List of escalated conversations
The Escalated Conversations page displays all interactions that were transferred from Smart Assistant to a human agent. It provides a detailed overview of how and when escalations occur, helping managers evaluate service efficiency, automation quality, and customer satisfaction.
This section helps identify recurring reasons for escalation, optimize automation coverage, and improve handover experiences between the assistant and human teams.
There are three available subsections: All, Automated and Requested escalations.
All Escalations
it shows every conversation that has been escalated, regardless of how the transfer was triggered.
Automated Escalations
Occur when Smart Assistant identifies that it cannot resolve the user’s request automatically, for example due to a missing answer or unsupported intent. These escalations often happen after a series of fallback responses or when predefined conditions are met (e.g., sensitive topics, complaint detection).
Monitoring automated escalations helps assess how effectively the assistant handles requests and where improvements are needed in content or logic.
Requested Escalations
Triggered when users explicitly ask to talk to a human agent (e.g., typing “I want to speak to someone” or “Can I contact support?”). This behavior indicates that the assistant correctly detects user intent and ensures a smooth transition to an agent.
Period selection and filters
At the top of the page, you can select the Time Range you want to analyze. The available options are Last 24 hours, Last week, Last month, Last year, or a Custom range.
To refine your search, use the Filter panel to focus on specific parameters:
- Rated: displays only conversations that received user feedback.
- Client: filters conversations belonging to a specific identified user.
- Assigned to: allows filtering by the Agent responsible for the conversation.
- Language: limits results to conversations held in a specific language.
- Topic: filters by the main subject of the conversation.
- Subtopic: refines the selection within a specific Topic , helping locate the exact type of user request that led to escalation.
Filters can be combined to narrow results — for example: “French conversations about payments, escalated to agent Marie, that include user feedback.”
Click Apply to refresh the list or Reset to clear all filters and return to the complete dataset.
Conversations table
The main table displays all escalated conversations matching your selected criteria. Each row represents one Conversation that required human intervention at some point.
| Column | Description |
|---|---|
| ID | The unique identifier assigned to each conversation, used for tracking and reporting. |
| Start / End | Indicates when the conversation began and when it concluded. Useful for understanding resolution time. |
| Duration | Total time elapsed between the start and end of the conversation. |
| Queries | The number of unique user questions detected during the exchange. High numbers can indicate complex issues. |
| Messages | The total number of messages exchanged between the user, Smart Assistant, and human agents. |
| Rating | Displays the user’s satisfaction score (1 to 5). A red dot indicates that the user added a feedback comment. |
| User | The name or alias of the person who initiated the conversation. |
| User ID | The internal identifier assigned to that user in your system. |
| Escalation Type | Specifies whether the escalation was automated (triggered by the assistant) or requested (initiated by the user). |
| Assigned | Shows the Agent or team who handled the escalated conversation. |
| Language | The detected language used during the conversation. |
| Location | The geographical location inferred from user metadata. |
| Channel | The Channel where the conversation occurred (e.g., web, WhatsApp, Messenger). |
| Device | Indicates whether the user was on desktop, mobile, or tablet. |
Click on any conversation ID to open its detailed view, where you can examine the full transcript, identify the escalation trigger, and review both user and agent messages.
Tracking requested escalations provides insights into user trust, frustration points, and how well Smart Assistant manages expectations.
How to use this page
The Escalated Conversations section is designed to help you analyze service efficiency and user satisfaction in cases where human involvement was required. It serves as a bridge between automation and human support, ensuring that every handover is timely and contextually relevant.
By studying both automated and requested escalations, you can detect knowledge gaps, refine escalation rules, and maintain an optimal balance between self-service and live support.
Notes and Recommendations
- Use Topic and Subtopic filters to identify areas generating the most escalations and prioritize improvements.
- Compare Automated vs. Requested escalation ratios to understand whether users or the assistant are initiating most transfers.
- Pay attention to Rating and Duration after escalation, long conversations with low ratings may signal process inefficiencies.
- Combine insights from this page with the Containment Rate and Fallback Rate in the Conversation Statistics section to evaluate automation quality.