Skip to content

August 2026 Release Note

These are the Treasure AI features and improvements released in August, 2026.


AI Studio

Enhancement to the Agent Sandbox: A Clean Environment for Every Chat

Posted on: Aug 12, 2026

  • Every chat in Treasure AI Studio now runs in its own environment: files in a work folder persist as before, while anything the AI writes elsewhere — working files, temporary files, packages it installs — belongs to that one chat and is removed with it.
  • Customer records, exports, and downloaded data the AI leaves outside a work folder no longer accumulate in a location no file browser showed, closing off a place where sensitive data could sit unnoticed long after the work was done.
  • Files that had accumulated in the AI's home directory were moved into a MIGRATED_HOME folder inside the work folders that had them, where you can browse, keep, or delete them like any other file.
  • Where your files appear in the sidebar, on the work folder page, and in the file viewer is unchanged, and no action is required.
  • Learn more by visiting the Security Overview page.

New Feature: Work Folder Actions

Posted on: Aug 13, 2026

  • Actions let you save a recurring workflow — a weekly report, a data check, a review — in a Treasure AI Studio work folder and rerun it anytime, with one click of Run on the work folder page or by typing /<name> in a chat.
  • Designed for anyone who repeats the same multi-step requests, actions replace re-typed prompts with a saved definition: Run starts a fresh chat that is already named after the action and marked with a ⚡ badge in your chat lists.
  • An action is a work folder skill — ask the AI to save a workflow as an action, and convert between skill and action anytime from the work folder page.
  • Learn more by visiting the Actions page.
Actions and Skills section on a work folder page, with an action row showing its description and Run button above a collapsed Skills sub-group

New Feature: Access All Work Folders from Any Chat

Posted on: Aug 13, 2026

  • In every Treasure AI Studio chat, the AI can now read and write files in all of your work folders, so cross-folder tasks — comparing datasets, consolidating reports, drawing on shared reference data — happen in a single conversation instead of being split across chats or re-uploaded.
  • Point the AI at a file in another folder by naming the folder and file in your message, or by dragging the file from the sidebar file tree into the composer to insert its exact path; the AI's output still lands in the chat's own work folder unless you ask for another destination.
  • Files also move between work folders now: drag a file or folder onto another work folder in the sidebar or on a work folder page, and if the destination already has an entry with the same name the move is refused with an error, so nothing is overwritten.
  • Learn more by visiting the Work Folders page.

New Feature: Action Schedules

Posted on: Aug 25, 2026

  • Action schedules run a saved action in a Treasure AI Studio work folder on a repeating cadence — daily, weekdays, weekly, monthly, or a custom cron expression — starting a fresh chat for each run without anyone clicking Run.
  • Designed for anyone who already runs an action on a routine, schedules remove the last manual step, so the work is finished before you go looking for it and only a failed run needs your attention.
  • Each run happens under your account, in the action's work folder, and the chat it produces carries a clock badge so you can tell scheduled runs from the ones you started yourself.
  • The new Automations section on the work folder page lists each schedule with its next run, turns one on or off with a switch, and opens a run history that merges scheduled and manual runs so you can see what every run produced.
  • Learn more by visiting the Action Schedules page.
Automations section on a work folder page showing a schedule that runs an action every day at 9:00 AM in Asia/Tokyo, with its next run time, an enable switch, and a succeeded last run linking to the chat it produced

Treasure AI Voice

Enhancement: Improved Connection Stability and Faster Uploads

Posted on: Aug 17, 2026

  • Treasure AI Voice for iOS improves Bluetooth connection stability, reducing cases where reconnecting to a device would get stuck without ever completing or failing.
  • Wi-Fi file transfers are more reliable: an interrupted transfer now resumes automatically instead of requiring a device restart.
  • Recordings are now compressed before upload, making uploads faster and use less data.
  • The login page now supports resetting a forgotten password.
  • This release also includes a number of smaller bug fixes and stability improvements.
  • Learn more by visiting the Treasure AI Voice page.

Engage Studio

New Feature: LINE OA Always-on Campaigns

Posted on: Aug 03, 2026

  • LINE Official Account (LINE OA) is now a first-class delivery channel for Always-on Campaigns in Engage Studio, bringing LINE messaging into continuous, Journey-orchestrated campaign workflows.
  • This extends the Always-on Campaign lifecycle (Draft → Planned → Live → Paused → Finished) to LINE OA, placing it on parity with Email and Push for marketers running recurring promotions, lifecycle journeys, and triggered activations.
  • Supported message types include text, Imagemap, and Flex messages. Flex messages enable flexible message formatting.
  • Learn more by visiting the LINE OA Campaigns page.

