Release Summary 25.20 | Oct 07, 2025
The following key features and improvements, along with bug fixes, have been released in Algonomy CXP products in the release version 25.20.
Recommend
Use Wishlist and Cart Items as Personalization Seeds in Configurable Strategies
We’ve introduced the ability to use a shopper’s cart and wishlist items as personalization seeds within Configurable Strategies. This enhancement allows you to generate more relevant recommendations by considering a shopper’s current interests and intent, not just their historical activity. Items from the cart and wishlist are sourced from the calculated States in the User Profile Service (UPS) response.
This feature is supported across several models, including Viewed Together, Bought Together (By User and In Order), Viewed Then Bought, Similar Products, Complete the Look, Image and NLP Similarity, NLP Cross Sell, and Attribute-based Top Sellers and Top Views. A configuration option is now available to specify the number of seed items from the cart or wishlist to tailor the depth of personalization.
The user interface has been enhanced accordingly. Cart and Wishlist now appear in the Personalization Seeds dropdown for the supported models.
Region-Seeded Strategies Now Supported in Data Science Workbench
You can now create strategies in the Data Science Workbench (DSW) that are tailored by region, allowing recommendation outputs to vary based on a shopper's regional context. This enhancement gives digital optimization teams the flexibility to align product suggestions with local preferences, availability, or business goals.
When configuring a new DSW strategy, "Region" is now available as a seed type under the Composable strategy option. This enables region-aware recommendations to be built and refined just like any other seed-driven approach, ensuring relevance across geographically diverse audiences.
Jira: ENG-30939
Enterprise Dashboard
Expanded Context Definitions: Now Supports Price and Language Conditions
The Context Library has been enhanced with two powerful new condition types: Price and Language. These additions give optimization teams more flexibility in defining when and how Contexts should apply across personalization features.
With the Price condition, you can specify a currency and set a price range that reflects the shopper’s activity. This operates similarly to the price filters used in recommendation rules. If a site does not define currencies or regions, the currency field becomes optional.
The Language condition lets you target based on the shopper’s browsing language, using the site’s configured language tags. Both options are now fully integrated into Context Definitions and can be used to drive more personalized and relevant experiences across the platform.
Easily Import Request URLs in Test Drives
A new "Import request" option has been added to the Recs and Content Test Drives, allowing users to paste a complete request URL and automatically populate the corresponding fields like placements, user ID, and product ID. This streamlines the debugging process by removing the need to manually enter individual inputs.
The system parses valid request URLs, maps the parameters appropriately, and validates that the request matches the current site. Once imported, the Test Drive runs as usual, displaying results based on the selected test type. This enhancement speeds up testing and makes troubleshooting real scenarios more efficient.
Jira: ENG-30726
Social Proof
Log Badges Shown to Shoppers for Tracking & Reporting for Social Proof Badging API
The Social Proof Badging API now logs the badges shown to shoppers as part of the track experience events. Whether a single badge or multiple badges are returned in the API response, they are logged accordingly. Multiple badges are recorded as comma-separated values, similar to how social proof messages are handled.
This update ensures consistent tracking across badge-based and message-based experiences. All standard parameters captured for messaging are now logged for badges as well.
Jira: ENG-30981
Social Proof and Guided Selling
Currency-Based Reporting Now Available in Social Proof and Guided Selling
You can now view Social Proof and Guided Selling reports by specific currency. A new Currency dimension has been added, allowing merchandisers to filter and analyze performance metrics based on individual currencies or across all currencies.
By default, reports will show data in the site’s default currency. Users can switch to other available currencies, with the selected currency passed through the API to ensure accurate results. This enhancement brings reporting capabilities in line with Dynamic Experience reports, improving consistency across analytics tools.
Jira: ENG-30942
Engage
Introducing Content Test Drive
The Content Test Drive helps digital optimization managers preview and validate content placements. The interface mirrors the Recs Test Drive experience, allowing users to select multiple content placements for a single page type and preview the content served through the Personalize API.
Users can customize test parameters, including user ID, region, channel, tag filters, tag refinements, and blocked content, to simulate real-world conditions. The tool supports both production and integration environments, with full support for advanced tag-based filtering and refinement logic.
