Release Summary 26.23 | Nov 12, 2026

The following key features and improvements, along with bug fixes, have been released in ADA Global Digital Experience Personalization (DXP) in the release version 26.23.

Recommend

Recs Response Builder

You can now extend the recsForPlacements API response with your own structured data, set per placement. This wasn't possible before. Your tech team builds a response template on the new Recs Response Builder page, choosing the components and attributes it holds, each attribute's data type, and its default value. Changes can be saved as a draft and checked in your integration environment first. The page is available to users with the Manage Placements and Layouts role.

To get the data back, pass includeConfigSchema=true on your recsForPlacements request. Each placement then comes back with a configuration array holding one object per live template, with its ID, name, version, and every attribute filled in. Values are validated against your template, and where nobody has set a value for a placement, we return the default. The parameter defaults to false, so existing integrations see no change. You can also pin the response to a specific template version with schema_version, which helps mobile apps running several versions at once.

Jira: ENG-32620, ENG-32621

Ensemble AI

Instant Exclusion of Outfits and Products in Ensemble AI

Excluded outfits are now removed from the API response right away, without waiting for the next model run. Exclusions are saved per seed and stay in effect until they are reversed or replaced by the next model run’s output. Un-excluding an outfit brings it back just as quickly.

Merchandisers can also exclude a product, for example when Virtual Try-On does not represent it well. Every outfit that contains the product is removed immediately, whether-or-not it is a seed product, and the product can be included again at any time to restore those outfits. Some outfit slots may stay empty until the next model run refreshes the results.

Merchandising decisions take effect in near real time (typically within 10–15 minutes), so shoppers only see outfits that match what merchandisers have approved.

Jira: ENG-33222, ENG-33539

Review All Seed Products with Pagination in Ensemble AI

Ensemble AI now lets you review every seed product in a category, not just the first 100. Seed products load 100 at a time, with Previous and Next buttons on the Outfit and All Ensembles pages. The page label shows the range on display, such as “Page 2 (101 - 200)”. When you change pages, the first product is selected automatically, its ensembles reload, and the seed grid scrolls back to the top.

The Outfit page breadcrumb also shows the style you are reviewing, in the format “Ensembles (Style Name)”, with the style name and ID on hover. Merchandisers can review ensembles across a full category and always know which style they are working on.

Jira: ENG-33551

Exclude Outfits and Products Across Styles in Ensemble AI

The “View ensembles across styles” page now includes an option to exclude outfits and products, so that they can review all the outfits for a given seed at one place, and hence reduces lot of effort to go into each style for review process.  Merchandisers/stylists can review and exclude and restore outfits and products without opening each style. Outfits can be excluded from multiple styles at once. When excluding a product, you choose either “Only this style” or “Across all styles”, and styles where the product is part of the style definition are skipped with a notice.

The Exclusions tab, shown with a count, lists excluded ensembles grouped by style and excluded products with an option to include them back. Including a product from this tab applies to the style where it is listed. Changes are applied when you save and stay in sync with the style-level pages. Merchandisers can review and manage exclusions across styles in one place, which saves time.

Jira: ENG-33484

Discover

Discover Reports Menu in Reporting

Reporting now includes a dedicated Discover menu with a submenu for Discover reports. The menu appears automatically for clients who have Discover enabled, so only relevant options are shown. It provides access to the Discover Performance Report and the API Call Count Report.

Having both reports in one place makes it easier to review how personalized browse and navigation is performing and how much it is being used.

Jira: ENG-33542

Social Proof

All Social Proof Intervals Enabled by Default

Social Proof now enables metrics for every interval by default for all clients, from real-time intervals up to 4 weeks. No additional configuration is needed, so each client has data available across all time ranges from the start.

Merchandisers can choose the interval that best fits their messaging, whether it is real-time urgency or longer-term popularity, without any setup effort.

Jira: ENG-33462

Product Coverage in Social Proof Campaign Reports

Social Proof campaign reports now include a “products shown social proof” column and trend, showing the number of unique products that displayed a message. The count is available for each variation and across campaigns in the report that spans all campaigns. The reporting API also returns this count as a new response parameter.

Merchandisers can see how widely social proof messaging is reaching the catalog and judge the coverage of each campaign.

Jira: ENG-33292

Other Feature Enhancements

The following feature enhancements and upgrades have been made in the release version 26.23.

Jira #

Module/Title

Summary

General Availability

ENG-33617

Discover, Find:

Consistent Boost Calculation in Discover 2.0

Discover 2.0 now uses a boost factor of 1 when calculating personalization boosts, matching the existing Discover behavior. Boost values are no longer scaled, so the results from personalization and global ranking are merged and reranked exactly as configured.

Shoppers see browse results that stay consistent with the existing Discover experience, and merchandisers can rely on boost settings behaving as expected.

12-Nov-26

ENG-33520

Discover, Find:

More Accurate Facet Counts in Find and Discover

Find and Discover now support an improved facet method that keeps facet counts consistent with the total number of results when products are grouped in browse. Standard facet functionality continues to work as before, and the improved method can be turned on with a configuration switch.

