Agent Behavior Configuration
The Behavior tab allows you to customize how your AI agent interacts with users, including conversation prompts, initial messages, and suggested responses.
Agent Prompt Configuration
System Instructions
- Agent Prompt: Define the core instructions that guide your agent’s behavior
- Use the prompt template system or write custom instructions
- Example: “Use the following pieces of context to answer the question at the end.”
Knowledge Integration
- Agents can access and utilize knowledge bases during conversations
- Configure how the agent should reference and apply knowledge
- Set up context-aware responses based on available information
Initial Conversation Setup
Welcome Messages
- Initial Agent Messages: Set the first message users see when starting a conversation
- Default example: ”👋 Hi, how can I help you?”
- Customize to match your agent’s personality and purpose
Suggested Prompts
Configure suggested questions that appear to guide users:- “Give me a summary of this knowledge base”
- “Can you provide an overview of the main topics covered here?”
- “What categories or sections does this knowledge base cover?”
Message Flow Options
Interactive Elements
- Show initial messages as tooltip: Display welcome messages in a non-intrusive way
- Remove ‘Powered by Aivah’: Clean up the interface for branded experiences
- Hide microphone: Disable voice input if not needed
User Experience Settings
- Configure message display preferences
- Set up conversation flow patterns
- Customize interaction styles based on your use case
Agent Identity
Naming and Branding
- Agent Name: Set a custom name for your agent (e.g., “Product Presenter”)
- Companion Name: Define the companion identity
- Avatar Selection: Choose from available avatars or upload custom ones
Voice and Personality
- Companion Voice: Select Gemini voices (Sportsman, Customer support, Sarah, etc.) or OpenAI Realtime voices (Alloy, Echo, Shimmer, Ash, Ballad, Coral, Sage, Verse, Cedar, Marin) based on the active model
- LLM Model: Choose the underlying model family (OpenAI GPT 4.1/5 series, Gemini 2.5 series, OpenAI Realtime, or Groq-hosted models)
- Configure personality traits through prompt engineering
Best Practices
- Clear Instructions: Write specific, actionable prompts
- User-Friendly Welcome: Create welcoming initial messages
- Helpful Suggestions: Provide relevant suggested prompts
- Consistent Identity: Maintain consistent naming and voice
- Purpose-Driven: Align behavior with intended use case