# October 2026 Release Note These are the Treasure AI features and improvements released in October, 2026. ## AI Studio ### New Feature: Agent Channels with SDK Apps *Posted on: Oct 21, 2026* - AI Studio now offers Agent Channels, starting with SDK Apps, so an application you host can run Treasure AI Studio chats against one shared work folder. - Designed for teams that want to bring an AI Studio agent into their own product or service. - Folder managers create SDK Apps under **Settings → Agent Channels**, and account administrators can limit creation to administrators. - Learn more by visiting the [Agent Channels](/products/ai-studio/agent-channels/) page. ### New Feature: MCP Endpoints in Agent Channels *Posted on: Oct 22, 2026* - Agent Channels now include MCP Endpoints, which publish one shared work folder as a remote MCP server so an MCP client can run the folder's Actions, start and continue chats, and read the results. - Designed for teams that already work in an MCP client and want to use their team's Actions from it, with every chat running in Treasure AI Studio as the endpoint's executor. - Folder managers create MCP Endpoints under **Settings → Agent Channels**, and the client connects with the endpoint's URL and a connection token sent in the `Authorization` header. - Learn more by visiting the [Agent Channels](/products/ai-studio/agent-channels/) page. ### Enhancement to MCP Endpoints: OAuth Sign-In *Posted on: Oct 28, 2026* - MCP Endpoints now support OAuth, so each person connects their own MCP client by signing in to Treasure AI Studio and allowing access, and the client's chats and Actions run as that person, with their own permissions and credits, in the work folders they can already use. - Designed for accounts where many members want to use Studio from the MCP client they already work in, without sharing a connection token or running everyone's work as one executor. - Account administrators create OAuth endpoints under **Settings → Agent Channels**, and each person adds only the endpoint's URL to their client; no token or header is needed. - Learn more by visiting the [OAuth MCP Endpoints](/products/ai-studio/agent-channels/mcp-oauth) page. ### Enhancement to AI Studio: Ask and Auto for tdx Command Approval *Posted on: Oct 30, 2026* - The command approval selector in Treasure AI Studio now lets you choose **Ask** (the default), which instructs the agent to request approval before `tdx` commands, or **Auto**, which runs them without the built-in approval card. - Marketers and data analysts can keep an approval checkpoint for CDP commands or let a trusted task continue without repeated command approval prompts. - Choose the mode in the Web or Desktop chat composer; the setting is saved per chat and does not turn off ordinary clarification questions. - Learn more by visiting the [AI Chat Interface](/products/ai-studio/chat/chat) page. ## Engage Studio ### New Feature: Landing Pages *Posted on: Oct 01, 2026* - Engage Studio now supports Landing Pages, letting you publish completed static website packages from ZIP files to Active custom subdomains as branded public campaign destinations. - Marketers prepare the page files outside Engage Studio, while web administrators configure and activate the custom subdomains. - Landing Pages can also host a branded unsubscribe confirmation page on an Active custom subdomain, but the page does not process the unsubscribe request itself. - Learn more by visiting the [Landing Pages](/products/marketing-cloud/engage-studio/landing-pages) page. ### Enhancement to Mobile Push: Time-Sensitive Delivery *Posted on: Oct 02, 2026* - The **Time-sensitive delivery** switch lets you ask Firebase Cloud Messaging (FCM) to use high priority for Android notifications in one-off and always-on Mobile Push campaigns. - Designed for time-sensitive alerts, the setting asks FCM to attempt delivery as soon as possible, including when an Android device is in Doze mode, but it does not guarantee immediate delivery because FCM, network, and device conditions can affect timing. - The switch is off by default, is included in test sends, and leaves iOS delivery unchanged. - Learn more by visiting the [Mobile Push Campaign](/products/marketing-cloud/engage-studio/channels/mobile-push) page. ### New Feature: Duplicate In-App and In-Browser Messages *Posted on: Oct 22, 2026* - Engage Studio now lets you duplicate