Shoppers see accurate counts next to each filter option, which makes browsing and refining results more reliable.

12-Nov-26

ENG-33480

Find:

Early Alerts for Stalled WOC Data

WOC jobs now report their true status, so a job that fails internally is no longer shown as successful. Jobs for sites that are not subscribed to WOC calculation are marked as failed, and the regular maintenance check sends an alert to the maintenance Slack channel when a site’s WOC data has stalled.

Issues are detected early and can be resolved before they affect data quality.

12-Nov-26

ENG-33588

Find:

Smoother Catalog and Query Vector Processing in Find

The catalog and query vector jobs in Find now complete successfully when there is no new catalog or query data to process. Earlier runs with no updates could be reported as failures, which triggered unnecessary alerts in the maintenance channel.

Maintenance alerts and job status reports now highlight only the items that need attention.

12-Nov-26

ENG-33558

Shopping Assistant:

Compare More Than Two Products in Shopping Assistant

Shopping Assistant can now compare two or more products side by side using the comparison recommendations service, which is turned on through a page setting. The table shows the image, name, prices, savings, brand, rating, category, and all product attributes. The lowest price and highest rating are highlighted, and a “Show differences only” option hides identical rows.

Shoppers can select products in the table to ask follow-up questions, and comparisons are saved in the chat history. Labels are available in English, Spanish, French, and German, and the table adapts to small screens. Shoppers can weigh several options at once and decide with confidence.

12-Nov-26

ENG-33234

Enterprise Dashboard:

Price Range Boost and Bury in New Discover

Merchandisers can now boost or bury products by price range in New Discover boost and bury. After selecting a category, choose the price attribute to use (such as price, effective price, or sale price), set a minimum and maximum value, and enter a boost amount from -100 to +100. Positive values boost products in the range, negative values bury them, and the default is 1.

Settings apply per category, and the effect is relative to other active rules. For example, lower-priced products can be ranked above higher-priced ones, and a value of 100 places products above everything else. Merchandisers get precise, price-based control over how products are ranked in browse.

12-Nov-26

ENG-33447

Enterprise Dashboard:

Search Test Drive Improvements

The new Search Test Drive has been refined based on user feedback. The Hybrid Search tab now supports a TopK value of up to 200, and search suggestions appear as you type.

Query Understanding is easier to manage. The page scrolls to the top when you create or edit a rule, and brand and category rules update existing entries instead of creating duplicates. Merchandisers can test and tune search with fewer steps and cleaner rules.

12-Nov-26

ENG-33455

Enterprise Dashboard:

Save as Draft in Recs Response Builder

You can now save a response template, or changes to one, as a draft. Drafts are not visible in production, so you can validate them in the integration environment first. Drafts are marked “Draft” on the list page and “Unpublished changes” on the detail page.

The JSON preview now matches the recommendations response, without descriptions or allowed values, and the “Required” column has been removed. Schema Builder is now Recs Response Builder, with “response template” used in place of “schema” across the page. Changes can be validated before they go live, and the terminology is easier to follow.

12-Nov-26

ENG-33485

Enterprise Dashboard:

Streamlined Model Choices in Configurable Strategies

NLP Cross Sell and Complete the Look are temporarily hidden while their pipelines are being updated. They no longer appear as options when creating or editing a strategy, and ImageCTLModel is hidden on the Model Options page. Existing strategies that already use these models continue to show them as selected and open and save normally.

The models are hidden rather than removed, so they can be made available again once the updates are complete. Customers build strategies only on models that return results.

12-Nov-26

ENG-33150

Recommend:

Automated Ingredient-Based Recipe Recommendations Job

The job that generates recipe recommendations from ingredients is now automated and can be run on demand for any site. Merchandisers get ingredient-based recommendations without manual setup steps, and results are produced more quickly and consistently.

12-Nov-26

ENG-33486

Engage:

Fair Rotation for Equal-Importance Campaigns in Engage

Content and recommendations campaigns with the same importance level are now shown in rotation, regardless of their seed attribute type. Previously, the most recently activated campaign took priority.

Every eligible campaign gets its share of exposure, so shoppers see a wider mix of content and merchandisers get the reach they planned for each campaign.

12-Nov-26

Bug and Support Fixes

The following issues have been fixed in the release version 26.23

Jira #

Module/Title

Summary

General Availability

ENG-33445

Enterprise Dashboard:

Fixed Preview Loading for Dynamic Experiences on Sites with Regions

We have fixed the Dynamic Experiences preview so that recommendations load for sites that use regions. The Recommendations widget template now includes a “Default Region ID” field with a region dropdown, and the preview uses this value when retrieving recommendations. During live execution, the region passed in the instrumentation is always used, and the default region applies only when none is provided.

12-Nov-26

ENG-33602

Enterprise Dashboard:

Fixed Currency Field in Chatbot and Ensemble AI Dynamic Experiences

We have fixed the Currency field in the Chatbot and Ensemble AI Dynamic Experience template, so it is now editable. When a region is selected, the Currency field updates to that region’s currency each time the region changes, and manual edits are honored and saved. Saved experiences open with the correct currency, and the preview and script use it.

12-Nov-26