Tell Mode — Answering Questions
Searches your knowledge base and provides clear, helpful answers to user questions. How it works: The agent finds relevant information from your knowledge base and generates a response. Best for: Customer support questions, product information, policy explanations, general inquiries. Examples: “What are your business hours?”, “How do I reset my password?”, “What’s your return policy?”Show Mode — Step-by-Step UI Guidance
Highlights UI elements on the page and provides step-by-step guidance to walk users through tasks. How it works: The agent identifies the relevant UI elements (buttons, links, form fields) and highlights them sequentially, guiding the user through each step of a workflow. Best for: Tutorials, feature demonstrations, user onboarding, how-to instructions. Examples: “Show me how to create an account”, “Show me how to use this feature”, “Show me how to update my profile”Do Mode — Browser Automation
Performs browser automation on behalf of the user, including clicking buttons, typing in fields, scrolling, and navigating between pages. How it works: The agent takes direct control of browser interactions — clicking, typing, scrolling, and navigating — to complete tasks for the user. Best for: Form filling, account setup, process automation, complex multi-step workflows. Examples: “Help me fill out this form”, “Complete the checkout process for me”, “Set up my account”Choosing the Right Mode
| Situation | Recommended Mode |
|---|---|
| User needs information | Tell |
| User needs to learn how to do something | Show |
| User wants a task done for them | Do |
| Customer support agent | Tell + Show |
| Onboarding agent | Show + Do |
| General assistant | Tell + Show + Do |
Configuring Modes
Tell mode is always enabled. Show and Do modes are optional and can be enabled in your agent settings:- Go to your agent’s configuration
- Select “Interaction Modes”
- Enable the modes you want
- Save your settings
Best Practices
- Tell Mode: Ensure your knowledge base is comprehensive and well-organized
- Show Mode: Run simulations so the agent learns your actual UI elements and workflows
- Do Mode: Ensure user consent before performing actions, provide clear progress feedback, and handle errors gracefully
Next Steps
- Create agents — Build agents with the right modes for your use case
- Test in the playground — Verify each mode works correctly
- Run simulations — Help agents learn your UI for Show and Do modes