Release Summary 25.18 | Sep 04, 2025
The following key features and improvements, along with bug fixes, have been released in Algonomy CXP products in the release version 25.18.
Recommend
New Placement Profile and List Pages for Centralized Management
Two new features have been introduced to streamline the management of Recommend placements and their configurations:
Placement Profile Page
Offers a centralized view of all rules and configurations applied to a specific placement. Optimization managers can easily review rule types like Advanced Merchandising, Strategy Rules, Boosting, and Restrictions, all organized in dedicated sections with metadata and environment statuses. A new checkbox option also allows specific placements to be excluded from deduplication in the same request, ensuring results aren’t impacted by other placements. This initial release focuses on rule visibility and is available to employees only. It will be rolled out to customers after an evaluation period.
Placement Profile List Page
Displays all Recommend placements in a searchable, filterable table that includes page type, layout, item count, and status for both Integration and Production environments. Users can navigate directly to individual Placement Profile pages, add new placements, or update statuses using a guided modal. Together, these enhancements improve placement visibility, rule traceability, and operational efficiency across environments.
Jira: ENG-30451, ENG-30665, ENG-30565, ENG-30025
Auto-Suggest for User ID in Configurable Strategies Preview
An enhancement has been made to the Configurable Strategies preview interface to improve usability for digital optimization managers. The User ID field on the "View Results" tab now supports auto-complete suggestions, enabling quicker and more efficient testing with previously used user IDs.
This feature uses the same list of stored user IDs as found in the User Affinity Configuration and Test Drives, presented in a dropdown format. Users can also manually enter a new user ID not previously used, ensuring flexibility while previewing recommendation results.
Jira: ENG-30816
Ensemble AI
Style ID Now Displayed in Ensemble AI Interface
Ensemble AI now includes Style IDs in both the list view and style detail page of the user interface. This enhancement helps merchandisers easily identify and copy Style IDs when configuring style-specific outfits via API, especially for server-side integrations or Active Content.
Style names are now shown alongside their corresponding Style IDs using a consistent display convention. Additionally, users can quickly copy the Style ID for reuse.
Jira: ENG-30757
Dashboard Chatbot
Conversational Context and History Now Supported in Dashboard Chatbot
The dashboard chatbot now supports thread-based conversational context using a session ID (thread_id). When a user initiates a conversation, the backend returns a unique session ID, which is automatically used for all follow-up messages within that session. This enables the chatbot to retain and apply context across interactions, leading to more natural, relevant, and continuous responses. Resetting the chat clears the session ID and starts a fresh session with new context.
Additionally, the chatbot now remembers the last three exchanges—each consisting of a user message and the chatbot’s response—within an active session. This short-term memory improves its ability to handle follow-up questions, reducing the need for users to repeat themselves. Once the chat is reset, this history is cleared. These enhancements make the chatbot more intuitive and user-friendly, aligning it with conversational expectations similar to platforms like ChatGPT.
Note: This release is currently available to employees only. It will be made available to customers in an upcoming release.
Other Feature Enhancements
The following feature enhancements and upgrades have been made in the release version 25.18.
|
Jira # |
Module/Title |
Summary |
General Availability |
|---|---|---|---|
|
Engage: Enhanced Tag Content Now Triggers ps_catalog Job Automatically |
When new content with enhanced tags is published via the dashboard, the system now automatically triggers the ps_catalog job. This ensures that the content gets properly associated with the relevant category, brand, or product attribute based on its tags. This enhancement addresses issues where tagged content was not appearing in campaigns .Once the job completes, the content is added to the Product Catalog and becomes eligible for use in affinity-sorted campaigns. |
04-Sep-25 |
|
|
Engage: Catalog Associations Automatically Removed When Enhanced Tags Are Cleared |
Enhanced tag management has been improved to ensure accurate catalog associations in Engage campaigns. When enhanced tags (such as category, brand, or product attribute) are removed from a content item, the corresponding catalog associations are now also automatically removed. This prevents outdated associations from affecting affinity sorting results. |
04-Sep-25 |
|
|
Ensemble AI: Informational Message Update for Free Form |
The UI for Ensemble AI’s Free Form input has been updated to prevent errors caused by sending unexpected IDs in the seed filters. This enhancement ensures the payload is properly structured and provides a clearer message to guide users during configuration. |
04-Sep-25 |
|
|
Mail Replication Now Configurable for Cloud Deployments |
To streamline cloud-based deployments and reduce overhead, mail replication is now controlled by a configurable flag in the MailRecCache service. When the disable.mail.replication flag is enabled, the service writes directly to Cassandra, bypassing Kafka and eliminating the need for MailRecCacheConsumer. By default, replication remains unchanged, and this setting is currently enabled only for the cloud implementation using Cloudflare proxy and custom endpoints. |
04-Sep-25 |
|
|
Ecommerce Chatbot: Chatbot Now Supports Attribute-Based Filtering |
The chatbot has been enhanced to support attribute-based filters such as GenderDescription and Color for more relevant product recommendations. A new “Filter Attributes” section on the LLM Configuration page allows merchandisers to define which attributes to use during vector catalog creation. These filters are now integrated with the portal API and Milvus, improving search precision without affecting embeddings. |
04-Sep-25 |
Bug and Support Fixes
The following issues have been fixed in the release version 25.18.
|
Jira # |
Module/Title |
Summary |
General Availability |
|---|---|---|---|
|
Dashboard Chatbot: Incorrect Date Displayed in Release Search |
Fixed an issue in Algo AI where searching for release documentation (e.g., "Release Summary 25.16") showed an incorrect future release date—August 25, 2025. The system now accurately reflects the correct release date for summaries in search results. |
04-Sep-25 |
|
|
Placement Deduplication Control Enhancement
|
You can now configure specific placements to be excluded from deduplication within a multi-placement request. This ensures that product recommendations in a placement—such as those used for category recs—are not affected by deduplication with results from other placements in the same call. When enabled at the placement or strategy rule level, the selected placement avoids deduplication with others, while still applying short-name deduplication within itself. This provides greater control over recommendation consistency and placement behavior. |
04-Sep-25 |