AI feature discovery

Definition

Use cases

• The "empty state" problem: A user opens a new document and stares at a blank page, unsure where to begin. AI feature discovery might suggest prompts like "Start with a draft about..." or "Outline my ideas."

• The hidden power feature: A design tool can automatically generate a color palette from an image, but the feature is buried in a submenu. The system surfaces the feature when a user uploads an image for the first time.

How it's used in practice

• Contextual tooltips: Introduce AI features with non-intrusive tooltips or guided tours.

• In-app messaging: Use targeted in-app messages to announce new AI features or remind users of existing ones.

• Progressive disclosure: Show users advanced AI features only after they've mastered the basics.

• Clear value proposition: Explain AI features in simple language so users understand what the feature helps them do.

Challenges & limitations

• Balancing visibility and intrusiveness: Discovery prompts need to feel noticeable without interrupting the user’s task.

• Managing user expectations: Ensuring that users understand what AI can and cannot do.

• Measuring success: Teams often need to track whether discovery patterns actually improve feature adoption or long-term usage.

Commonly used frameworks

• Jobs-to-Be-Done (JTBD): Best for defining why users hire a feature before you teach AI to surface it.