Conversations Guide

Conversations track all interactions between visitors and your AI chatbot. This guide will help you understand, monitor, and leverage conversation data to improve customer engagement and lead generation.

A conversation is a complete interaction between a website visitor and your AI chatbot. Each conversation contains:

• Messages: All messages exchanged (visitor and assistant)
• Visitor Token: Anonymous identifier for the visitor
• Status: Active or Ended
• Timestamps: When started, last updated, and ended
• Leads: Any leads created during the conversation

Key Benefits:

• Track visitor engagement and questions
• Identify common pain points and topics
• Measure lead conversion from conversations
• Improve chatbot responses based on real interactions
• Provide customer service insights

Conversations have two possible statuses:

Active:

• Conversation is currently ongoing
• Visitor can continue adding messages
• Chatbot will respond to new messages
• Appears in "Active Conversations" list

Ended:

• Conversation has been closed
• No new messages can be added
• Visitor can view past messages (read-only)
• Appears in "Ended Conversations" list
• Can be used for analysis and reporting

Conversation Lifecycle:

1. Visitor Opens Chat: Clicks the chatbot button on your website
2. Conversation Created: System creates a new conversation with status "active"
3. Initial Greeting: Chatbot displays "Hello! How may I assist you today?"
4. Messages Exchanged: Visitor asks questions, chatbot responds using AI
5. Lead Capture (Optional): Chatbot may create a lead if visitor provides contact info
6. Conversation Ends: Visitor closes chat or conversation becomes stale
7. Archived: Ended conversation is available for review and analysis

To View All Conversations:

1. Navigate to Leads → Conversations in the sidebar
2. See a list of all conversations in reverse chronological order
3. View conversation count badge
4. Click any conversation to see details

Conversation List Shows:

• Visitor token (anonymous identifier)
• Status badge (Active or Ended)
• Start time and last update time
• Number of messages
• Quick actions (View Details, Delete)

Click any conversation to see complete details:

Message History:

• All messages in chronological order
• Visitor messages (gray background, right-aligned)
• Assistant messages (blue background, left-aligned)
• Timestamps for each message
• Formatted content with links and lists

Conversation Metadata:

• Visitor token
• Status (Active/Ended)
• Started at timestamp
• Last updated timestamp
• Ended at timestamp (if ended)
• Associated leads (if any)

Location Information:

• City and state (from visitor's IP)
• Helps understand geographic reach
• Used in dashboard analytics

The Conversations Dashboard provides comprehensive analytics and insights.

To Access Dashboard:

1. Navigate to Leads → Conversations
2. Click Dashboard button

Dashboard Metrics:

Summary KPIs:

• Total Conversations: All-time conversation count
• Active Conversations: Currently ongoing conversations
• Ended Conversations: Completed conversations
• Average Messages: Average messages per conversation

Geographic Distribution:

• Interactive US map showing conversation locations
• Color-coded by conversation volume
• Helps identify high-engagement regions
• Click states to see details

30-Day Trend Chart:

• Line chart showing conversations over time
• Identify patterns and trends
• See impact of marketing campaigns
• Track growth over time

Lead Conversion Metrics:

• Conversations with leads created
• Conversion rate percentage
• Helps measure chatbot effectiveness
• Identify opportunities for improvement

Status Breakdown:

• Active vs. Ended conversation counts
• Visual progress bars
• Percentage distribution

Conversations can be ended manually by visitors or automatically by the system.

Visitor Ends Conversation:

1. Visitor closes the chatbot window
2. System prompts: "Are you sure you want to end this conversation?"
3. Visitor confirms
4. Conversation status changes to "ended"
5. Ended timestamp is recorded

What Happens When Ended:

• No new messages can be added
• Visitor can view past messages (read-only)
• Conversation appears in "Ended" list
• Visitor can start a new conversation anytime

The system automatically ends conversations that have been inactive for 24 hours.

How It Works:

• Background job runs periodically
• Identifies conversations with no activity for 24+ hours
• Automatically ends these "stale" conversations
• Frees up resources and keeps data clean

Manual Cleanup (Administrators):

• Run: make end-stale-conversations
• Or specify custom hours: make end-stale-conversations HOURS=48
• View stats: make conversation-stats

When the chatbot creates a lead during a conversation, it's automatically linked to that conversation.

How Lead Tracking Works:

1. Visitor provides contact information during chat
2. Chatbot uses CreateLeadTool to capture the lead
3. Lead is created with conversation_id field set
4. Lead appears in Leads dashboard with conversation link
5. Conversation shows associated lead in details

Benefits:

• Track which conversations convert to leads
• Calculate conversion rates
• Review conversation context when following up on leads
• Understand what questions led to conversions
• Improve chatbot prompts based on successful conversations

Viewing Lead-Conversation Links:

• In Conversations: See "Associated Leads" section
• In Leads: See "Conversation" link (if lead came from chat)
• In Dashboard: View conversion metrics

Visitors can request an email transcript of their conversation.

How It Works:

1. Visitor asks chatbot to email the conversation
2. Chatbot uses EmailConversationTool
3. Visitor provides email address
4. System sends formatted transcript

Email Contains:

• Complete message history (excluding system messages)
• Formatted with visitor and assistant labels
• Color-coded messages for easy reading
• Brand name and contact information

Use Cases:

• Visitor wants to save information for later
• Visitor wants to share conversation with others
• Visitor needs reference for quotes or details
• Follow-up documentation

Monitoring Conversations:

• Review conversations daily to identify common questions
• Look for patterns in visitor inquiries
• Identify topics that need better content
• Monitor conversion rates and optimize chatbot prompts

Using Conversation Data:

• Create FAQ content based on common questions
• Improve chatbot instructions based on real interactions
• Identify service areas with high interest
• Track geographic distribution of inquiries

Lead Follow-Up:

• Review conversation before calling leads
• Reference specific questions visitor asked
• Provide personalized responses based on chat history
• Use conversation context to build rapport

Privacy and Data Management:

• Conversations use anonymous visitor tokens
• No personally identifiable information unless visitor provides it
• Stale conversations are automatically cleaned up
• Delete conversations as needed for privacy compliance

Conversation won't end:

• Refresh the page and try again
• Check browser console for errors
• Verify visitor has active internet connection

Messages not appearing:

• Check that conversation is still active
• Verify chatbot agent is configured
• Check for API errors in conversation details

Dashboard shows no data:

• Verify conversations exist in the system
• Check date range filters
• Ensure location data is being captured

Lead not linked to conversation:

• Verify lead was created via chatbot (not manual entry)
• Check that CreateLeadTool is enabled for chatbot agent
• Review conversation details to see if lead creation succeeded

Now that you understand conversations, explore these related features:

• AI Agents: Learn how to configure the visitor chatbot agent
• Leads: Understand lead management and Salesforce integration
• Visitor Chatbot: Configure chatbot appearance and behavior