The results section displays rendered content, enabling teams to fine-tune, validate, and troubleshoot content campaigns quickly and confidently before publishing.
Jira: ENG-31146
Chatbot
Session Management for Ecommerce Sales Assistant Chatbot
The Ecommerce Sales Assistant Chatbot now supports session management to maintain chat history even when the shopper navigates away or reloads the page. This enhancement ensures that the shopper can return to the chat and continue from where they left off — including reviewing previous product suggestions — without losing the conversation thread.
When a session is active, chat history is reloaded automatically with helpful notification. If a session expires after inactivity, or is manually reset, the shopper is clearly informed that a new session has started. This update makes the chat experience smoother and more consistent across the shopping journey.
Jira: ENG-31089
Other Feature Enhancements
The following feature enhancements and upgrades have been made in the release version 25.20.
|
Jira # |
Module/Title |
Summary |
General Availability |
|---|---|---|---|
|
Enterprise Dashboard: Support for Product IDs with Double Hyphens in Merchandising Reports |
We’ve improved the handling of product external IDs in the Merchandising Report. Previously, when a product ID included double hyphens (--), users encountered a 403 Forbidden error while accessing the report. This enhancement ensures that whenever a product ID contains double hyphens, it is automatically enclosed in double quotes in the UI. |
07-Oct-25 |
|
|
Related Searches Job Now Focuses on Top and Zero-Result Queries |
The related searches generation job has been refined to use more targeted inputs, specifically top-performing and zero-result queries, instead of pulling all search terms. This behavior can now be configured per site, allowing selection of key metrics over a rolling date range to drive more relevant related search suggestions. |
07-Oct-25 |
|
|
Find: CTR and Findability Enhancements in Find Reports |
Click Through Rate (CTR) and Findability metrics in both Find and Find Search Terms reports now include options to view data with or without deduplicating product clicks. This enhancement helps users analyze shopper behavior more accurately when multiple clicks occur on the same page. The rollup logic is complete, and the update is awaiting UI integration, after which users can toggle between counting all clicks or only unique ones. |
07-Oct-25 |
|
|
Find: Improved Handling of Solr Requery for Personalization Decisions |
A new configuration flag psdOptimize has been introduced to better manage Solr requery behavior during personalization, particularly when the number of psdItems is greater than or equal to the number of rows requested. With this enhancement, when rows are less than psdItems, the system now automatically adjusts psdItems to match the row count, ensuring efficient query execution without errors or unnecessary retries. |
07-Oct-25 |
|
|
Find: Related Searches Now Powered by Azure OpenAI |
The Related Searches job has been upgraded to use Azure OpenAI for generating related search terms. This change addresses prior reliability issues encountered with the standard OpenAI endpoint, ensuring more stable and consistent performance. Both the Related Searches Generator and Publisher jobs now leverage the Azure endpoint for generating terms. |
07-Oct-25 |
|
|
Recommend: Region-Based Strategy Type Now Supported in DSW |
You can now create region-seeded strategies in Data Science Workbench (DSW), enabling region-specific recommendations for improved relevance and targeting. When setting up a new strategy, "Region" is now available as a strategy type and also as a seed type when using Composable strategies. This enhancement ensures that the model is trained using region-based data and returns recommendations tailored to the shopper’s region, based on the regionId passed in the request. |
07-Oct-25 |
|
|
Recommend: User-Based Filters Moved to User Tab in Merchandising Rules |