an In-App Message or an In-Browser Message, creating a new message in the same channel from the content of the original so you can reuse proven content for a similar initiative without rebuilding it. - Duplication copies only the Message Content tab, including message text, images, links, and HTML. The Target Audience, Personalization section, and delivery settings are not copied, so the new message does not inherit targeting or delivery settings that could conflict with your new campaign. - The new message is named after the original with **(Copy)** appended, the same as campaign duplication. Review and configure it before you publish. - To duplicate a message, select **Duplicate** from the message list menu or from the message detail page. - Learn more by visiting the [Mobile In-App Messaging](/products/marketing-cloud/engage-studio/channels/mobile-inapp) and [Engage Studio Experiences](/products/marketing-cloud/engage-studio/experiences) pages. ## Personalization Studio ### New Feature: Pop-up Messages *Posted on: Oct 12, 2026* - Personalization Studio now supports Pop-up messages (in-browser messaging), so marketers can display promotions, announcements, and onboarding guidance to selected audiences on selected website pages. - Marketers can build messages with text, images, links, and buttons, choose one-, two-, or three-column layouts (including a 3/7 split), set one of seven positions, and preview the result on desktop, tablet, or mobile before launch. - Pop-up messaging is in Release Preview (Beta); the feature must be enabled for your account, and campaigns require Real-Time Personalization 2.0. - Learn more by visiting the [Creating Campaigns](/products/marketing-cloud/personalization-studio/creating-campaigns) page. ### Update: Audiences Built in the Rule Editor *Posted on: Oct 30, 2026* - Audiences in Personalization Studio are now built in the rule editor from batch attributes and real-time attributes, instead of importing batch segments from Audience Studio. You can also rename an audience and edit its rules. - Batch and real-time conditions go in separate rule sets within the same audience. - Batch behaviors, lookup catalog, real-time drag-on rules, and importing batch attributes aren't supported as audience conditions yet. - Learn more by visiting the [Managing Audiences](/products/marketing-cloud/personalization-studio/managing-audiences) page. ## Data Workbench ### New Feature: CREATE TABLE AS SELECT Support for Hive *Posted on: Oct 21, 2026* - The Hive query engine now supports the `CREATE TABLE AS SELECT` (CTAS) statement, so data engineers can create a table directly from the result of a `SELECT` query using HiveQL, without switching to Trino. Together with the `CREATE TABLE` support added in July and the `DROP TABLE` support added in September, you can create tables, save query results as tables, and delete tables with Hive alone. - You can specify `COMMENT`, `TBLPROPERTIES` (`expire_days` and User-Defined Partitioning), and `IF NOT EXISTS`. If a query fails during execution, the created table is dropped automatically. For what to do in the rare case that a table remains, see the failure behavior described on the page. - Learn more by visiting the [CREATE TABLE AS SELECT Statement Syntax](/products/customer-data-platform/data-workbench/queries/sql-reference/create_table_as_select) page. ## Integration Hub ### Enhancement to OneTrust Import Connector: V4 Collection Point Filtering and Rolling Incremental Loads *Posted on: Oct 01, 2026* - The OneTrust Import Connector's Data Subject Profile (API V4) data type now supports Collection Point filtering and rolling incremental loads, bringing V1-style per-Collection Point ingestion to V4. - Set `collection_point_guid` to retrieve data for one Collection Point; if you omit it, the connector keeps the existing behavior of retrieving data across all Collection Points. - With `incremental: true`, `end_time` is optional. The connector fixes an upper bound for each run and advances the next run's start time to that bound only after all pages are retrieved and written successfully, even when the run returns no records. - This enhancement is designed for data engineers moving OneTrust Data Subject Profile jobs from API V1 to API V4 who need to keep website consent data separate and avoid repeatedly requesting the same empty time range. - Learn more by visiting the [OneTrust