Agent Builder - Advanced Configuration¶

Advanced features and power-user settings for fine-tuning your AI agent.

Overview¶

This guide covers advanced configurations for power users who want to enhance their agent with:

• Dynamic behavior based on user context
• Custom personality traits
• Voice processing capabilities
• External tool integrations
• Advanced analytics and quality monitoring

For basic agent setup, see Agent Builder - Basic Configuration.

Advanced Configuration Sections¶

1. Pre Processors and Tags - Dynamic tagging and pre-processing
2. Instructions (Advanced) - Conditional instructions, special instructions, current message prompt
3. Agent Parameters - Dynamic variables for instructions
4. Prompt Shortcuts - Quick actions for users
5. Personality Traits - Agent personality configuration
6. Documentation - User-facing agent documentation
7. Behavior - Advanced behavior settings
8. Audio and Speech - Voice processing configuration
9. Context - Advanced context management
10. Tools - External integrations and capabilities
11. Post Processors - Response analysis and evaluation

1. Pre Processors and Tags¶

Overview¶

Pre Processors execute JavaScript code before the user query is sent to the LLM. They can dynamically set tags based on user context, authentication status, session data, or any custom logic.

Pre Processors and Tags configuration

How Pre Processors Work¶

Workflow: 1. User sends a message 2. Pre Processor executes (JavaScript snippet) 3. Pre Processor returns tags based on logic 4. Tags determine agent behavior, shortcuts visibility, tool availability 5. Query is sent to LLM with tag context

Adding Pre Processors¶

Pre Processor added to agent with tag configuration

Steps:

1. Configure Pre Processor in Settings (see Settings > Processors)
2. Click "Add" button in Pre Processors section
3. Select configured pre processor from list
4. Configure tags returned by the pre processor

Use Case Example: Authentication-Based Behavior¶

Pre Processor Logic:

// preprocessor example
const subscriberid = ld.get(data, 'user_data.subscriberid', null);
if (subscriberid !== "") {
  ctx.addTag('authenticated');
} else {
  ctx.addTag('unauthenticated');
}

Tag Configuration:

• Tag: authenticated → Shows "View My Account", "Order History" shortcuts
• Tag: unauthenticated → Shows "Sign Up", "Learn More" shortcuts

Benefits:

• Dynamic agent behavior based on user state
• Personalized experience
• Controlled feature access

2. Instructions (Advanced)¶

Overview¶

The Instructions section in Agent Settings provides advanced instruction management beyond the basic Agent Instructions and Initial Message.

Special Instructions¶

Special / Additional instructions to be used in the agent. These are typically set by the Co-Admin role but can also be set by Owner, Admin, Workspace Admin, and Co-Admin.

Use {special_instructions} as a placeholder in your Agent Instructions to control where special instructions are inserted.

Each Special Instructions field supports Conditional Instructions (see below).

Current Message Prompt¶

A secondary mechanism to steer the agent's responses. The Current Message Prompt allows you to append or prepend instructions to each user prompt message.

• To create a dynamic prompt, use {message} as a placeholder in your text
• If {message} is not included, the current message will be appended at the end automatically

Each Current Message Prompt field supports Conditional Instructions (see below).

Conditional Instructions¶

Conditional Instructions allow you to apply different instruction blocks based on tag criteria. For example, a separate instruction can be applied when the user is "Authenticated" vs. "Unauthenticated."

Conditional Instructions apply to Agent Instructions, Special Instructions, and Current Message Prompt.

Key Features:

• Tag-based conditional instruction blocks
• Support for common instructions preceding conditional blocks — reducing redundancy
• Backward compatible with existing instruction sets
• Use the "Insert Condition" button to add a condition block
• Use the "Verify" button to validate the condition syntax

End of Conversation Message¶

When enabled, this message is sent exactly as written when the conversation ends due to system idle timeout. The system injects an internal user turn with the tag internal_end_of_conversation, allowing the agent to perform final actions.

Add Instruction¶

Click "Add Instruction" to add additional instruction blocks. This allows multiple instruction sets to be configured for the same agent.

3. Agent Parameters¶

Overview¶

Agent Parameters are dynamic variables that can be used within agent instructions using {botParams.<key>} syntax.

Agent Parameters configuration

How to Use¶

Step 1: Define Parameters

Add key-value pairs in the Agent Parameters section:

company_name: fifthelement.ai
product_name: AI Agent Platform
support_hours: 9 AM - 6 PM EST
support_email: support@fifthelement.ai
documentation_url: https://docs.fifthelement.ai

Step 2: Reference in Instructions

Use parameters in your instructions:

You are a support agent for {botParams.company_name}.
Our product, {botParams.product_name}, is an enterprise AI platform.

If users need help:
- Support hours: {botParams.support_hours}
- Email: {botParams.support_email}
- Documentation: {botParams.documentation_url}

Benefits: - Easy updates without touching instructions - Reusable agent templates - Multi-environment configuration (dev, staging, prod)

4. Prompt Shortcuts¶

Overview¶

Prompt Shortcuts are predefined prompts or questions displayed as buttons in the end-user chat interface. By assigning a short label and an actual prompt for each button, Agent admins can expedite interactions and steer users through conversations with pre-made options.

