Analytics & Chat History

View all conversations with your AI chatbot, track usage metrics, and export data for analysis.

Accessing Chat History

Go to your app’s Insights page → Conversations tab. It has three sub-tabs:

• Chats: Browse and view conversations, and export chat data
• Calls: Phone call history
• Notifications: Email notifications when consumers start new conversations

Chat List

The left side of the Chats sub-tab shows all chat sessions with filters:

Filters

Chat Session Info

Each chat session shows:

• User name/email (if collected) or anonymous
• Source channel badge
• Preview of messages
• Timestamp
• Unread indicator

Viewing Conversations

Click a chat session to view the full transcript:

• Full message history
• User and AI responses
• Timestamps for each message
• Attachments and images shared

Facts extracted for the User Memory feature are on the Insights page → Memory tab.

Usage Metrics

The Insights page → Overview tab shows key metrics:

Below the metrics, you’ll see usage trends, Sessions by Source, and Recent Activity.

Downloading Chat Data

Download as CSV

1. Go to Insights > Conversations > Chats
2. Click the CSV button above the chat list
3. A CSV file downloads with every conversation that matches the current source and date filters

The CSV includes:

• Session ID
• User info (if available)
• Source channel
• Timestamps
• All messages with roles

Export is unavailable when Zero Data Retention is enabled for your organization.

Google Sheets Integration

Workspace owners can send chat logs to Google Sheets. First connect Google Sheets from the Integrations tab. Then, above the chat list:

• Click Google Sheets for a one-time export
• Click Auto-sync to Google Sheets and paste a Google Sheets URL to sync new chat logs every 15 minutes. Each message is added as a row.

See Google Sheets Integration for setup details.

Managing Conversations

Mark as Read

Conversations are marked as read when you view them. The unread count in the Status filter updates automatically.

Filter by Source

Use the Source filter to see conversations from specific channels, for example:

• Web: Your app’s chat page
• Widget: The embedded chat widget
• API: API integrations
• Slack: Slack bot messages
• WhatsApp: WhatsApp integration
• Email: Email conversations
• Voice: Voice conversations

Improving Your Bot

Use chat history insights to improve your AI:

1. Find common questions: Search for patterns in what users ask
2. Identify confusion: Look for long back-and-forth conversations
3. Update knowledge: Add missing information to your knowledge base
4. Refine prompts: Adjust your system prompt based on real conversations

Troubleshooting

No Conversations Showing

1. Make sure your app is published and accessible
2. Check that users have actually chatted with your bot
3. Clear all filters to see all conversations

Can’t Find a Specific Chat

1. Use the search box to search message content
2. Filter by source if you know the channel
3. Check the status filter isn’t hiding it

Download Not Working

1. Wait for the download to complete (may take time for large datasets)
2. Check your browser’s download folder
3. Try a different browser if issues persist

Using with MCP

You can access analytics programmatically using the Chipp MCP Server, enabling AI-assisted analysis and bulk operations.

Available Tools

Example Prompts

"Show me analytics for my support bot for the past week"
"Find conversations where users mentioned 'refund'"
"What are the top 10 questions users ask my FAQ bot?"
"Export all conversations from December"

Workflow: Weekly Analytics Review

You: Show me a summary of all my apps' performance this week

AI: [Calls get_app_analytics for each app]
Here's your weekly summary:
- Support Bot: 1,247 conversations, 812 unique users
- Sales Assistant: 432 conversations, 301 unique users
- FAQ Bot: 298 conversations, 255 unique users

You: What are users asking about most?

AI: [Calls get_top_queries]
Top queries across your apps:
1. "How do I reset my password?" (142 times)
2. "What are your business hours?" (98 times)
...