Skip to main content
Documentation
Documentation
Last Updated: 2025-11-14
Add Models from the Custom AI Model Library

Add Models from the Custom AI Model Library

Your account must have Custom AI enabled. You can request it by contacting your Customer Success Manager.
The Custom AI Model Library is a searchable, categorized repository of over 120 pre-configured Custom AI models. This feature offers instant access to fully tested models across functions like Legal, HR, Finance, IT, and Sales, saving time by eliminating the need to manually build fields.
You can browse models tailored to specific contract types and industries, preview their details, and publish them instantly, or first save them as a draft for testing and custom refinement.
  1. Click
    Automation
    .
  2. On the
    Fields
     tab, click
    Browse Model Library
    .
  3. Search for models or use the filters to browse categories.
  4. Click on the model(s) you would like to use and click
    Add Model
     to add it to your cart. Alternatively, click
    Add All Models
     to bulk-add models from a category.
  5. Set the scope.
    Choose between
    Apply to All
     (applies filters to all selected categories), or
    Set Custom Scope
    , (defines a scope for each category).
  6. Define which documents the models should run on by adjusting the available filters.
  7. Click
    Save & Test
     to test the models before publishing, or click
    Confirm & Publish
     to publish the models to your environment.
    Note: Publishing models from the library affects your account balance.
Publishing a model automatically creates a corresponding data field under the
Data Fields
 section in the
Admin
 tab.
If a field with the same name already exists in your environment, Workday appends
(Custom AI Library Model)
 to the name of the new field added with the model.
Once you publish a model or save it as a draft, you can
Add Instructions
 to the model to provide additional guidance and customize it to your context and documents. These instructions don’t replace or overwrite the existing library model instructions.
After evaluating 10 samples, a model added from the Library is eligible for
AI Boost
 to improve accuracy based on customer feedback.