritize the actions that are most likely to be relevant at any given time.
  • Use nudges to build skills, not dependence. Early nudges can teach users how to find and use capabilities on their own. Avoid designing nudges that lock value behind repeated hand-holding where more advanced capabilities and deeper engagement rely on users learning the tool themselves.
  • Avoid growth theater. Nudges should not be thinly veiled upsells. If a feature is premium, make that clear up front rather than baiting users with inaccessible actions.