Monitor usage, assess agent adoption, and identify consumption patterns in Agent Activity.
Requirements
- Available with any Expertise Engine package
- Access Group Permissions: Kantata AI - Agent Activity Access
Overview
Agent Activity provides visibility into agent usage across your entire account, allowing you to monitor usage, assess agent adoption, and identify consumption patterns.
Note: Account Administrators do not have access to Agent Activity by default and must be added to a custom Access Group with the appropriate permission checked.
Activity Tab
To access Agent Activity, hover over Settings, then select Agent Studio. Select the Activity tab.
You can use the filter at the top to change the period (i.e. a month) to view usage for.
Key Metrics
The key metrics at the top let you understand your account’s usage at a glance. You can quickly see whether your account is approaching its usage limit for the current period or compare usage between different periods and identify patterns.
The following metrics are displayed:
- Consumed—Amount of transactions used for the selected period and the percentage of total available transactions used. The secondary number displayed here is the amount of total available transactions for the selected period.
- Remaining—Amount of transactions remaining for the selected period. The secondary number displayed here is the number of days left in the period.
- Daily average—Average amount of transactions per day for the selected period.
-
MCP Transactions—Amount of transactions via the MCP server for the selected period. This metric is only visible if your account has access to the MCP server.
Note: Transactions from the MCP server do not count toward account usage.
Billing Period Consumption
The Billing Period Consumption chart provides a visualization of Custom Agent and Expertise Agent usage compared to the total available amount of transactions. This chart helps you understand how your organization is utilizing your account’s custom agents overall compared to the Expertise Agent.
Below the chart are the amounts for Custom Agent transactions used, Expertise Agent transactions used, and Remaining transactions, along with percentages.
To the right of the chart is the percentage consumed and total transactions used vs total transactions available. These are the same values shown in the Consumed key metric.
Agent Usage Breakdown Table
The Agent Usage Breakdown table displays usage data for the Expertise Agent and custom agents. Quickly identify which custom agents your members are seeing value from and which ones have not been adopted. The data in this table can also help you understand whether team members need further training or help you identify custom agents that can be refined or deactivated.
The following information is available in the table:
-
Agent Name
- For custom agents, the agent ID is displayed below the agent name.
-
Transactions—Total transactions used for the agent for the selected period.
- Direct—Total transactions made directly in the agent for the selected period.
- Workflows—Total transactions made through a workflow that calls on the custom agent.
- Status—Whether the agent is Active, Paused, or Deleted.
NOTE
- Agents without any transactions for the selected period are not displayed in the table.
- Transactions from creating custom agents in the Agent Studio are not included in this table.
Tip: Expertise Agent usage includes open-ended conversations that are not routed to a custom agent. This data can help you assess whether users are leveraging custom agents or are routinely engaging in open-ended chat.
Configure Columns
You can determine which columns to display in the table and in what order.
- Select the Configure Columns icon that appears to the upper right of the Agent Usage Breakdown table.
- In the modal that appears, check or uncheck columns to control which ones appear in the table.
- Drag and drop columns to change the order in which they are displayed.
- Select Save to update the view.
The view will reset if you leave the page or log out.
Comments
0 comments
Please sign in to leave a comment.