Manage learning instances

A learning instance is an IQ Bot structure that describes the data in a single document type, for example an invoice. It contains information such as the language used in the document and the fields to extract.

A learning instance can be either in staging or production mode. These modes determine whether the learning instance is a draft or live version.

Staging: In the staging mode, you can create and edit learning instances, upload sample documents, and train bots to extract data from the sample documents you uploaded. You can test and improve the accuracy of a learning instance and its bots before deploying it to production mode.

Note: Because the data generated by learning instances in the staging mode is based on sample documents, it is kept separate from data pertaining to learning instances in production mode. For example, sample documents are not counted against your organization's purchased pages. For more information, see Review the dashboard.
Production: In the production mode, a learning instance can automatically process real business documents, extract data, and send documents requiring manual validation to the validation queue.

To find out which mode a learning instance is in, navigate to the LEARNING INSTANCES tab, where the My learning instances page displays the learning instances in the environment. Each learning instance has a toggle in the Actions column, which indicates whether it is in staging (toggle to the left and is colored gray) or production (toggle to the right and colored orange).

A learning instance with the toggle in staging mode and a learning instance with the toggle in production mode

Based on the mode that a learning instance is in, different user types contribute to IQ Bot. For example, in the staging mode, the Data engineer creates learning instances and the Bot builder builds supporting RPA bots, and in the production mode, the Validator manually corrects the fields in the documents that IQ Bot could not process. To learn more about how each user contributes to IQ Bot, see IQ Bot process overview.

Note: The following are the file size restrictions in each stage:

In staging, you can upload a maximum of 150 documents of 10 MB file size per learning instance.
In production, you can upload a maximum of 50 MB file size per document. However, the maximum number of documents allowed per learning instance depends on the license.
There are no limitations on the number of pages per document in a pdfbox OCR.
You can upload 60 pages per document in an image-based OCR.
With the Tesseract4 OCR, there is a known limitation that restricts the number of pages per document to fewer than 60 pages.

The learning instance lifecycle

The staging and production modes are independent from the IQ Bot environment when it comes to the traditional DevOps workflow of separate development, testing, and production environments. A learning instance can be in either mode in any of the environments in the development life cycle.

Note: When you transfer a learning instance, it remains in the mode in which it was exported. For example, if you export a learning instance in staging mode from the development environment and import it to the testing environment, the learning instance will still be in staging mode and it will not perform document processing automatically.

The following image shows the typical lifecycle of a learning instance:

Lifecycle overview of a learning instance

You can transfer learning instances across environments by using the Migration Utility. This helps you to avoid having to recreate similar learning instances and better manage the lifecycle of the associated bots. To learn more, see Migrate learning instances.

Automation 360