Changed behavior

Data quality rule associations

• •When you run ingress or file import jobs, MDM SaaS runs all migrated rule associations in the load step.
Previously, when you ran ingress or file import jobs, MDM SaaS ran all basic, advanced, and DaaS rule associations in the transform step.
• •When you import or ingress records without mapping input fields from migrated rule associations configured to run even if all input fields are empty, MDM SaaS runs these associations.
Previously, MDM SaaS didn't run these rule associations.
• •When import or ingress records with picklist values, MDM SaaS retains the values that you import even if they aren't available in the enterprise code list.
Previously, MDM SaaS converted these picklist values to null.

Adaptive AI match model training process

• •When you create a match model, the Adaptive AI match model doesn’t include any match fields by default. Match fields added in the candidate selection criteria and Directed AI match rules appear as recommendations for inclusion in the Adaptive AI match model training.
Previously, the match fields added in the candidate selection criteria and Directed AI match rules appeared by default as match fields in the Adaptive AI match model.
• •You can now modify the configuration of a match model even after you start training its Adaptive AI match model. For example, you can add or remove match fields from existing Directed AI match rules or add new Directed AI match rules to the match model. The match process displays the newly added match fields as recommendations for inclusion in the Adaptive AI match model training.
Previously, you couldn’t modify the configuration of a match model after you started training its Adaptive AI match model.
• •If you modify the following settings of a match model during or after training the Adaptive AI match model, the match process discards the training and requires you to retrain the model from the beginning:
• - Default population
• - Candidate selection criteria
• - Match fields in the Adaptive AI match model
Previously, you couldn’t modify these settings during or after training the Adaptive AI match model.
• •When you train an Adaptive AI match model, the match process now requires a minimum of 500 candidate record pairs to generate sample match candidates for accurate predictions. You can contact Informatica Global Support to modify this minimum requirement based on your business needs.
Previously, no minimum candidate record pairs requirement was enforced.
For more information about sample match candidates generation, see Viewing Match – Start Training jobs.

Purge data

Dynamic pools

Reprocess master data jobs

Enrichment and Validation Orchestrator

• •In the response that an Application Integration process returns for a validation request, groupId is a required property in root field objects. When you create a process for validation, ensure that the process object for root fields includes the sys object that contains a groupId field.
Previously, groupId wasn't a required property in root field objects in the fields array of a response that a validation process returned.
The following JSON code is a sample response that includes the groupId property for the drugCode and drugManufacturer root fields:
[
{
"docId": "business_entity-master_id1",
"validation": [
{
"passed": false,
"fields": [
{
"fieldname": "drugCode",
"dtsp": 23,
"sys": {
"groupId": "f88114db-461c-34c3-87b6-d2cebc76e85e"
}
},{
"fieldName": "drugManufacturer",
"dtsp": 23,
"sys": {
"groupId": "f88114db-461c-34c3-87b6-d2cebc76e855"
}
},
{
"fieldName": "regulatoryInformation",
"dtsp": 23,
"sys": {
"groupId": "f88114db-461c-34c3-87b6-d2cebc76e85e"
}
}
]
}
]
}
]
• •You can now configure an objective to run when all its input fields are empty. When a user creates or updates a record, the objective with the setting enabled runs if the input fields of the objective are empty.
Previously, during record creation, if a value was added to a field at the same level as the input fields, the objective ran even if its own input fields were empty. Objectives did not run during record updates in this scenario.
For example, consider the Demography and Employment Details objectives created after the October 2025 release. The Demography objective was configured to run on the First Name and Last Name fields, and the Employment Details objective was configured to run on the Designation field. In the October 2025 release, when a user created a record and entered a value in the Designation field, the Demography objective also ran because the Designation field is at the same level as the First Name and Last Name fields.
From the April 2026 release, all existing objectives have the option to run on empty fields enabled by default. As a result, the Demography objective runs during both record creation and record updates when a value is entered in the Designation field, even if its input fields are empty.
If you want an existing objective to behave as it did in the October 2025 release, disable the Update trigger action in the objective. Then, create another objective with the Update trigger action and disable the option to run on empty fields.
• •If the option to reject records on validation failure is enabled for an objective and a rule association in the objective fails with the Exception status, records are rejected.
Previously, users could submit records even when rule associations failed with the Exception status.

Relationship REST API