Human-AI Collaboration

Intent Preview

Before any significant action, the agent presents a clear, scannable summary of what it intends to do - showing planned steps, reversibility status, and edit controls for user approval.

What is Intent Preview?

The Intent Preview pattern addresses a core anxiety in agentic AI: users need to understand what will happen BEFORE it happens. Unlike traditional AI where the user explicitly types a prompt and evaluates the response, agentic actions may be initiated proactively or involve consequences that are difficult to reverse - sending emails, booking flights, modifying files. This pattern shows a clear, scannable summary of planned actions using plain language (not technical jargon), with each step marked for reversibility and editable by the user. The preview must be sequential for multi-step operations, highlight irreversible actions visually, and never auto-dismiss. This transforms the approval moment from a binary yes/no into a structured review that builds trust and catches misunderstandings before they cause harm.

Problem

When an agent is about to take a multi-step action, users need to understand what will happen before it happens. Without an intent preview, users experience anxiety leading to constant monitoring or blind trust that erodes at the first mistake.

Solution

Before any significant action, present a clear, scannable preview showing planned steps in plain language, with reversibility indicators, edit controls for individual steps, and explicit approve/reject buttons. Never auto-dismiss the preview.

Real-World Examples

Implementation

AI Design Prompt

Guidelines & Considerations

Implementation Guidelines

1

Use progressive disclosure within the preview itself - show the summary first ('3 actions planned'), let users expand for detail.

2

Highlight irreversible actions visually with different color or warning icon. Sending an email deserves more visual weight than drafting a document.

3

Include a time estimate. 'This will take approximately 2 minutes' sets expectations for multi-step operations.

4

For recurring tasks, offer 'always approve this type' to graduate to higher autonomy over time.

5

Never auto-dismiss a preview. The user must explicitly approve, even if it's just a single tap.

6

Make individual steps editable - users should be able to modify, reorder, or remove steps, not just approve or reject the entire plan.

7

Use plain language descriptions, not technical jargon. Show 'Cancel your flight to San Francisco' not 'Execute cancel_booking(id: 4A7B)'.

Design Considerations

1

Preview-to-approval rate: how often users approve without modifications indicates preview quality

2

Preview-to-edit rate: how often users modify the plan indicates the preview caught misunderstandings

3

Preview skip rate: if users start skipping previews, consider graduating them to higher autonomy

4

Balancing preview frequency with user fatigue - not every action needs a full preview

5

Determining the right level of detail for each action type to avoid information overload

6

Ensuring the preview accurately reflects what will happen to maintain trust

7

Supporting both quick-scan and deep-review modes for different user contexts

Want More Patterns Like This?

Get 6 essential AI design patterns (free PDF) + weekly AI/UX analysis

One-page PDF for design reviews + weekly AI/UX analysis. Unsubscribe anytime.

Related Patterns

About the author

Imran Mohammed is a product designer who studies how the best AI products are designed. He studies and documents AI/UX patterns from shipped products (36 and counting) and is building Gist.design, an AI design thinking partner. His weekly analysis reaches thousands of designers on Medium.