Release Summary 26.15 | July 23, 2026
The following key features and improvements, along with bug fixes, have been released in Digital Experience Personalization (DXP) in the release version 26.15.
Social Proof
Fallback to Individual Metrics When Aggregated Metrics are Unavailable
The Social Proof Metrics API now falls back to individual product metrics when aggregated attributes, such as master SKU, are not available for a product. Previously, products lacking these attributes, common for non-recommendable products, could not return metrics through the API at all. The system now treats the product itself as the aggregation point when no master SKU is present, ensuring metrics are still returned.
This ensures Social Proof messaging can display accurate metrics across a broader range of products, including those without full catalog attribute data.
Jira: ENG-32962
Contextual Parameters in Prediction API
The Social Proof prediction API now accepts category, channel, and price as context parameters for each product, allowing optimized messages to be returned based on the specific context in which a product is being viewed. Category is derived from the product's primary category, price reflects the product's price, and channel is determined from the user's visit. These parameters are also captured in logging alongside the prediction calls. When the site configuration is enabled, this feature is enabled as part of the existing optimization.
This enables Social Proof messaging to be optimized precisely for the context each shopper is in, supporting more relevant messaging and better engagement across categories, price points, and channels.
Jira: ENG-32575
Configurable Contextual Optimization
Contextual optimization for Social Proof, which factors in category, channel, price, and customer segments, is now controlled through a new site configuration flag, enable_social_proof_contextual_optimization, defaulting to false and requiring explicit enablement per site. The flag has no effect unless site-level optimization is already active, and enabling it for one site has no impact on others. If contextual parameters are partially unavailable, the system degrades gracefully, applying optimization based on whatever signals are available.
This gives clients control over when contextual optimization is introduced, preventing unintended changes to existing sites while allowing selective rollout.
Jira: ENG-32870
Optional Placement Parameter in Social Proof Messaging API
The placement parameter, previously required to specify page type, is now optional in the Social Proof Messaging API for both badging and messaging, using a new parameter that defines single or multi product templates instead. Existing experiences that already define a placement continue to work as before.
This gives more flexibility in configuring Social Proof experiences without requiring changes to experiences already relying on page type.
Jira: ENG-32607
Ensemble AI
Region Error Message Shown Only When Enabling Region
The region-related error message in Ensemble AI now displays only when a user attempts to enable the region option, rather than appearing by default. Sites with 10 or fewer regions enable the option without error; sites with more show the existing error message and the option is not enabled. The region dropdown is also now available for free form styles when regions are enabled.
This removes an error message that previously appeared regardless of user action, reducing unnecessary confusion for merchandisers configuring styles.
Jira: ENG-32823
Enterprise Dashboard
Consistent Product Naming Across the Portal
Main menu items in the portal have been updated for consistent naming, aligning navigation labels with the product names used elsewhere in reporting and configuration. "Search | Browse" is now labeled "Find and Discover," "Content" is now labeled "Engage," and "Recommendations" is now labeled "Recommend," matching how these products are already referenced in related reports. Icons have also been updated to reflect current branding and help documentation is being updated to match.
This creates a more consistent experience across the portal, reducing confusion between naming used in navigation versus reporting and documentation.
Jira: ENG-32822
User Profile Page: Human-Readable Channel Names
The User Profile page now displays the API client key name instead of the raw API key when showing the channel associated with an event, making channel information easier to read and understand. This update applies to Item Views, Recommendation Clicks, Category Views, Searches, and Orders Placed, as well as the channel labels on the session detail page.
This improves clarity for merchandisers reviewing user activity, making it easier to identify which channel an event occurred on without needing to interpret an API key.
Jira: ENG-32758
Shopping Assistant
Chat Reporting: Add to Cart Metrics in Shopping Assistant Transcript Report
The Shopping Assistant Transcript report now includes add to cart data as part of its funnel-level metrics, capturing products added to cart within a chat session alongside existing metrics such as questions, responses, and product clicks. Three new columns have been added to the report: atc_product_ids, ext_user_id, and zone_related_question_ts.
This gives merchandisers a more complete view of how chat engagement translates into shopper actions, connecting product discovery through chat to actual cart activity within the session.
Jira: ENG-32953
MVT
Add to Cart Metrics in Trend Visualizations
Add to carts and ATC Rate are now available as trend metrics in the MVT report, alongside existing metrics like RPV and CVR. Users can select either metric from the dropdown to view its trend visualization, with the current default dropdown selections remaining unchanged. Both metrics are also included in the Daily Data and MVT Visit Data views for conversion metrics.
