Edit

Share via

Customize the bot dashboard

  • Article

Note

Copilot Studio bot is renamed as Copilot agent (agent or AI agent). Human agent is now renamed as customer service representative (service representative or representative). You might come across references to the old and new terms while we update the product UI, documentation, and training content.

You can customize the out-of-the-box real-time and historical bot dashboards with additional filters and metrics to effectively visualize your bot metrics. Learn more in customize visual display.

The table describes the filters and metrics that you can add to the bot dashboards to help visualize key performance indicators (KPI).

Filters

Perform the steps in Add a filter to an entire page to add the following filters to the bot dashboard.

Title Definition Applies to Channel Data
Dialed number identification service (DNIS) Choose a customer-facing phone number from the list to see bot metrics for that number.
You can track call volumes for different campaigns or services, analyze marketing effectiveness, customize Interactive Voice Response (IVR) experiences, and generate detailed reports on call patterns, ultimately helping to optimize resource allocation and improve customer service.		 Real time and historical Voice only DimPhoneNumber: DNIS
Language Filter and view bot metrics by the last language used.
This metric helps you understand your callers' language preferences and optimize multilingual support.
For example, a conversation can start in English before the customer switches to Spanish or a conversation begins and ends in Spanish. If you select Spanish as the last language, the report displays the metrics for all conversations that ended in Spanish. In our example, the dashboard displays metrics for both the conversations.
Note: In the real-time bot dashboard, setting the Last language filter displays metrics for conversations that were escalated to an agent or an external number and are in the closed state. The metrics aren't updated when the bot conversation is ongoing.		 Real time and historical Chat and voice DimLanguage: Language

Screenshot of realtime bot dashboard with filters.

Fallback action calls

Perform the steps in Add visualizations to a report to represent FactSession : Failed bot conversation data in a Single number card visual for fallback action calls on the bot dashboard.

Title Definition Applies to Channel Data
Fallback action calls The number of conversations initiated by the customer but couldn't be connected to an AI agent due to a system failure. The application registers a call only after the classification rules in the workstream run and the work distribution system routes the call to the AI agent. This indicates that an AI agent was assigned to the call. Calls that fail before this step don't appear on the dashboard. Real time and historical Voice only FactSession: Failed bot conversation

Bot session level outcome reason

Perform the steps in add a matrix visualization to represent Session level outcome reason in a matrix visual to view metrics by outcome reason for AI agents to the report.

Title Definition Applies to Channel Data
Metrics by outcome reason The number of engaged conversations grouped by the outcome reason.
  • User exit: The number of conversations that end either because the customer ends the conversation or the session times out while waiting for the customer's response.
  • AgentTransferConfiguredByAuthor: The number of bot conversations transferred to an agent or external number based on the AI agent's configuration. For example, if an article ends with "Transfer to an agent" when the user selects "No," the AI agent transfers the conversation without the user requesting escalation, per the AI agent's business rules.
  • Resolved: The number of bot conversations that were resolved by the AI agent.
  • User error: The number of bot conversations that ended because of incorrect AI agent design.
 Historical Chat and voice Outcome reason