Only the label is visible to the end user — not the prompt behind it.

Shortcut Configuration Fields¶

When to Display This Prompt¶

Control when each shortcut is visible to users:

Static Context for Shortcuts¶

When a static context is provided for a shortcut, it overrides the agent's configured context. The response will be based solely on the static context provided with the shortcut.

User Defined Shortcuts¶

End users can create their own shortcuts for messages they use frequently. This feature must be enabled by the Agent admin.

How end users create shortcuts:

1. Within the chat interface, hover over or select a message from chat history
2. A + icon appears on the right side of the message
3. Click or tap and hold the icon to create a shortcut
4. The agent auto-generates a short label based on the message
5. The shortcut appears at the end of the chat area

Managing User Shortcuts:

• Edit: Click and hold the shortcut to edit the label inline
• Delete: Click the trash can icon that appears when editing
• Mobile: Long press on a message to create a shortcut

AI-Generated Prompt Shortcuts¶

After each agent response, the system can analyze conversation context and automatically generate relevant follow-up suggestion buttons alongside static shortcuts.

Configuration (under Prompt Shortcuts):

Available Template Variables:

Initial Session Tagging¶

A system-level tag internal_no_message is automatically applied at conversation start and removed on the first user message. This can be used to trigger initial prompt shortcuts before a user has typed anything — for example, displaying a welcome menu of options.

5. Personality Traits¶

Overview¶

Personality Traits allow you to fine-tune your agent's conversational style and tone. Select up to 5 traits from various personality dimensions.

Personality Traits configuration with multiple trait dimensions

Available Traits¶

Each dimension has three options: Left Extreme, None (Neutral), Right Extreme

Example Configurations¶

Maximum 5 traits can be selected to avoid conflicting personalities.

6. Documentation¶

Overview¶

The Documentation section allows you to create user-facing information about your agent. This content is displayed when users open the agent in Web App modality via the "How to use this agent - Quick Guide" button.

Documentation configuration with rich text editor

Documentation shown to users in Web App modality

What to Include¶

User Documentation Label: - Title displayed above the documentation - Example: "Quick Start Guide", "Agent Capabilities", "How to Use"

User Documentation (Rich Text):

Recommended Content:

1. Agent Purpose:

This agent helps you find information about our products,
troubleshoot common issues, and connect with support.

2. What It Can Do:

✓ Answer product questions
✓ Provide troubleshooting steps
✓ Search documentation
✓ Escalate to human agents

3. How It Works:

1. Ask your question in natural language
2. The agent will search our knowledge base
3. Receive instant answers with source links
4. Use shortcuts for common tasks

4. Best Practices:

• Be specific in your questions
• Use shortcuts for faster results
• Provide error messages when troubleshooting
• Rate responses to help improve the agent

5. Limitations:

This agent cannot:
- Process refunds (contact support@company.com)
- Modify account settings directly
- Provide pricing quotes (contact sales)

Welcome Screen HTML¶

Replaces the previous blank state shown on a new chat with a configurable Welcome Screen. Admins author the landing experience for an agent — a greeting, a short description of the agent's purpose, documentation, suggested starter questions, or anything else they want users to see first. Applies to Web App and Sidebar modalities. Other modalities (Floating Embed Widget, etc.) are unaffected.

Documentation section with the Welcome Screen HTML field alongside User Documentation Label and User Documentation — Preview, Use template, and Generate with AI affordances are exposed directly on the field

Configuration¶

The Welcome Screen HTML field sits inside the Documentation section of Agent Settings, alongside the existing User Documentation Label and User Documentation fields. Admins author arbitrary HTML — without being constrained to a fixed tile layout.

The ? icon next to the field documents allowed tags and provides sample code.

Author Affordances¶

Preview mode renders the authored HTML — greeting, agent purpose, and clickable starter questions appear exactly as the end user will see them. The ← Edit link returns to the raw HTML editor Generate welcome screen modal — describe what the welcome screen should emphasize, click Generate, and the AI writes the full HTML into the field

End-User Experience¶

• Saved Welcome Screen HTML renders as the initial state of every new chat on the Web App and Sidebar modalities, above the message input.
• Any clickable links or starter questions authored in the HTML behave as the admin defined them — for example, underlined sample questions that submit on click.
• Leaving the field empty preserves the existing default empty state — no regression for agents that don't configure this.

Welcome Screen as the initial empty state of a new chat — Web App (left) and Sidebar (right). Clickable starter questions submit on click; the message input remains active below

7. Behavior¶

Overview¶

The Behavior section contains toggles and settings that control how your agent interacts with users, handles data, and integrates with external systems.

Debug and Development¶

Language Model Control¶

User Experience¶

User Capabilities¶

When Upload to Repositories is enabled:

• Configure Allowed Repositories for AI Response Upload — select one or more repositories where users can upload AI responses
• Only repositories where users have write access will be available at runtime

Context and Title Generation¶

Conversation Management¶

Security and Encryption¶

Encrypt Conversation:

When enabled, only users who participated in the conversation can see messages in clear text. Supervisors monitoring conversations will see encrypted values.

