AI Admin Center Performance Explorer dashboard

  • Release version: Australia
  • Updated July 30, 2026
  • 2 minutes to read
    • Summarize
    Summarized using AI
    This content was generated using new OpenAI-powered functionality. Results are provided on an as is basis and are not guaranteed to be accurate or complete.

    Summary of AI Admin Center Performance Explorer dashboard

    The AI Admin Center Performance Explorer dashboard enables ServiceNow customers to monitor and analyze detailed execution data for AI assistants and AI agents deployed across their organization. It provides granular insights into individual executions, helping to evaluate performance, identify trends, and troubleshoot AI asset behavior.

    Show full answer Show less

    Key Features

    • Two focused sub-tabs:
      • Assistants tab: Displays execution records for AI assistants, with filters for date, assistant name, result type, conversation end state, deflection outcome, and deflection state. Users can drill down by selecting an assistant name to view full execution details.
      • Agents tab: Lists executions of AI agents with filters for date, execution state, latency, and agent name search. This view highlights tool calls, large language model (LLM) calls, and latency metrics.
    • Execution details and metrics:
      • For assistants, metrics include conversation end state, inferred customer satisfaction (CSAT), transfer/escalation status, assist actions, deflection outcomes, and effort scores reflecting user interaction effort.
      • For AI agents, metrics cover execution state, number of tool and LLM calls, end-to-end latency, individual latency contributions, assist credits consumed, and inferred CSAT scores.
    • Filter and reset options: Allow users to narrow down execution lists based on relevant criteria and reset filters to default views.

    Practical Benefits

    ServiceNow customers can leverage this dashboard to:

    • Investigate specific assistant or agent execution details to understand behavior and outcomes.
    • Analyze performance trends over time using latency, satisfaction, and outcome metrics.
    • Identify opportunities to improve AI assistant deflection success and reduce live agent escalations.
    • Monitor resource consumption such as assist credits and tool calls for cost and efficiency management.

    Overall, the AI Admin Center Performance Explorer dashboard equips organizations with actionable insights to optimize AI deployments, enhance customer satisfaction, and maintain operational efficiency.

    Use the AI Admin Center Performance Explorer dashboard to review and analyze the execution details of assistants and AI agents across your organization.

    AI Admin Center Performance Explorer dashboard

    The AI Admin Center Performance Explorer dashboard displays execution-level details for assistants and AI agents. Use the dashboard to investigate individual executions, analyze performance metrics, and identify patterns across your AI asset deployments.

    The Performance Explorer dashboard includes two sub-tabs: Assistants and Agents. Each sub-tab displays a table of individual executions for the selected asset type.

    Assistants

    The Assistants tab displays a list of individual assistant executions. Use the Date, Assistant Name, Result Type Offered, Conversation End State, Deflection Outcome, and Deflection State filters to narrow results. Select Reset Filters to clear all applied filters.

    Figure 1. Assistants tab of the AI Admin Center Performance Explorer dashboard
    Assistants tab showing a table of assistant executions with columns for Assistant Name, Executed On, Result Type Offered, Conversation End State, Inferred CSAT, Deflection Outcome, Deflection State, and Effort Score.
    Assistant Name
    The name of the assistant that was executed. Select the assistant name to view the full execution record.
    Executed On
    The date on which the assistant execution occurred.
    Result type Offered
    The type of result that the assistant offered during the execution, such as an answer, a deflection, or a transfer.
    Conversation End State
    The state of the conversation at the end of the execution, such as Open or Faulted.
    Inferred CSAT
    The inferred customer satisfaction score for the execution, calculated based on conversation signals.
    Transfers and escalation
    Indicates whether the conversation was transferred or escalated to a live agent during the execution.
    Assist
    The number of assist actions performed by the assistant during the execution.
    Deflection Outcome
    The outcome of the deflection attempt, indicating whether the conversation was successfully deflected.
    Deflection State
    The state of the deflection for the execution, such as deflected or not deflected.
    Effort Score
    A score reflecting the level of effort required by the user to complete the interaction, based on conversation signals.

    Agents

    The Agents tab displays a list of individual AI agent executions. Use the Date, State, and E2E Latency (S) filters, or the Search Agent field, to narrow results. Select Reset Filters to clear all applied filters.

    Figure 2. Agents tab of the AI Admin Center Performance Explorer dashboard
    Agents tab of the Performance Explorer dashboard, showing a table of AI agent executions with latency, call, and CSAT columns.
    Agent Name
    The name of the AI agent that was executed.
    Executed On
    The date on which the AI agent execution occurred.
    State
    The state of the AI agent execution, such as Completed or Terminated.
    Tool Calls
    The total number of tool calls made by the AI agent during the execution.
    LLM Calls
    The total number of large language model (LLM) calls made by the AI agent during the execution.
    E2E Latency (S)
    The end-to-end latency of the execution, in seconds, measured from the start to the completion of the AI agent run.
    Tool Latency
    The cumulative latency contributed by tool calls during the execution.
    LLM Latency
    The cumulative latency contributed by LLM calls during the execution.
    Assists Consumed
    The number of assist credits consumed by the AI agent during the execution.
    Inferred CSAT
    The inferred customer satisfaction score for the execution, calculated based on interaction signals. See Exploring Conversation Insights for more information.