Agentic AI – Optimizer helps organizations gain deep, actionable insights into the real-time performance of their AI agents and workflows within the Agentic AI framework on ServiceNow.
Unlike platform-level dashboards that focus on adoption or governance, this application is purpose-built for operational observability by answering questions like:
-
Which workflows are actively used and which are idle?
-
Where are failures happening at the agent, tool, or flow level?
-
Which agents are slow, underperforming, or unused?
-
How often is assist or similarity being triggered?
-
Are my AI workflows improving over time?
The built-in analytics page surfaces live metrics on workflow usage, failure patterns, execution times, assist interactions, and more. These insights help platform owners, AI product teams, and developers continuously tune and improve performance without relying on logs or scripts.
This is not a governance dashboard. It’s an operational command center for Agentic AI filling a critical gap for organizations that care not just about whether AI is adopted, but whether it’s working well.
-
Centralized Performance Dashboard: Presents a unified, real-time view of critical AI execution metrics including Published AI Workflows, Active AI Agents, and Average Execution Time. Administrators gain instant visibility into system activity and workflow readiness from a single dashboard.
-
Execution Insights & Diagnostics: Continuously surfaces operational intelligence through reports like Top Used Agents, Slowest Agent, Failed Tool Calls, and Failed Agent Calls. These insights eliminate manual analysis by automatically highlighting failure counts, execution delays, and agent usage trends.
- Workflow Usage Visualization: The Usage of Workflows visualization provides a quick overview of how frequently AI workflows are executed, helping teams track adoption and identify underutilized or overloaded processes for optimization.
This initial release introduces the Agentic AI – Optimizer application, designed to streamline the monitoring and analysis of AI workflows and agent performance within Agentic AI. The application features a unified dashboard that provides real-time visibility into key execution metrics, including Published AI Workflows, Active AI Agents, Average Execution Time, Top Used Agents, Usage of Workflows, Slowest Agent, Failed Tool Calls, and Failed Agents Calls. This release empowers administrators and developers to quickly assess system readiness, identify execution issues, and the need of improving workflow efficiency— to ensure consistent AI performance across customer environments.
Instance must be on at least Yokohama Patch 3.
Instance should have "Now Assist AI Agents" - (sn_aia) Plugin installed.