Amazon Q Business Analytics dashboard metrics - Amazon Q Business

Amazon Q Business Analytics dashboard metrics

The Amazon Q Business dashboard provides a comprehensive view of key metrics to help admins understand the performance and usage of every Amazon Q application environment from the Amazon Q Business console.

The following analytics are available in the dashboard for the specified time period.

  • Average daily active users

    This tile gives the average number of unique daily users for the application environment that participated in at least one chat session for the chosen period.

    Note

    This metric is calculated using the Maximum statistic. For more information, see CloudWatch statistics definitions.

  • Total users

    All the users of this application environment.

  • Weekly active users

    This tile gives the total number of unique users for the application environment who participated in at least one chat session for the week in the chosen period.

  • Monthly active users

    This tile gives the total number of unique users for the application environment who participated in at least one chat session for the month in the chosen period.

  • Total conversations

    This tile gives the total number of chat conversations that have occurred in this application environment. Conversations are different from individual chat queries and represent a collection of related queries and responses over a relevant session or subject.

  • Total queries

    This tile gives the total number of chat message queries received by this application environment.

  • Average queries per user

    This tile gives the average of daily chat queries for divided by the daily active unique users for that application environment for the chosen period.

  • Average queries per conversation

    This tile gives the average of total chat queries to an application environment divided by the total individual conversations in the application environment. Conversations refer to new chat threads, which are listed on the left navigation pane of the web experience for the application environment.

  • Total plugins

    This tile shows the total number of plug-ins currently used by this application environment.

  • Most frequently used plugin

    This tile shows the plug-in most frequently used by this application environment.

  • The Overview trend

    This line chart displays two trend lines - one showing the daily active users for each day for the application environment, and another showing the total daily individual chat queries received by the application environment.

  • The Customer feedback trend

    This line chart displays three trend lines for the application environment.

    • The number of thumbs up feedback.

    • The number of thumbs down feedback.

    • The number of chat message responses that did not receive any feedback thumbs up or thumbs down).

  • The Thumbs down reasons chart

    This pie chart represents the different feedback reasons that users choose when providing a thumbs down feedback, along with the count or percentage of times each reason was selected. You can also filter by feedback reasons to further see the top queries that resulted in the feedback or choose view details to resolve issues to see all the queries for further evaluation.

    • The Thumbs-down feedback queries

      This page shows the list of queries that resulted in an unsuccessful response for further evaluation. You can choose any query to view details and see Recommendations on how to possibly resolve the issue.

      Note
      • A maximum of 5000 query responses will be loaded initially for evaluation. You can always search for specific queries.

      • The empty columns in some rows indicate that you need to update the Vended log deliveries configuration in Admin-controls-guardrails page.

  • The Unsuccessful query responses chart

    This pie chart represents the breakdown between unsuccessful query responses because the answer was not found or the answer was blocked due to your guardrails settings. You can also filter by response type further to see the top queries that resulted in the response or choose view details to resolve issues to see all the queries for further evaluation.

    • The Unsuccessful response queries

      This page shows the list of queries that resulted in an unsuccessful response for further evaluation. You can choose any query to view details and see Recommendations on how to possibly resolve the issue.

      Note
      • A maximum of 5000 query responses will be loaded initially for evaluation. You can always search for specific queries.

      • The empty columns in some rows indicate that you need to update the Vended log deliveries configuration in Admin-controls-guardrails page.

  • The Queries trend

    This chart displays three trend lines.

    • The number of chat queries that included a file upload.

    • The number of instances where the chat queries was blocked by the guardrails set.

    • The number of instances where the application environment was unable to find a relevant answer to the user's query.

  • The Plugins chart

    This bar chart shows the number of times users have selected each of the available plugins per day. For example, ServiceNow, Salesforce, JIRA, Zendesk, Custom, etc. when sending a query to the application environment.

  • The Conversations chart

    This bar chart shows the total number of conversations per day for the specified date range. Conversations are different from individual chat queries, and represent a collection of related queries and responses over a relevant session.

  • The Average queries per conversation trend

    This bar chart calculates the average number of queries per conversation by dividing the total number of daily queries by the total number of conversations per day for the application environment. Conversations are listed on the left navigation pane of the web experience, and each conversation represents a collection of related chat messages. The chart indicates the average length of conversations, as measured by the total number of chat messages from the first to the last message in a given conversation thread.