Overview
How might we design progressive feature unlock so people can trust and act on AI output?
When to use
- Ideal for applications with extensive AI features, enterprise tools, and platforms where gradual introduction prevents user overwhelm and improves learning outcomes.
When to skip
- Expert tools where hiding features frustrates the core audience.
- When unlock criteria are opaque or feel like dark-pattern paywalls.
- Capabilities that must be visible for safety (export, delete, revoke) even for new users.
Rules
Locked UI chrome with no explanation of how to unlock.
Unlocking only via payment while marketing it as “learning progress.”
Never letting users peek at advanced features, so they cannot aspire or upgrade intentionally.
Unlocking high-risk agent write access before users understand undo and approvals.
Evidence
| Product | Implementation |
|---|---|
| ChatGPT | Gradually surfaces tools and modes after first successful chats; some via plan. |
| Duolingo-style / consumer AI tutors | Unlocks harder exercises as competence grows. |
| Cursor | Advanced agent and model options sit behind experience and plan tiers. |
| Notion AI | Introduces AI entry points after core workspace habits form. |