Import Integration](/int/onetrust-import-integration) page. ### New Connector: OpenAI Ads Conversion API Export *Posted on: Oct 21, 2026* - The OpenAI Ads Conversion API Export connector sends server-side conversion events from Treasure AI to an OpenAI Ads Pixel, supporting both standard and custom events. - Designed for marketing teams that measure conversions in OpenAI Ads, the connector lets you send events from your own data instead of relying only on the Pixel, with field mapping from query results, pre-delivery validation, and normalization and SHA-256 hashing of user identifiers such as email addresses and phone numbers. - Each delivery is split into batches of up to 1,000 events, and setup requires an existing OpenAI Ads Pixel ID and Conversions API key. - The connector is available in Early Access. - Learn more by visiting the [OpenAI Ads Conversion API Export Integration](/int/openai-conversion-api-export-integration) page. ### New Connector: Equativ Export *Posted on: Oct 24, 2026* - The Equativ Export connector sends segment taxonomy and user profile membership from Treasure AI to Equativ, so audiences you build in Treasure AI can be activated for ad targeting in Equativ. - Designed for teams that use Equativ for ad delivery, the connector writes export files to the Google Cloud Storage buckets that Equativ provides, and profile exports can be routed to the US, EU, or APAC Equativ environment. - Start a new setup with a taxonomy export, because Equativ needs each segment to exist before profile data that references it is sent. - The connector is available in Early Access. - Learn more by visiting the [Equativ Export Integration](/int/equativ-export-integration) page. ### Enhancement to Microsoft Advertising Export Connector: REST API Support *Posted on: Oct 25, 2026* - The Microsoft Advertising Export connector now supports the Microsoft Advertising REST API as an alternative to the SOAP API, which Microsoft is retiring. - Designed for marketing operations and data engineers who export audiences to Microsoft Advertising and need to move off SOAP before it is retired. - To switch, select **Use REST API** in the export settings before Microsoft Advertising retires the SOAP API on January 31, 2027. - Learn more by visiting the [Microsoft Advertising](/int/microsoft-advertising) page. ## Treasure AI Voice ### Update: AI Voice App 1.9.2 - Treasure AI Voice app 1.9.2 simplifies the interface for bulk-transferring recordings, making it easier to transfer multiple recordings at once. - Metadata attached through deep links is now displayed in the app. - This release includes improvements to device connectivity. ## Composable CDP ### Enhancement to Composable Audience Studio for BigQuery: Workload Identity Federation (WIF) Authentication *Posted on: Oct 14, 2026* - Composable Audience Studio for BigQuery now supports Workload Identity Federation (WIF) authentication, using short-lived credentials to connect to your warehouse. - Designed for data engineers and IT administrators whose security policies restrict long-lived service account keys, WIF lets teams connect audience workflows to BigQuery without storing or rotating service account JSON keys. - To get started, select **Workload Identity Federation** in a **Google BigQuery** authentication in **Integration Hub** and follow the setup guide to configure Google Cloud and supply the **ADC Keyfile**. - Learn more by visiting the [Set Up BigQuery Authentication](/products/customer-data-platform/composable-cdp/audience-studio/bigquery/set-up-bigquery-authentication) page. ## Real-Time ### New Feature: Delete All Real-time Profile Data - **Delete All Real-time Profile Data** is currently in beta. On the **Administration** tab of a Parent Segment's real-time configuration, select **Delete All Real-time Profiles** to delete the real-time profiles and start again from an empty state. To get beta access, contact your account representative. - Designed for teams in a proof of concept or early implementation who need to delete test data or recover from misconfigured attributes without opening a support request. - Only users with the MasterSegmentConfig permission and edit access can see the **Administration** tab. You must type `DELETE` to confirm. - Learn more by visiting the [Delete All Real-time Profile Data](/products/customer-data-platform/real-time/delete-all-real-time-profile-data) page.