AI Generation

Arc uses OpenAI models to analyze your plain text descriptions and extract the underlying structure of a journey map.

How it works

Arc uses a multi-step process to ensure the generated map is both high-quality and structurally sound.

When you paste text into the Arc plugin:

1. Parsing: The AI identifies the logical stages (Phases) and individual actions (Steps).
2. Enrichment: It infers emotional state, friction points, and opportunities based on the context.
3. Validation: The output is validated against the Journey Map Schema to ensure consistency.
4. Rendering: The structured data is sent to Figma to be drawn as editable shapes.

Quality Flags

The metadata object in the generated map may include qualityFlags. These are signals from the AI about parts of the journey that might be weak, generic, or require human review.

Writing Effective Prompts

The quality of your input directly affects the quality of the generated journey map. See our Prompt Examples guide for detailed examples and best practices.

Key principles:

• Be specific: Include concrete steps rather than abstract concepts
• Add context: Mention user goals, constraints, or emotional state
• Describe phases: Natural breaks in the journey help Arc structure the map
• Include pain points: If you know where users struggle, mention it