This gives merchandisers greater visibility into how add to cart performance trends over time, helping them assess the consistency of test results beyond conversion and revenue metrics alone.
Jira: ENG-32442
Data Engineering
Analytics: User Segment Data Available for Reporting
User segment data is now brought into Redshift on a daily basis, making it available for reporting in the same way as custom segments.
This allows merchandisers to report on user segments alongside existing segment data, extending analytics visibility to a broader range of segmentation types.
Jira: ENG-32238
Recommend
MCP Server for AI Shopping Assistants
Recommend now offers an MCP (Model Context Protocol) server that connects the platform to AI-powered shopping assistants, letting large language models retrieve personalized recommendations and shopper affinity data during a live conversation. The server exposes two tools: one that returns ranked recommendations based on shopping intent, such as browsing a category, viewing similar products, or reviewing past purchases, and another that returns a shopper's behavioral affinity profile built from views, purchases, and cart activity, giving the assistant context to personalize the conversation from the first message. It connects to any MCP-compatible client, including off-the-shelf tools and custom chatbot applications, without exposing internal placement details to the LLM.
Session context is supplied by the calling application and threaded through each tool call, along with a tracking token that preserves personalization continuity, and click tracking is built into every recommended product so engagement data flows back into the recommendation engine.
This gives retailers a direct path to bring Recommend's personalization into conversational, AI-driven shopping experiences, without requiring deep integration work to get there.
Jira: ENG-32006, ENG-32381, ENG-32764, ENG-32382, ENG-32148
Other Feature Enhancements
The following feature enhancements and upgrades have been made in the release version 26.15.
|
Jira # |
Module/Title |
Summary |
General Availability |
|---|---|---|---|
|
Social Proof: Message-Level Reporting by Page Type with CTR |
Social Proof message-level reporting now breaks down performance by page type, calculated daily. Each message is tracked using its underlying text template, so variations with a dynamic count or interval are aggregated under a single message. For each message and page type, the report captures visits, views, clicks, CTR, add to carts, converted visits, sales, ATC rate, CVR, and RPV, attributed across all messages and page types a shopper encountered before converting. This gives merchandisers a clearer, more granular view of message performance, making it easier to identify which messages drive engagement and conversion. |
23-Jul-26 |
|
|
Recommend: Reporting on Boosting Rule Performance |
A new rollup report gives merchandisers visibility into how Recommendation Boosting rules impact performance, closing a gap where boosting changes previously had no measurable feedback. Boost data is now captured as part of visits, aggregated or de-aggregated by page type and placement, with the ability to select a specific rule. The report captures clicks, attributable sales, orders, and items per rule, with attribution split across rules when multiple rules boost the same product. This gives merchandisers a clear, rule-level view of which boosting configurations drive engagement and revenue. |
23-Jul-26 |
|
|
Recommend: Wildcard Support for Domain Whitelisting |
The whitelist host names setting for redirect validation now supports wildcards, allowing subdomains to be whitelisted using a pattern such as *.example.com rather than listing each subdomain individually. Any URL matching the wildcard pattern passes redirect validation. This is useful for clients whose development workflows generate dynamic subdomains, allowing them to whitelist an entire domain pattern rather than updating the configuration each time a new subdomain is created. |
23-Jul-26 |
|
|
Dynamic Experiences: Custom Template Support with "My Templates" |
The Dynamic Experiences template picker now includes a "My Templates" tab alongside "Engage Templates," allowing users to create their own templates. From this tab, users can create new templates with custom sections and variables, returned in JSON format as part of the experiences API response. Existing experiences can also be saved directly as templates using a new "Save as Template" button. This gives digital optimization managers a faster way to build new experiences based on their own patterns, rather than starting from scratch each time. |
23-Jul-26 |
|
|
Platform: Omnichannel Order API: Consolidated Status Endpoint |
A new endpoint retrieves omnichannel order processing status across a date range, without requiring a specific job ID. It returns jobs filtered by status, Finished, Failed, Failed with Warning, and Processing, along with order IDs, totals processed, and a breakdown of failed orders with the reason for each failure. Results are returned in JSON and paginated, with a page token for retrieving subsequent pages. This gives teams a straightforward way to identify failed orders across a period and take corrective action, such as re-ingesting after a fix, without looking up individual job IDs. |