Enhancement to LINE Delivery Logs: Test Send Identification

Posted on: Aug 10, 2026

  • LINE delivery logs now include a test_mode column in the line_events table of the delivery_line_{system_identifier} delivery log database, identifying whether a record was created by Send Test or by an actual campaign send.
  • Designed for marketers and analysts who query LINE delivery data, test_mode makes it possible to exclude test sends from delivery reporting and analysis, bringing LINE delivery logs to parity with email delivery logs, which already provide this field.
  • The column is populated automatically with no action required: test_mode is true for Send Test records and false for normal campaign sends.
  • Records created before this release have a null value for test_mode, so queries that span historical data should use a condition such as test_mode IS NOT TRUE rather than test_mode = false when filtering for campaign sends.
  • Learn more by visiting the LINE Campaign Reporting page.

Enhancement to Display Conditions: Highlight Conditional Rows

Posted on: Aug 20, 2026

  • Engage Studio now highlights rows with Display Conditions in the email and popup editors, making conditional content easier to identify while you edit a message.
  • Designed for marketers who tailor content to recipient or viewer attributes, row highlighting makes it easier to review conditional logic and avoid accidental edits to conditional content.
  • The visual indicator does not change how Display Conditions work in previews or rendered HTML.
  • Learn more by visiting the Email Personalization page.
Engage Studio email editor showing a row with a Display Condition highlighted by a blue outline

Integration Hub

New Connector: Google Enhanced Conversions Export (Data Manager API)

Posted on: Aug 04, 2026

  • The Google Enhanced Conversions Export connector now uses the Google Data Manager API, enabling customers to continue sending enhanced conversion events to Google Ads after Google's migration from the legacy Google Ads API.
  • Designed for marketers who track offline conversions and enhanced conversions for leads through Google Ads, the connector supports offline, online, and store sales conversions with a unified event model that requires no developer token.
  • The connector supports hashed user identifiers (email, phone, address), conversion metadata (action ID, timestamp, value, currency, transaction ID, GCLID, and consent), and request-level diagnostics via request IDs for troubleshooting.
  • Learn more by visiting the Google Data Manager for Conversions Export Integration page.

Enhancement to Microsoft Azure Blob Storage Import Connector

Posted on: Aug 05, 2026

  • The Microsoft Azure Blob Storage Import connector now supports Shared Access Signature (SAS) authentication as an alternative to Storage Account Keys.
  • Designed for enterprise customers with strict security policies that prohibit sharing Storage Account Keys, SAS authentication enables limited-scope, time-bound access to Azure Blob Storage containers.
  • Customers can now authenticate using either Storage Account Keys (existing) or SAS tokens, without any changes to existing workflows or existing Storage Account Key configurations.
  • Learn more by visiting the Microsoft Azure Blob Storage Import Integration page.

Enhancement to Salesforce Marketing Cloud V2 Export Connector

Posted on: Aug 11, 2026

  • The Salesforce Marketing Cloud V2 Export connector now supports composite primary keys (up to 4 fields) when creating new Data Extensions, eliminating the need for manual workarounds.
  • Designed for integration engineers and marketing operations teams using SFMC activations from Treasure Data, composite primary keys enforce data integrity for records with multiple unique identifier fields.
  • The connector uses a new primary_columns configuration parameter (a list) alongside the existing primary_column parameter for backward compatibility; existing single-primary-key configurations continue to work unchanged.
  • Learn more by visiting the Salesforce Marketing Cloud V2 Export Integration page.

Enhancement to Meta Dataset Quality Import Connector: Offline Event Type Support

Posted on: Aug 11, 2026

  • The Meta Dataset Quality Import connector now supports offline event type, enabling quality score tracking for offline conversions alongside existing web event quality metrics.
  • Designed for advertisers running omnichannel campaigns on Meta, offline event quality scores provide visibility into conversion signal quality from physical stores and offline channels, helping optimize Advantage+ campaigns.
  • The connector now supports four offline sub-scores (composite, match_key, frequency, freshness) and automatically maps the offline field set response from Meta's Dataset Quality API to your destination table.
  • Learn more by visiting the Meta Dataset Quality Import Integration page.

New Connector: SFTP Parquet Import

Posted on: Aug 26, 2026

  • The SFTP Parquet Import connector directly imports Parquet files from an SFTP server into Treasure AI, removing the need to convert files to CSV before loading them.
  • Designed for data engineers and integration teams whose data arrives over SFTP, the connector retains Parquet schema information and removes a conversion step that adds batch time, storage use, and operational work.
  • Create an SFTP Parquet connection in Integration Hub, use a folder path and optional regular expression to select files, and rely on incremental loading to skip already processed paths by default.
  • The connector is available in Early Access.
  • Learn more by visiting the SFTP Parquet Import Integration page.

Enhancement to Google Ads Import Integration V2: Manager Account Support for Custom Dates

Posted on: Aug 30, 2026

  • Google Ads Import Integration V2 now uses the configured Ads Manager Account for every Google Ads API request, including internal validation for custom date ranges.
  • Designed for customers whose Google Ads account is accessed through a manager account, this enhancement lets historical reports and Click View custom queries run for a specified date instead of failing during validation.
  • Existing Google Ads Import Integration V2 configurations with an Ads Account and Ads Manager Account continue to use their configured account IDs across report and custom-query workflows.
  • Learn more by visiting the Google Ads Import Integration V2 page.

