First Run Experience

Follow the steps below to access Document Understanding in Automation Suite:

1. Activate the license
2. Enable AI Center on the tenant
3. Create a project on AI Center
4. Upload the model base images
5. Create a data labeling session
6. Launch the data labeling session

Activate the License

Sign in to Automation Suite, then go to Admin > Licenses and input your license key. More information on how to access Automation Suite can be found here.

Upon successful activation, under Robots & Services, AI robots quantity should not be 0. A specified number of pages are shown for Document Understanding after clicking the Generate new button.

Enable AI Center on the Tenant

Please go to Admin > Tenants, click the three dot menu and select AI Center as a provision service under Tenant Setting.

Create a Project on AI Center

To create a project on AI Center, click AI Center on the left rail, and then click Create Project from the top bar.

Enter a project name and click Create.

Upload the Document Understanding ML Packages

Note: This step is required only for offline installation. For online installation, skip this step as all the models should already be under ML Packages > Out of the box Packages > UiPath Document Understanding.

Upload the Document Understanding ML Packages to AI Center. You need to upload only the models you want to use.

• Detailed instructions can be found here.

If you are using Form Extractor, Intelligent Form Extractor, and Intelligent Keyword Classifier, you can directly access the model without uploading it to AI Center. Detailed instructions can be found here.

Create a Data Labeling Session

Create a data labeling session to use Data Manager to label your dataset and retrain the base model.

To do so, go to Data Labeling > UiPath Document Understanding.

Name the data labeling session, enable the Create new dataset slider, name the dataset and then click Create.

This may take a few minutes to deploy. Please refresh the page if the status does not change.

Launch the Data Labeling Session

Launch the newly created data labeling session.

You should see an empty session such as the one below.