23-Jul-26 |
|
|
Platform: Job Status Reports Server with Explorer UI |
A new Job Status Reports Server provides a self-hosted dashboard and REST API for exploring job execution history from existing job and event status data, answering questions such as which orders failed in a given time window for a site, without requiring direct database access or engineering involvement. The API supports paginated job listings filtered by site, time range, application, and status, direct lookup by job ID, and event details per job, with configurable custom columns so relevant fields can be surfaced per service without code changes. The accompanying explorer UI supports dark mode, date and time filtering, load-more pagination, and direct job ID lookup, giving teams a straightforward way to investigate job and order status without querying the underlying data directly. |
23-Jul-26 |
|
|
Chatbot: Chat Transcript Consumer: Timezone and MVT Seed Type Alignment |
The Chat Transcript Consumer now accounts for site timezone when storing chat transcripts in HDFS, ensuring data lands in the correct date-based path for each site. The site-denoted user ID is also now populated based on the MVT seed type configured in the portal, rather than a fixed default. This ensures chat transcript data aligns accurately with each site's configured timezone and seed type, supporting reliable downstream reporting. |
23-Jul-26 |
|
|
Platform: Omnichannel API: Order and User ID Tracking for Successful Orders |
Common status event tracking now records the order ID and user ID for successful orders, in addition to failed orders, providing consistent detail across both outcomes. This gives teams complete visibility into order processing status, making it easier to trace and reference successful orders alongside failures. |
23-Jul-26 |
Bug and Support Fixes
The following issues have been fixed in the release version 26.15.
|
Jira # |
Module/Title |
Summary |
General Availability |
|---|---|---|---|
|
Data Engineering: Cart Page Not Showing Social Proof Messages with Inventory Messaging Enabled |
We have fixed an issue where the cart page did not display Social Proof messages when inventory messaging was enabled and the configured message did not include the @inventorycount placeholder. Messages now display correctly whether or not the @inventorycount placeholder is included in the message text. |
23-Jul-26 |
|
|
Enterprise Dashboard: Error While Saving New Template in Dynamic Experiences |
We have fixed an issue where saving a new template under My Templates in Dynamic Experiences failed with a JavaScript error. |
23-Jul-26 |
|
|
Ensemble AI: Style Name Showing as Null in Ensemble AI Report |
We have fixed an issue where the style name appeared as null in the Ensemble AI report. |
23-Jul-26 |
|
|
Ensemble AI: Ensemble AI Menu Positioning Issue |
We have fixed an issue where the Ensemble AI menu option did not appear in a consistent location within the Social Proof menu across navigation. The Ensemble AI icon has also been corrected to display properly. |
23-Jul-26 |
|
|
Chatbot: Chatbot Not Answering Valid Configuration Questions |
We have fixed an issue where the chatbot responded with a generic "not sure how to help" message for valid questions about specific configuration options, such as the "Enable backfill" checkbox for Preferred Strategies. |
23-Jul-26 |
|
|
Social Proof: Badge Selector Element Value Not Saving in Social Proof Badging Configuration |
We have fixed an issue where the selector element value in Social Proof badging configuration was not retained after saving, causing badge configurations to remain incorrect or incomplete. |
23-Jul-26 |
|
|
Platform: Omnichannel Order API Failures with Zero-Priced Line Items |
We have fixed an issue where Omnichannel Order API uploads failed for orders containing line items with a unit price of 0, such as discounted or free add-on items. Order uploads with a price or quantity of 0, missing, or non-zero now process successfully. |
23-Jul-26 |
|
|
Enterprise Dashboard: Custom Attribute Names Not Visible in Affinity Configuration |
We have fixed an issue where custom attribute names, such as hasCashback and over18Only, were not displayed in the Affinity Configuration UI after being added, even though their corresponding values appeared correctly. |
23-Jul-26 |
|
|
Recommend: XSS Validation Error When Saving Experience Templates |
We have fixed an issue where saving a valid experience template through the templates API was incorrectly blocked by XSS validation, due to legitimate HTML content such as stylesheet links, form elements, and template placeholders being flagged as invalid. |
23-Jul-26 |
|
|
Recommend: User Affinity Config Cache Failure with Hyphenated Attribute Names |
We have fixed an issue where the User Affinity Config cache failed to build when attribute names containing hyphens, such as "best-use," were used in combination, due to a SQL syntax error generated during query construction. |
23-Jul-26 |