AI Onboarding
Introduce an AI feature by showing what it can do with the user's own data, setting honest expectations about limits and privacy.
6 states4 edge casesDesktop + Mobile
01The problem
Most users meet an AI feature as an empty box with a sparkle icon. They don't know what it's good at, whether it can see their data, or whether its answers can be trusted. Some try one vague prompt, get a vague answer and never return. Others over-trust it. First use determines both adoption and appropriate trust.
02When to use
Use when
- Launching a new AI feature to existing users
- First-time use of an assistant or copilot
- AI features that use workspace or customer data
Avoid when
- AI that works invisibly behind a familiar control (e.g. smarter search ranking) — no ceremony needed.
03Recommended UX
Statecraft recommendation
On first use, show three or four example tasks built from the user's real context that run in one click, a single line on what the AI can access, and a plainly stated limitation.
04Desktop behaviour
On desktop, the first-use state replaces the empty assistant panel with contextual examples, a data-access line and the composer ready.
Desktop rules
- 3–4 example tasks generated from the current page or workspace.
- One line: what it can access, and a link to details.
- One honest limitation in plain language.
- No multi-step tour; a dismissible hint at most.
05Mobile behaviour
On mobile, keep first use to a short list of tappable examples and the access note.
Mobile rules
3 mobile rules are included in Pro.
Statecraft Pro
Unlock the full pattern
Get the complete interaction flow, states, edge cases, mobile rules and implementation prompt.
- Interaction flow · 5 steps
- 3 mobile rules
- 6 states
- 4 edge cases with recommendations
- 4 microcopy comparisons
- Accessibility notes
- Developer notes
- Claude Code / Codex implementation prompt
Edge cases covered in this pattern
- No data yet
- Admin hasn't enabled AI
- Data residency concerns
- First result is bad