Skip to content
Marketrix Docs

Interaction Modes

Agents support three interaction modes that give users different levels of assistance. Each mode serves a distinct purpose, and users can switch between them during a conversation.

Tell

Section titled “Tell”

The agent explains things conversationally. It provides text-based answers, instructions, and guidance drawn from its knowledge and instructions.

Best for: Answering questions, explaining concepts, and providing step-by-step written instructions.

Show

Section titled “Show”

The agent visually demonstrates how to complete tasks. It highlights UI elements and walks through steps directly in your application, guiding the user through the workflow.

Best for: Teaching users new workflows, onboarding walkthroughs, and showing where to find features.

Do

Section titled “Do”

The agent takes actions directly in the browser. It clicks, types, navigates, and completes tasks on behalf of the user.

Best for: Automating repetitive actions, filling out forms, and performing multi-step operations that the user would rather not do manually.

Configuring mode availability

Section titled “Configuring mode availability”

Mode availability can be configured per widget. From the widget settings, you can enable or disable Tell, Show, and Do independently. This lets you tailor the experience based on your use case:

  • A documentation widget might only need Tell mode.
  • An onboarding widget might use Tell and Show.
  • A power-user widget might enable all three modes.