---
title: Evaluate agentic AI assets
description: Agentic evaluations enable you to test agentic AI assets against defined datasets to verify quality before deployment. Run results score performance, identify issues, and provide opportunities for optimization.
locale: en-US
canonical_url: https://www.servicenow.com/docs/r/australia/intelligent-experiences/agentic-evals.html
release: australia
topic_type: reference
last_updated: "2026-03-18"
reading_time_minutes: 1
breadcrumb: [AI Agent Studio \(legacy\), Enable AI experiences]
---

# Evaluate agentic AI assets

Agentic evaluations enable you to test agentic AI assets against defined datasets to verify quality before deployment. Run results score performance, identify issues, and provide opportunities for optimization.

## Get started

<table id="table_gyk_3rj_q3c" class="nav-card presentation"><tbody><tr><td>

[Explore\[Omitted image "bus-explore.svg"\] Alt text:Explore what agentic evaluations are and general guidelines for running them.](https://raw.githubusercontent.com/ServiceNow/ServiceNowDocs/australia/markdown/australia/intelligent-experiences/explore-aia-eval.md)

</td><td>

[Evaluate\[Omitted image "bus-integration-and-apis.svg"\] Alt text:Create evaluation runs, track their progress, and analyze and act on results.](https://raw.githubusercontent.com/ServiceNow/ServiceNowDocs/australia/markdown/australia/intelligent-experiences/evaluating-aia.md)

</td><td>

[Reference\[Omitted image "bus-learn.svg"\] Alt text:Learn more about the roles, parser tool for custom metrics, and the results page.](https://raw.githubusercontent.com/ServiceNow/ServiceNowDocs/australia/markdown/australia/intelligent-experiences/ref-aia-eval.md)

</td></tr></tbody>
</table>Automated evaluations for agentic AI are a structured quality assurance capability for agentic AI assets built in AI Agent Studio. Each evaluation run tests your agents against a defined dataset, applies LLM-powered judges to score key quality dimensions, and surfaces issues with recommended fixes. Use automated evaluations to get objective, explainable evidence that your agentic AI is ready to deploy.

## Helpful resources

Some ServiceNow resources that can provide you with helpful information are:

-   **\[Omitted image "dcx-icon-community.svg"\]ServiceNow Community**

    [ServiceNow Community](https://community.servicenow.com/community)

-   **\[Omitted image "dcx-icon-dev-portal.svg"\] Developer**

    [developer.servicenow.com](https://developer.servicenow.com/)

-   **\[Omitted image "dcx-icon-impact.svg"\] Impact**

    [http://impact.servicenow.com](http://impact.servicenow.com)

-   **\[Omitted image "dcx-icon-learning.svg"\] ServiceNow University**

    [ServiceNow University](https://learning.servicenow.com/now/lxp/home)

-   **\[Omitted image "dcx-icon.svg"\] Best Practices**

    [Best Practices](https://mynow.servicenow.com/now/best-practices/home)

-   **\[Omitted image "dcx-icon-partner.svg"\] Partner**

    [https://www.servicenow.com/partners.html](https://www.servicenow.com/partners.html)

-   **\[Omitted image "dcx-icon.svg"\] ServiceNow**

    [http://servicenow.com](http://servicenow.com)

-   **\[Omitted image "dcx-icon-store.svg"\] ServiceNow Store**

    [http://servicenow.com](http://servicenow.com)

-   **\[Omitted image "dcx-icon-support.svg"\] Support**
    -   [https://support.servicenow.com/now](https://support.servicenow.com/now)
    -   [Known Error Portal](https://support.servicenow.com/kb?id=known_error_portal)