The dashboard includes analytics for: 

• Personalization reach. 

• Personalized and default experience deliveries. 

• Delivery errors. 

• Delivery health. 

• Experience activity across channels and languages.  

You can use these insights to investigate personalization issues, identify underperforming experiences, and monitor personalization delivery across your sites.

Audience Builder AI assistant

The new AI Assistant makes it easier to build audiences from natural-language descriptions while keeping full control over the resulting rules.

• Create audiences using natural-language prompts - generate audience rules from simple text descriptions.

• Refine rules through the Chat tab - add, remove, or modify rules in an existing audience.

• Use suggested prompts - start with suggested prompts for common audience segments.

• Switch between AI-assisted and manual editing - switch between AI-generated rules and manual configuration as needed.

Event and session rule limits

A new toggle gives you more control over whether rule limits are evaluated across a single session or multiple sessions.

• Apply limits to event and session rules - define how many times an event or session attribute must occur for a profile to qualify.

Once connected, SitecoreAI creates a dedicated AI Discovery folder within the brand context’s file structure. The data gives agents in Agentic studio richer brand context, including insights into visibility, audiences, competitors, topics, prompts, and cited sources. Agents can use this context to make content generation, recommendations, and other brand-related tasks more informed and relevant.

What's included

• Connect brand context to Scrunch - connect brand context to a configured  brand in Scrunch to generate the AI Discovery files.

• Access AI Discovery insights - explore the read-only AI Discovery files with insights into your brand’s visibility and performance across AI platforms, topics, audiences, competitors, and cited sources.

• Keep brand data up to date - AI Discovery files refresh automatically each day so your Brand context reflects the latest available data from Scrunch.

• Manage Scrunch connections - disconnect Brand context from Scrunch at any time to stop data synchronization.

What's included 

• Custom schema extensions - add profile attributes beyond the standard profile schema.  

• Flexible data structures - model customer data using simple and complex values. 

• Audience support - after the schema is published, use profile extensions as audience criteria when personalizing your website or targeting content. 

Profile import now supports bulk profile deletion in SitecoreAI. You can upload a JSONL deletion batch through the Profile Import API to permanently delete profiles for privacy and data-subject deletion workflows.

A deletion record requires only one identifier that matches the profile; you don't need to provide every associated identifier. You can also monitor batch status and download results to confirm deleted and unmatched profiles.

• When creating variants in A/B/n tests, you can now use Copy component and reuse items available for container components. This option copies the container and its associated content items without creating new content item versions, making it easier to test different content layouts and structures without duplicating or maintaining additional content items.

The latest release of Page builder includes the following resolved issue:

• Updated personalization status indicators to reflect live production behavior after publish, ensuring authors can accurately see when published personalizations are active and running in production.

Most implementations are not expected to be affected. Standard scenarios that use primitive values such as strings, numbers, and booleans will continue to function as before. The primary impact area is custom code that stores complex objects in PipelineArgs.CustomData and later serializes and deserializes those values.

After this change, custom classes, interface-based objects, abstract types, and heterogeneous object collections may no longer be automatically restored to their original .NET types during deserialization. Depending on the implementation, values may be returned as generic JSON objects, resulting in casting failures, missing data, or other compatibility issues.

Action required

Customers should review any serialization round-trips involving PipelineArgs.CustomData and consider using strongly typed models or explicitly reconstructing custom objects after deserialization.

To support validation and migration, a feature flag will be provided that allows customers to enable the new behavior before it becomes the default and temporarily revert if compatibility issues are discovered. We strongly recommend testing any customizations that serialize and deserialize PipelineArgs data during the notification period to identify and address potential impacts before rollout.

We also strongly recommend that you test the new security implementation before the base image release scheduled for November 29 by creating a patch configuration file and redeploying your environment. 

Example configuration to enable the changes:

<sitecore>
  <settings>
     <setting name="PipelineArgs.DisableUnsafeSerialization">
     <patch:attribute name="value">true</patch:attribute>
     </setting>
  </settings>
</sitecore>

• The selected value comparison for Droplist field types is now case insensitive, properly displaying the selected item in the UI.

• Fixed bug where Content Editor was not displaying child items under /sitecore/content/Collection/Site/Media in the content tree.

• Resolved defect where page designs list could display incorrect entries if they had been previously viewed by a user with limited access permissions.

• Dictionary fallback updates triggered by deleting dictionary entries are now processed asynchronously, reducing the likelihood of long-running requests and timeouts.

• Fixed issue with incrementally published items with non-default language not updating changes in paths when parent folder is moved.

• Upgraded the Telerik components included in the Sitecore AI Base Image from 2020.3.1021.45 to 2026.3.812 in order to improve platform quality, maintainability, and long-term supportability.

Breaking Change:  Changes to JSON serialization

SitecoreAI is introducing an improvement to PipelineArgs serialization and deserialization by changing the underlying configuration from TypeNameHandling.All to TypeNameHandling.None. This change removes embedded .NET type metadata from serialized payloads and enforces more restrictive deserialization behavior.

Most implementations are not expected to be affected. Standard scenarios that use primitive values such as strings, numbers, and booleans will continue to function as before. The primary impact area is custom code that stores complex objects in PipelineArgs.CustomData and later serializes and deserializes those values.

Action required

Customers should review any serialization round-trips involving PipelineArgs.CustomData and consider using strongly typed models or explicitly reconstructing custom objects after deserialization.

To support validation and migration, a feature flag will be provided that allows customers to enable the new behavior before it becomes the default and temporarily revert if compatibility issues are discovered. We strongly recommend testing any customizations that serialize and deserialize PipelineArgs data during the notification period to identify and address potential impacts before rollout.

We also strongly recommend that you test the new security implementation before the base image release on November 29, by creating a patch configuration file and redeploying your environment. 

As part of this release, the bundled Telerik assemblies have been upgraded to a newer supported version, 2026.3.812. Sitecore AI functionality that relies on these components has been validated and is expected to continue operating as before.

For most customers, no action is required. However, if you have custom solutions that directly reference specific Telerik assemblies or versions, we recommend reviewing and testing those customizations after upgrading. Standard .NET assembly versioning and binding redirect mechanisms can be used to maintain compatibility if needed.

We recommend upgrading to the latest release to benefit from these improvements.