Open Catalog Access

New Feature: Open Catalog Access

Posted on: Aug 16, 2026

  • Open Catalog Access shares tables managed in Complete CDP with your cloud data warehouse without copying data: you publish tables in Apache Iceberg format, and Snowflake, Databricks, or BigQuery queries them in place through a managed Iceberg catalog — no export jobs, no transfer delays, no second copy to govern.
  • Designed for organizations whose data science, BI, and AI engineering teams work in external warehouses, it replaces per-team custom exports with governed, read-only access to the same customer tables, while Complete CDP remains the system of record.
  • The Snowflake and Databricks integrations are generally available, and the BigQuery integration is available as an experimental feature.
  • Open Catalog Access is not enabled on accounts by default; contact Technical Support or your Customer Success representative to enable it.
  • Learn more by visiting the Open Catalog Access page.

Audience Studio

Enhancement to Behavior Activation: Multi-Table Support

Posted on: Aug 11, 2026

  • Behavior Activation now supports activating data from up to four behavior tables in a single activation, up from the previous limit of one table.
  • Designed for Marketing Operations teams whose behavior data spans multiple tables — such as separate tables for email addresses, phone numbers, and transactions — this removes the need to manually restructure that data into a single table before activation.
  • This capability is available now through the Audience API, for both segment activations and journey activation steps, using the new syndicationBehaviors array field in place of the deprecated syndicationBehavior field, which remains supported for single-table activations; no feature flag or enablement request is required.
  • The Treasure AI Console does not yet support creating or editing multi-table activations, and blocks saving an activation that has more than one behavior table configured until Console support ships in a future release, to prevent accidental loss of the additional tables' settings.
  • Learn more by visiting the Multi-Table Behavior Activations and Configuring Multiple Behavior Tables pages.

Journey Orchestration

New Feature: Re-randomize on Re-entry

Posted on: Aug 17, 2026

  • Re-randomize on re-entry lets profiles receive a fresh variant assignment when they re-enter a batch Journey and reach a configured A/B Test step.
  • Designed for recurring Journeys where Sticky assignments can create stale cohorts, the feature re-randomizes across all configured variants, including Control.
  • Configure the behavior independently for each A/B Test step; profiles already in the Journey are not reassigned mid-Journey, and Profile Cap and Declare Winner are not supported on re-randomized steps.
  • Learn more by visiting the Re-randomize on Re-entry page.

Platform

Deprecation: Remember Me Feature Removed from Login Page

Posted on: Aug 05, 2026

  • The Remember Me feature has been removed from the Treasure AI login page to enhance security and align with enterprise authentication best practices.
  • This change ensures that sessions respect configured idle timeouts and eliminates automatic re-authentication paths, meeting the security requirements of regulated enterprises and customers with strict session management policies.
  • Users will need to log in with their credentials each time they access Treasure AI. For organizations using SAML or federated authentication, login remains streamlined through your identity provider.
  • Learn more by visiting the Control Panel page.

New Feature: Overage Notification Settings in Treasure AI Console

Posted on: Aug 31, 2026

  • Overage Notification settings in the Treasure AI Console let customers view and update six account-level options: notification language, approaching-limit and overage alerts, weekly and monthly usage reports, and recipient email addresses.
  • Designed for customer account administrators, self-service settings make it easier to keep usage notifications aligned with each account's communication preferences without contacting your Customer Success Manager.
  • If you want to customize per-metric threshold settings, please contact your Customer Success Manager.
  • Learn more by visiting the Overage Notification Settings page.

Treasure Code (CLI)

New Feature: Personalization Studio Campaign Commands

Posted on: Aug 31, 2026

  • Treasure Code now includes the first commands for managing Personalization Studio campaigns from the command line, letting engineers and automation workflows view and interact with campaigns without opening the console.
  • You can view a campaign's full details — its status, assigned audiences, content spots, and active personalizations — with a single command.
  • A session context command stores the active campaign for your terminal session, removing the need to supply a campaign identifier on every subsequent command when running a sequence of steps against the same campaign.
  • Learn more by visiting the Treasure Code page.

Enhancement: Always-on Campaign Management via CLI

Posted on: Aug 14, 2026

  • Always-on campaigns in Engage Studio can now be pulled, edited locally as YAML files, and pushed back through Treasure Code, bringing them into the same version-controlled and scriptable workflow already available for email templates and standard campaigns.
  • Learn more by visiting the Treasure Code page.

Fix: Email Preheader Text Now Preserved on Push

Posted on: Aug 31, 2026

  • Email preheader text is now preserved when pushing email templates, campaigns, and always-on campaigns through Treasure Code; previously it was silently dropped, causing emails to render without preview text in recipients' inboxes.
  • No changes to existing configurations are needed.

Fix: Composable Audience Studio Activation Settings and Preview Count

Posted on: Aug 19, 2026

  • Two issues affecting Composable Audience Studio commands have been fixed: activation behavior settings were silently dropped on push and pull round trips, and campaign previews always returned a count of zero regardless of actual audience size.
  • Both fixes apply automatically with no changes to existing configurations.