Configure Data Quality > Orchestrating data enrichment and validations > Configuring data enrichment
  

Configuring data enrichment

To enrich records with additional data from data providers, configure rule associations that use enrichment plugins.
You can enrich records by transforming the data within them.
To perform advanced data enrichment, you can also connect to external data providers and enrich records with the data that they return. For example, you can connect to OpenAI ChatGPT to translate the product description in an Item record.
The following table lists the plugin that you can use to configure data enrichment:
Plugin Name
Category
Service that the plugin uses
Can connect to external data provider?
Application Integration-based Enrichment
Enrichment
Application Integration
Yes
Note: Enrichment is triggered and applied to a record even if the user doesn't have access to certain fields involved in the enrichment process.

Application Integration-based Enrichment plugin

To enrich records with data from external data providers, use the Application Integration-based Enrichment plugin and configure data enrichment.
When a rule association with a data enhancement rule that uses this plugin is triggered, the input fields are sent to an Application Integration process, which connects to an external data provider. When the data provider returns data for enrichment, you can automatically enrich or assign records to a hierarchy. You can also display the enrichment data as recommendations in business applications.
When you configure data enrichment, you can choose to enrich specific fields of records. For example, you can send a product description to an external data provider and extract the brand name from the description.
You can also add records to hierarchies based on the data from an external data provider. For example, to assign product records to a hierarchy, you can send a product description to an AI-powered external data provider to get relevant product categories. Based on the product categories that the data provider returns and the existing category records, you can either directly assign the records to a hierarchy or present the category records as recommendations to users.
This plugin uses an Application Integration process to perform enrichment. For more information about creating a process for Application Integration-based enrichment, see Payload formats for Application Integration-based enrichment.
The following table lists the properties of the plugin:
Property
Description
CAI Process Identifier
Identifier of the Application Integration process that the plugin uses.
CAI Process Name
The API name of the process that's specified in the CAI Process Identifier property.
Accept Threshold
The confidence score above which data from the data provider is automatically accepted and applied. To enable rule associations to enrich records automatically, enter 0 in the Accept Threshold and Reject Threshold fields. To display the information from the data provider as CLAIRE recommendations, specify the accept and reject thresholds with a value greater than 0.
For more information about the accept threshold, see Determining the data enrichment type.
Reject Threshold
The confidence score below which recommendations for enrichment are automatically rejected.
For more information about the reject threshold, see Determining the data enrichment type.
Confidence Score Field
The name of the field that contains the confidence scores from the data provider. To display the confidence scores along with the recommendations in business applications, specify the name of the field. You can get the field name from the Output Fields tab. If you don't specify the name of the field that contains the confidence scores, CLAIRE displays all recommendations without applying the accept and reject thresholds.

Determining the data enrichment type

When you configure Application Integration-based enrichment, you can configure to enrich records automatically or to provide CLAIRE recommendations to the user.
When you configure a rule association to provide CLAIRE recommendations, CLAIRE displays enrichment data as recommendations that are based on confidence scores from data providers. When you configure data enrichment, you can set accept and reject thresholds in a data enhancement rule.
For example, you can send a product description to an AI-powered data provider and request the data provider to suggest brand names based on the description. When the data provider suggests multiple names for a brand, CLAIRE displays the suggestions that fall within the Accept and Reject thresholds as recommendations.
The following image shows how the accept and reject thresholds impact the recommendations from a data provider:
The image shows that recommendations from a data provider are automatically rejected if the confidence score falls within the reject threshold. Recommendations are automatically accepted if the score exceeds the accept threshold, and recommendations are displayed in business applications if the score falls between the accept and reject thresholds.

Data enrichment configuration scenario

You work for Acme and want to manage its product information on Amazon. You want to translate the English descriptions of products into French for the customers in France after business users create Item records. To ensure that the products listed on Amazon reflect the correct product classification, you also want to automatically classify the records.
To translate English product descriptions into French, design a custom process named Translate Product Description in Application Integration. Design the process to send English product descriptions to OpenAI ChatGPT for translation. Additionally, design a process named Product Classification for Amazon to connect to an external service that provides product categories based on product descriptions.
The following image shows how two Application Integration processes connect to two external services:
Two Application Integration processes named Translate Product Description and Product Classification for Amazon send requests to two external services named ChatGPT and Product Classification Service.
In Business 360 Console, you create an objective group named Enrich Production Information on Amazon. Within the objective group, you define objectives named Enrich Brand Information and Classify Products.
In the Enrich Brand Information objective, you create a rule association named Translate description to French with a data enhancement rule named Translation rule. You configure this rule association to send product descriptions to the Translate Product Description process and to update the Item records with the translated descriptions that ChatGPT returns.
The following image shows a sample data enrichment configuration for Acme Pvt Ltd:
A sample data enrichment configuration in Business 360 Console includes an objective named Enrich Brand Information that contains a rule association named Translate description to French. It includes a data enhancement rule and field mappings within the rule association.
Similarly, create a rule association named Mappings for Product Classification within the Classify Products objective. You can configure the rule association to send product descriptions to the Classify Products for Amazon process. Also, you configure the rule association to classify the Item records based on the product categories that the process returns.