💬 Conversation Model
Track customer conversations with AI-powered sentiment analysis.
Auto-created with Intercom
When you connect Intercom, this model is automatically created and configured!
Overview
The Conversation Model tracks conversations across chat, email, and phone channels. Like support tickets, Meza AI automatically analyzes conversation sentiment to understand customer satisfaction.
Conversations provide real-time insight into customer mood and can surface issues before they escalate to formal support tickets.
Key Fields
| Field | Type | Description | Required |
|---|---|---|---|
conversation_id | String | Unique conversation ID | Yes |
account_id | String | Customer account | Yes |
user_id | String | User who initiated | No |
channel_type | String | email, chat, phone | No |
subject | String | Conversation subject | No |
created_at | Timestamp | When conversation started | Yes |
sentiment | String | AI-detected sentiment | Auto |
Channel Types
Meza AI supports multiple conversation channels:
Chat
Live chat, in-app messaging
Email conversations
Phone
Call transcripts
AI Sentiment Analysis
Meza AI analyzes conversation content to determine overall sentiment. For multi-message conversations, sentiment is calculated from the aggregate of all messages.
- Positive — Customer expresses satisfaction, thanks
- Neutral — Questions, normal inquiries
- Negative — Frustration, complaints, escalation language
Conversation Metrics
Once configured, Meza AI tracks:
- Conversation Count — Total conversations per account
- Response Time — Average time to first response
- Sentiment Distribution — Positive vs negative ratio
- Active Conversations — Currently open conversations
- Channel Breakdown — Distribution by channel type
✅ Tip