User history filters like Categories Pending Purchase and Categories Already Purchased have been moved from the Category tab to the User tab for better organization. The category picker now appears only when these options are selected, and required fields are clearly marked. To prevent conflicts, these filters cannot be combined with other filter options—selecting one will disable the rest. |
07-Oct-25 |
|
|
Recommend: Margin Sorting Tooltip in Configurable Strategies |
A tooltip has been added to explain Margin sorting: products are ranked by model score multiplied by profit. If Margin isn’t enabled, hovering over the option shows: “Please enable the Margin Model in Model Options to use Margin sorting.” |
07-Oct-25 |
|
|
Dashboard Chatbot: Dashboard Chatbot Now Available on All Pages |
The AlgoChat icon is now visible across all pages in the dashboard, making it easier to access help wherever you are. The chatbot’s welcome message has also been updated to remove product-specific wording. It now reads: “Hi, I’m AlgoChat. I’m here to answer questions about Algonomy and guide you through setup, features, and best practices. How can I help?” |
07-Oct-25 |
|
|
Streaming Catalog: Fix for NCA Database Connection Leak |
A connection leak in the Native Catalog Adapter (NCA) was resolved by improving how PostgreSQL connections are handled during streaming. This ensures stability across ingest and scoped action flows. |
07-Oct-25 |
|
|
Platform: MailRecCache: Configurable Datadog Agent Address |
The MailRecCache component now supports a configurable Datadog agent address. This enhancement allows for greater deployment flexibility, making it easier to direct telemetry data to custom monitoring endpoints as needed. |
07-Oct-25 |
|
|
Recommend: User Filter Options Consolidation in Merchandising Rules
|
User-related filter options in Merchandising Rules have been moved from the Category tab to the User tab for better clarity. These include filters like Categories Pending Purchase and Categories Already Purchased. The category picker appears only when these filters are selected, with an asterisk added to denote required fields. To prevent conflicting settings, these user-specific options are now mutually exclusive with other filters—activating one disables the others automatically. |
07-Oct-25 |
|
|
Find: Normalize Character Accents for Hybrid Search Matching |
Hybrid search now handles accented and case-variant keywords more effectively. Queries like Tricô and Trico will return consistent results, regardless of accents or letter casing. This is achieved by normalizing keywords during both storage and retrieval of query vector embeddings, ensuring uniform behavior across languages and input styles. |
07-Oct-25 |
|
|
Chatbot: Chatbot Help Page for First-Time Shoppers |
A help page is now shown the first time a shopper interacts with the chatbot, explaining how to use the conversational assistant. It highlights how to ask product-related questions and what kind of responses to expect. The help page is dismissed once the shopper closes it and won’t be shown again. |
07-Oct-25 |
|
|
Dynamic Experience: Updated Role Permissions |
Role permissions for Dynamic Experience templates have been updated to include MANAGE_API_CLIENTS and VIEW_API_CLIENTS privileges. This ensures users have the necessary access to manage and view API client settings when configuring chatbot-related templates. The enhancement simplifies setup and customization for chatbot and Dynamic Experience configurations. |
07-Oct-25 |
|
|
Chatbot: Support for API Key and API Client Key
|
The chatbot front-end has been enhanced to pass both the API key and API client key, ensuring proper site-specific configuration and click-through functionality. The chatbot dynamic experience template now supports using click URLs returned from the API response and automatically hides or discards products without valid URLs. This ensures accurate tracking and a consistent experience across client implementations. |
07-Oct-25 |
Bug and Support Fixes
The following issues have been fixed in the release version 25.20.
