NLU Workbench - Advanced Features

NLU Workbench - Advanced Features expands the functionality of NLU Workbench to help you manage and improve your models.

Summary usage

The NLU Workbench - Advanced Features application includes the following features:

• Multi-model Batch Testing: Test large groups of utterances against your NLU models to see how the model predicts intents.
• Cross-model Conflict Review: Identify conflicting intents within or across models to improve model performance.
• NLU Model Performance: Use NLU Model Performance to see how well your models predicted intents in Virtual Agent based on end-user confirmation.
• NLU Expert Feedback Loop for VA: Provide feedback on Virtual Agent chat log utterances to help the system continuously learn and to better predict user input.
• Irrelevance detection in NLU: train your Virtual Agent models to ignore utterances that are not relevant.
• Issue Auto Resolution Tuning in NLU (Expert Feedback Loop for IAR): provide feedback to your ITSM Issue Auto Resolution model to improve its predictions.

After the application is installed and activated, Model Performance, Expert Feedback Loop for VA, and Expert Feedback Loop for IAR appear under All > NLU Workbench. Multi-model Batch Testing and Cross-model Conflict Review appear under All > NLU Workbench > NLU Advanced Features.

Installation

NLU Workbench - Advanced Features is available from the ServiceNow Store. For instructions on how to purchase and download, see Install NLU Workbench - Advanced Features.

Note:

The available Intent Discovery ServiceNow Store application is installed separately from NLU Workbench - Advanced Features.

Related features

• Install NLU Workbench - Advanced Features
  You can install the NLU Workbench - Advanced Features application (com.snc.nlu.workbench.advanced) if you have the admin role.
• Multi-model Batch Testing
  Test multiple Natural Language Understanding (NLU) models against a large set of utterances to evaluate the performance of the models. Add test sets, test multiple models, and see test results.
• Cross-model Conflict Review
  Identify conflicting intents within or across models so you can take corrective actions, resolve such conflicts, and improve your NLU model performance.
• NLU Model Performance
  Use NLU Model Performance to see how well your models predicted intents in Virtual Agent (VA) based on end-user confirmation.
• NLU Expert Feedback Loop
  Provide feedback on Virtual Agent chat log utterances to help the system continuously learn and to better predict user input.
• Issue Auto Resolution Tuning in NLU
  Use the NLU Workbench homepage to support Issue Auto Resolution (IAR) tuning in NLU.
• Intent Discovery
  Use the Intent Discovery application to help identify opportunities for incident deflection. For example, you can use it to identify which Virtual Agent conversations to activate next.