v2025.18 Release Notes – September 9, 2025

v2025.18 Release Notes – September 9, 2025

Here’s a summary of what we shipped in Release v2025.18 (September 9, 2025):

Improvements

Platform Area

Description

Platform Area

Description

Attributes

Users can now delete unused Custom Attributes directly within the Data Mapper workflow, streamlining clean-up and reducing the number of clicks required.

Content Designer

Improved handling of LLM rate limits in tactic and tone tagging. Publishes now retry with deferrals instead of failing, reducing errors during high-volume publishing.


Fixes

Platform Area

Description

Platform Area

Description

Reporting

Overdue Balance on Entry - Fixed an issue where this value could be null if the first valid overdue balance arrived after a Playbook execution began. It now correctly captures the first non-null overdue balance during execution.

Platform-wide

Fixed an issue where users accessing a deep link while logged out were redirected to the organization switcher after login instead of the intended page. Deep links now correctly take users to their target location once authenticated.

Conversions

Event Type Selection - Fixed an issue where creating a new Conversion failed with an error when retrieving event types. The Campaign service has been updated to include the new Customer Trail event types.

Prevent

Insurance Claim Payments - Fixed an issue where payment amounts from Walnut’s InsuranceClaimCreated events were stored in cents instead of dollars. The system now correctly handles these values in dollars to ensure accurate claim and payment data.

Insurance Claim Processing - Fixed an issue where valid insurance claims were incorrectly flagged as duplicates and skipped. The uniqueness check has been updated to include the Payment To field, ensuring all valid claims and payments are processed correctly.

Delinquency Archetypes

Baseline Delinquency Archetype Inference - Fixed an issue where BDA inference jobs failed due to missing run IDs. The process now correctly retrieves the previous successful run ID, ensuring inference jobs complete as expected.