Sessions Reporting for Conversational AI on Digital Channels

Updated 

You can view reports on conversational AI session counts where a session is defined as all the messages within a 24-hour window in one conversation, counted as a single session. The next session begins from the first message received after the 24-hour window of the previous session is completed.

To plot session reporting for WhatsApp as a channel, see Session & Pricing Reporting for WhatsApp.

Enablement Note

To learn more about getting this capability enabled in your environment, please work with your Success Manager.

Create a Reporting Widget

Create a reporting widget and select the Data Source as Social Analytics.

Metrics and Definitions

Plot the following metrics and dimensions.

Metric/Dimension

Description

Account

View the number of sessions corresponding to conversations related to a specific account.

Brand Message Count

The total count of messages, encompassing those handled by both the conversational AI application and human agents in the sessions.

Bot Message Count

The total count of messages solely handled by the conversational AI application in the sessions.

Agent Message Count

The total count of messages handled exclusively by human agents in the sessions.

Session Count

The overall count of sessions in the specified time range.

Session Start Time

The timestamp indicating when each session begins.

Session Handled By

Bot: All messages in the session are managed solely by the conversational AI application.

Bot + Agent: Some messages in the session are handled by the conversational AI application, and some by Agents.

Points to Remember

  • Any filters based on case level custom fields consider the first case in the sequence.

  • Session reporting includes

    • Only Bot

    • Bot + Agent

    • Only Agent (not included)

  • Session reporting considers data only after the feature enablement. Any sessions occurring prior to this are not included in the reporting metrics.

  • There is a reporting lag of 30 minutes in the metrics. This means that the data displayed in the reports reflects a 30-minute delay from real-time activity.