Skip to main content

Prompt Examples

This guide provides practical, tested prompt examples for various AI agent use cases. These examples are based on the suggested messages and prompts visible in the agent configuration interface.

Knowledge Base Agents

Based on the initial suggested messages shown in the configuration interface, knowledge base agents commonly use these prompt patterns: Initial Suggested Messages

Core Knowledge Base Prompts

The interface shows these fundamental knowledge base prompts:
  1. “Give me a summary of this knowledge base”
    • Perfect for users who want an overview
    • Helps users understand what information is available
    • Provides a starting point for deeper exploration
  2. “Can you provide an overview of the main topics covered here?”
    • More structured than a simple summary
    • Organizes information by topic areas
    • Good for users who want to navigate to specific areas
  3. “What categories or sections does this knowledge base cover?”
    • Provides organizational structure
    • Helps users understand how information is categorized
    • Useful for systematic exploration

Extended Knowledge Base Prompts

Building on the core patterns shown in the interface:

System Prompt Templates

Knowledge Base Agent System Prompt

Based on the context instruction shown in the interface (“Use the following pieces of context to answer the question at the end”):

Customer Support Agent System Prompt

Sales Assistant System Prompt

Industry-Specific Examples

Healthcare Assistant

Professional Avatar Example

Technical Support Examples

Educational Content Assistant

Conversational Flow Examples

Opening Conversation Patterns

Based on the ”👋 Hi, how can I help you?” pattern shown in the interface:

Clarification Prompts

When users need guidance:

Transition Prompts

Moving between topics smoothly:

Voice-Optimized Prompts

For voice-enabled agents, prompts should be conversational and natural: Voice Configuration

Voice-First System Prompt

Voice User Prompts

Form Integration Prompts

For agents with data collection capabilities: Data Collection Form

Lead Qualification Prompts

Data Collection Flow

Multi-Modal Integration

Scene-Specific Prompts

Different visual environments may call for different conversational styles: Scene Selection

Testing Your Prompts

Evaluation Criteria

When testing prompts, consider:
  1. Clarity: Do users understand what the agent can do?
  2. Engagement: Do the prompts encourage interaction?
  3. Relevance: Are suggested prompts appropriate for the use case?
  4. Flow: Do conversations progress naturally?
  5. Completeness: Do responses satisfy user needs?

A/B Testing Approaches

These examples provide a foundation for creating effective prompts across different use cases. Remember to customize them based on your specific domain, audience, and objectives while maintaining the conversational patterns that work best for your users.