Analytics¶

GTM Id (Google Tag Manager Id):

Enter a GTM ID to enable analytics data transmission and website insights. This tracks agent events like widget open, message sent, AI response received, new conversation created, and sharing of conversation transcripts.

Email Notifications¶

Administrators can customize notification email templates from Settings > Email Settings > Create Email Template for both "Conversation Flag" and "Message Dislike" notifications. Placeholders for dynamic content (conversation ID, user comment, disliked message text) are available.

8. Audio and Speech¶

Overview¶

Configure real-time voice processing capabilities for your agent, enabling voice interactions via SIP (telephony) or WebSocket (web-based).

Audio and Speech configuration section

Configuring audio and speech settings

Realtime Voice Processing¶

Enable Realtime Voice Processing:

Once enabled, you can configure:

SIP Realtime Voice Processing:

• For telephone/SIP integration
• Inbound/outbound call handling

Websocket Realtime Voice Processing:

• For web-based voice interactions
• Browser-based voice chat

Enable SIP or Websocket realtime voice processing

Configuration Details¶

Refer to on-screen documentation for detailed configuration of: - Audio codecs - Voice models (ASR/TTS) - Latency settings - Call routing - Interruption handling

9. Context¶

Overview¶

The Context section allows you to add information that your agent can reference when generating responses. Context can be provided as text, documents, or knowledge repositories.

Context configuration with Text, Document, and Knowledge Repository tabs

Context Types¶

Text:

• Direct text input
• Token count displayed
• Example: Product specifications, company policies

Document:

• Upload documents (PDF, DOCX, TXT, etc.)
• Processed and embedded for retrieval
• Example: User manuals, technical documentation

Knowledge Repository:

• Connect to pre-configured knowledge bases
• Vector search enabled
• Example: Full documentation sites, FAQs, support articles

Context Options¶

Limit Context Token:

• Restrict context size to prevent exceeding model limits
• Recommended for cost control

Allow User Provided Context:

• Let users upload documents per session
• Use Case: Custom document analysis

Important: Add 'Context Tool' under the 'Tools' section to enable context usage.

10. Tools¶

Overview¶

Tools extend your agent's capabilities beyond conversation, enabling it to retrieve data, perform actions, and integrate with external systems.

Tools configuration section

Adding tools to the agent

Adding Tools¶

Steps:

1. Click "Add" button in Tools section
2. Select a tool from available options:
3. Knowledge Retrieval - RAG: Access knowledge bases
4. Transfer to Human: Escalate to human agents
5. API Integration: Call external APIs
6. Context Tool: Use context from Context section
7. Custom Tools: JavaScript-based custom tools
8. Configure tool settings:
9. Scope Limiting: Set tags to control when tool is used
10. Instructions: Add guidance for tool usage
11. Parameters: Configure tool-specific settings

Tool Configuration Options¶

Scope Limiting (Tag-Based):

• Control tool availability based on tags
• Example: "Transfer to Human" only for tag needs_escalation

Tool Instructions:

• Provide specific guidance on how/when to use the tool
• Example: "Use this tool when user asks for order status"

11. Post Processors¶

Overview¶

Post Processors analyze agent responses by sending them to an LLM for evaluation, quality checks, or additional processing.

Post Processors configuration

Types of Post Processors¶

Turn Post Processor:

• Executes after each user-agent exchange
• Analyzes individual responses in real-time

Use Cases: - Sentiment analysis per response - Compliance checking - Quality scoring - Real-time flagging

Conversation Post Processor:

• Executes after the entire conversation ends
• Triggered when conversation expires (after Conversation Inactivity Duration)

Use Cases: - Overall conversation summary - CSAT prediction - Conversation categorization - Post-conversation analytics

Configuration¶

Steps:

1. Configure Post Processors in Settings (see Settings > Processors)
2. In Agent Builder, navigate to Post Processors section
3. Click "Add" to select configured post processors
4. Choose processor type:
5. Turn Post Processor
6. Conversation Post Processor
7. Save configuration

Summary¶

You've now configured advanced features for your AI agent:

• Dynamic behavior with Pre Processors
• Conditional and special instructions
• Reusable parameters
• User-friendly shortcuts with conditional visibility
• Custom personality
• User documentation
• Advanced behavior settings
• Voice capabilities
• Knowledge context
• External tool integrations
• Quality monitoring with post processors

Don't forget to click "Save" to apply all configurations!

Related Topics¶

• Agent Builder - Basic Configuration - Essential setup steps
• Agent Reports - Monitor agent performance and analytics
• Agent's Look & Feel - Customize branding and appearance
• Cloning an Agent - Duplicate and modify existing agents
• Version Control - Manage agent versions and rollback
• Knowledge Overview - Configure knowledge bases for RAG
• Settings > AI Models - Manage available language models
• Settings > Processors - Configure pre/post processors