---
sourceDocument: Xanadu Enable AI
sourceDocumentLink: https://servicenow-prod.fluidtopics.net/r/xanadu/intelligent-experiences

 Release :

    - xanadu

ft:locale :

    - en-US

ft:publication_title :

    - Xanadu Enable AI

ft:clusterId :

    - platai

bundleId :

    - platai

workflow :

    - Platform


---

# Test your model

# Test your model {#ariaid-title1}

* Release version: Xanadu
* 
* Updated August 1, 2024
* 
* ![](https://www.servicenow.com/docs/portal-asset/ico-clock) 2 minutes to read

Test your Natural Language Understanding (NLU) model against its default test set.
Testing helps determine how your model is performing with the current content.

## Before you begin

* Make sure that the NLU Model Builder - Core plugin, NLU Model Builder plugin, NLU Workbench - Advanced Features plugin and Predictive Intelligence plugin are all installed and activated.
* Have a trained model for Virtual Agent or AI Search. For more information, see [Build and train your model](https://servicenow-prod.fluidtopics.net/JsS8ENhBUq~PRpyhSM2UIg "After creating a model, build the model's content by adding intents, entities, vocabulary, and test set utterances. Your NLU model content determines how the model responds to user inputs.").
* Have a test set for testing models. For more information, see [Test set creation and management](https://servicenow-prod.fluidtopics.net/y3xFPXipXEZ~RyjkoUWSRg "Use the default test set of your NLU model to test the model's performance and accuracy. Manage your test set over time by building or updating its content in the NLU Workbench.").
* Role required: nlu_editor, nlu_admin, or admin. The editor must be assigned to the model.
{#test-your-model__ul_hbx_rzc_wrb}

## About this task

Run a test of your model using its default test set, which contains test utterances and their expected intents.

* If an expected intent in your test set does not correspond to any intent in the model, the utterances with those intents are not used for testing. They are not included in the test results.
* The mid-conversation responses of Dialog Acts can't be tried or tested in NLU Workbench.
* When the model returns no prediction for utterances which are marked as <kbd class="ph userinput">Not relevant</kbd>, that result will be counted as Correct.
* If your test set does not cover at least 60% of intents in the model, the system will not recommend a confidence threshold. However, you can still run the test.

In this example scenario, you've trained your model and want to assess the
performance.

## Procedure

1. Navigate to AllNLU WorkbenchModels.  
   The Virtual Agent tab opens by default.
2. Select the tab for your model's application, then the name of the model you want to test.
3. On the model overview's Model details tab, locate the card for the Test and publish your model phase, then click View phase.
4. Select Run new test.  
   Note:  
   Other phases are unavailable during a test.
5. The Confirm run new test dialog box opens, so select OK.  
   The test begins and the Testing is in progress... page loads.

## Result

When the test is finished, the Test and publish your model
page reloads. The Test run date field reflects the date and
time of this test.

The Overview tab displays a chart of the test results. It
also displays a list of the top 5 incorrect intents and the top 5 missed
intents.

The Detailed results tab lists all of the test utterances and
their prediction outcomes.

You can see previous test
results by clicking view the test history on Test and publish your model, or by navigating to Batch TestingTest results.

## What to do next

Use the results to edit and improve your model's content. When you are satisfied with the results, publish your model to make it available to consuming applications such as Virtual Agent.

*[\>]: and then