The Interactions page shows conversations between users and AI Agents.
It helps review user messages, assistant responses, interaction metadata, sentiment, CSAT, topics, and technical details for each conversation.
Overview
Use the Interactions section to:
- Review chat and voice transcripts
- Check interaction status
- See the assigned AI Agent
- View model and usage details
- Analyze sentiment and CSAT
- Review topics, summaries, and keywords
- Inspect customer environment details
Interaction Transcript
The transcript displays the full conversation between the user and the assistant.
Each message includes:
| Field | Description |
|---|
| Role | Shows whether the message was sent by the user or assistant. |
| Timestamp | Shows when the message happened inside the interaction. |
| Message | Full text of the user or assistant message. |
| Sentiment | Displays detected user sentiment when available. |
Interaction Details
The details panel contains metadata about the selected interaction.
| Field | Description |
|---|
| Agent | AI Agent used in the interaction. |
| Organization | Organization where the interaction belongs. |
| LLM Model | Model used to generate assistant responses. |
| Type | Interaction type, such as Text Chat or Voice Call. |
| Status | Current interaction status, for example Finished. |
| Messages Count | Total number of messages in the interaction. |
| Billable Messages | Number of messages counted for billing. |
| Start Time | Date and time when the interaction started. |
| Duration | Total interaction duration. |
| Browser | User browser. |
| OS | User operating system. |
| IP Address | User IP address. |
| Country | Detected user country. |
| City | Detected user city. |
| URL | Page or platform URL where the interaction happened. |
| Total Costs | Estimated interaction cost. |
| CSAT | Customer satisfaction score. |
| CSAT Description | Explanation of the detected satisfaction score. |
Topics Summary
The Topics Summary section provides an AI-generated overview of the interaction.
It may include:
| Field | Description |
|---|
| Topic | Main conversation topic. |
| Summary | Short summary of the interaction. |
| Sentiment | Detected customer sentiment. |
| Sentiment Description | Explanation of the detected sentiment. |
| Keywords | Important keywords extracted from the conversation. |
Sentiment
Sentiment helps understand the customer’s emotional tone during the interaction.
Possible sentiment values may include:
- Positive
- Neutral
- Negative
Sentiment is generated automatically and should be used as a support signal, not as the only source of evaluation.
Notes
- Interaction details may vary depending on chat, voice, or integration type.
- Some fields may be empty if the data was not available.
- Costs are calculated based on platform billing logic.
- CSAT, sentiment, topics, and summaries are generated automatically.