|
Jira # |
Module/Title |
Summary |
General Availability |
|---|---|---|---|
|
Ensemble AI: Ensemble Report: Fix for Missing Style Performance Data |
An edge case was identified where some products present in the data were missing from the catalog. This mismatch led to failures in the style performance rollup, resulting in missing report data from mid-August. This issue has now been resolved. The report can handle such cases gracefully, and the style performance data is available as expected. |
07-Oct-25 |
|
|
Chatbot: Chat Suggestions: Regionalised Pricing and Sale Price Display |
Chat suggestions now support region-specific pricing and can differentiate between base and sale prices using the regionId in the getProducts API. Products with a valid sale price lower than the base price are flagged as on sale. The chatbot display can be configured to show both prices, only one (sale or base), or no price. Sale prices can also be visually styled to align with brand preferences. |
07-Oct-25 |
|
|
Chatbot: Expanding Input Box for Long Queries |
The chatbot input field now automatically expands to display the full query as the user types, improving visibility and ease of editing. It grows and shrinks dynamically based on the content and shows a scrollbar when the input exceeds a certain height. This ensures long messages are no longer clipped or hidden within the input area. |
07-Oct-25 |
|
|
Enterprise Dashboard: Font Face Quotes Render Properly |
Fixed an issue where font names in single quotes were converted to special characters due to double HTML encoding. Encoding is now applied only once, so saved font faces appear correctly. |
07-Oct-25 |
|
|
Chatbot: Chatbot No Longer Sends Empty Queries |
Resolved an issue where the chatbot would trigger a message even when no question was entered. It now waits for valid user input before responding. |
07-Oct-25 |
|
|
Chatbot: Product Recs in Chatbot Now Clickable |
Fixed an issue where product suggestions from the chatbot were missing links, preventing users from clicking through to product pages. The recommendation cards now include proper href attributes, just like the default recs, ensuring all suggestions are fully interactive. |
07-Oct-25 |
|
|
Enterprise Dashboard: Exclusion Tags in Content Catalog |
Resolved an issue where the exclusion tag field disappeared when no inclusion tags were selected, making it unmaintainable. The Exclusion box is now always visible, and a message prompts users to add at least one inclusion tag for filtering to apply. Tag selection colors have also been updated to clearly indicate inclusion (green) or exclusion (red). |
07-Oct-25 |
|
|
Enterprise Dashboard: User Management Page Selection Issue |
Resolved an issue in the User Management portal where selected checkboxes on one page remained selected across pagination, causing confusion during user deletion. The UI has been corrected to ensure selections are specific to each page, preventing accidental deletions. |
07-Oct-25 |
|
|
Enterprise Dashboard: Unsaved Campaign Changes Persisting |
Fixed an issue where unsaved edits in Campaign settings appeared to persist after navigating away from the page. Users were seeing changes reflected as if they were saved, even though no save action was taken. The Campaigns page now properly reflects only saved configurations. |
07-Oct-25 |
|
|
Enrichment Feed File with No Data Now Rejected |
Addressed a critical issue where zero-byte enrichment feed files were being processed, resulting in unintended product data removal. The system now includes guardrails to automatically reject feed files with no content or missing expected files after unzipping. This prevents processing of incomplete or empty data feeds. |
07-Oct-25 |
|
|
Chatbot: Chatbot Screen Clipping on Mobile Resolved |
Fixed a visual issue where the chatbot interface was clipping on the left side of mobile screens. This occurred due to use of vw and vh CSS units, which caused overflow when vertical scrollbars were present. The layout has now been updated to use percentage-based units, ensuring proper rendering across all screen sizes. |
07-Oct-25 |
|
|
Find: Incorrect Parsed Query in Solr Debug for B2B Sites |
Fixed an issue where the parsedQuery and parsedQuery_toString fields were incorrectly displayed or missing when using the findDebug=solrDebugAll parameter in the Find API for B2B-enabled sites. This bug only impacted the debug section and did not affect search functionality or API responses used by clients. The system now correctly returns the expected query details during debugging. |
07-Oct-25 |
|
|
Enterprise Dashboard: Sorting in User Affinity Graphs |
Resolved an issue where values in the user affinity graphs were not correctly sorted by total score, including scores from browsed category pages. The graphs now accurately reflect sorted values based on total affinity, improving data clarity and usability. |
07-Oct-25 |
|
|
Discover, Find: Region Parameter Handling in New Browse API |
Resolved an issue where the new browse API incorrectly used the RELEASE_DATE field instead of product_region_id, resulting in no results when the region or rid parameter was used. The API now correctly uses the region field, ensuring accurate results based on shopper location. |
07-Oct-25 |
|
|
Chatbot: Splash Page Design Customization for Chatbot |
Addressed a request to align the chatbot splash screen with a client-provided design. Updates included UI adjustments to match the intended layout, such as customizing the title bar color and font styling. |
07-Oct-25 |
|
|
Chatbot: Default Recommendations Removed on Chat Open |
A fix has been implemented to address the issue where the chatbot opened with two default recommendation widgets already displayed. A new configuration variable, Show Initial Recs, has been introduced to control whether these initial recommendations appear when the chat launches. This ensures a cleaner and more controlled chatbot experience. |
07-Oct-25 |