Configuring Business Information Extraction in ESM – Optional

Objective

After completing this lesson, you will be able to configure and activate Business Information Extraction scenarios in SAP ESM and enable intelligent data extraction from case content using Generative AI.

Configuring Business Information Extraction

This lesson explains how to configure Business Information Extraction scenarios using Generative AI within SAP Enterprise Service Management (ESM). It guides administrators through the necessary steps to activate and set up extraction scenarios, map relevant Case fields, and link business documents to Case types. These configurations enable the automatic extraction of key data from Case content and attachments, streamlining business processes such as memo requests, dispute management, and more.

Note

Pre-requisite:

The following section is optional and only applicable if your organization has subscribed to the SAP AI Units license.

Configuring Business Information Extraction

The first activity that you as an Administrator must perform, once you have addressed the commercial and technical pre-requisite, is to activate and configure the business information extraction at the following path: Settings -> All Settings -> Machine Learning -> Generative AI -> Business Information Extraction.

The Generative AI Scenarios configuration screen in ESM is shown with Machine Learning and Business Information Extraction highlighted.

Each extraction scenario you define can have a different scope and business purpose. For example, you can define an extraction scenario for the following:

  • Create a debit or credit memo request

  • Dispute Management

  • Invoice

  • Payment advice

  • and others …

The main goal is to use information within the Case, such as the Case description and subject, plus attached documents, and map that information to the corresponding business document attributes of the target back-end system. This field mapping activity is required as part of the Input Options within the business document settings.

The next image shows an example of existing scenarios within the business information extraction AI capability of SAP ESM. New scenarios can be created by selecting on the plus icon.

The Generative AI Scenarios configuration screen in ESM is shown. The active tab is Business Information Extraction, located among other tabs such as Case Summary, Email Drafter, and Account Duplicate Check. At the top, the Activate Business Information Extraction toggle is enabled. Activation details indicate Activated On: Mar 10, 2025 A table in the Scenarios section lists seven scenarios with columns for Scenario, Active, Created By, Created On, Updated By, and Updated On. Scenarios like Create Memo Record, Dispute Management, and Price and Availability Check are included. The Active toggle is switched on for multiple scenarios, and the Action column features an edit (pencil) icon for each. On the right-hand side, a search icon (magnifying glass) and a + icon, highlighted with a yellow arrow, allow users to search or add new scenarios. At the bottom-right corner, Save and Cancel buttons are visible, providing options to apply or discard changes.

Once the business information extraction scenario within the Generative AI section is activated and defined, you as an Administrator can move to the configuration of the corresponding business document extraction settings at the following path: Settings -> All Settings -> Enterprise Service -> Business Document.

Select the existing business document from the list, enter the desired version and navigate to the Input Options Tab.

Select the business information extraction scenario activated previously from the drop-down-list, and mark which Case fields must be extracted, from the following options: Case Subject and Case Description.

The next image shows an example of the "Create Memo Request" scenario where the Subject and Description fields from Case are selected for extraction.

The Business Information Extraction configuration screen is displayed under the Input Options tab, focusing on setting up business information extraction capabilities. On the left, the Business Information Extraction toggle is enabled. On the right, the Scenario dropdown is set to Create Memo Record, highlighted by an arrow. Below, under Extract from case fields, checkboxes for Subject and Description are selected, with one arrow pointing to them. An unchecked option for Show confidence level is also highlighted by an arrow. The Field Mappings section follows, showcasing rows for Business Document Field values such as Account Type, Document Number, and Paid Amount, all marked as (Not Mapped) in the Extracted Header Field column; another arrow highlights these dropdown menus. At the bottom-right corner, Save and Cancel buttons allow for saving or discarding the configuration changes.

By selecting the gear icon, you can define a rule with the condition that must be met for triggering the selected business information extraction scenario when a case gets processed by agents. The subject and description can be extracted from the case and a confidence level will be displayed as well next to the related data, to help agents take further action while handling the case.

Once you have completed the field mapping, select the Save button.

If no additional configuration is required, you can now go back to the Designer Tab and select the Validate button. If there are no inconsistencies, the business document can be activated.

the designer tab is open and the validate button is highlighted

Connecting the Business Document to a Case Type

Now that you have created, configured and activated the business document, the last step would be to connect it to a case type. You as an Administrator should navigate to the following path: Settings -> All Settings -> Case -> General Tab.

Under features settings, you must activate the Business Document flag as shown in the next image.

The General settings screen for a case type in ESM is displayed. At the top, the Case Type section shows Name as ZSRC, Description as Service Request, Version Description as Sample Case Type, and Version Status as DRAFT. The Self-Service toggle is set to Yes. Below, under General Settings, the Status Schema is Default, Number Range is listed as cxmcase with a clickable View/edit number range table link, Party Schema is Case, and Catalog is BestRun Customer Support. In the Features Settings section, toggles for Case Hierarchy and Forms are enabled, while Task Management and Business Information Extraction are disabled. A yellow arrow highlights the Business Documents toggle, which is enabled.

Next, you should navigate to the Additional Feature tab, select the plus icon and select the required business document from the pick list.

The Additional Features settings screen for a case type in ESM is displayed. At the top, the Case Type section specifies Name as ZSRC, Description as Service Request, Version Description as Sample Case Type, and Version Status as DRAFT. A yellow arrow highlights the Additional Features tab, which is selected. Below, in the Business Documents section, there is a message stating, No business document available. It looks like you haven’t selected a business document yet, accompanied by a small illustration. On the right side of this section, a dropdown menu labeled Search Business Documents is open, listing options such as Action Automation Test and several Alex-named business documents (e.g., Alex 20241104). A yellow arrow highlights this dropdown menu. At the bottom-right corner of the screen, Activate and Close buttons are visible for confirming or exiting the screen.

Select the Activate button.

Now you have successfully configured a business document and linked it to the respective case type in SAP ESM.

Lesson Summary

In this lesson, you learned how to configure Business Information Extraction using SAP ESM’s Generative AI features. This optional setup—available only with an active SAP AI Units license—allows administrators to create extraction scenarios that pull structured data from unstructured Case content (like descriptions and attachments).