Skip to main content
Week of March 2, 2026
Rosetta StoneMajor ReleaseNew Feature

Custom Attributes UI

The Rosetta Stone Attributes page now includes a Create Attribute button, making it easy to define your own custom attributes directly from the UI. Custom attribute creation has been available via the API, and now you can do it without writing any code.
  • Rosetta Stone can auto-generate the display name, name, and description for you using AI-assisted fields
  • Custom attributes are private to your company by default, with the option to share more broadly
  • View documentation →
Create Attribute panel showing AI-generated display name, name, and description for an ACME Partner Tier attribute
Week of March 2, 2026
IntegrationsNew Feature

Meta Conversions API Connector

You can now send conversion events — purchases, leads, checkouts, and more — directly to Meta’s Conversions API for improved ad attribution and campaign optimization on Facebook and Instagram.
  • Supports offline, online, and app conversion events
  • Map your conversion data to the meta_conversion_event Rosetta Stone attribute for automatic transformation, hashing, and delivery
  • PII fields are SHA-256 hashed before transmission — raw customer data is never sent to Meta
  • View documentation →
Facebook Connector showing the Deliver Conversion Events connection type
Week of March 2, 2026
PlatformImprovement

Data Planes Updates

The Settings → Data Planes page has been refreshed with an updated list view and a new Insights tab. The list view now includes sortable columns for ID, display name, unique name, and creation date. The Insights tab provides at-a-glance operational metrics — total jobs, completed and failed counts, average runtime, P95 runtime, and average queue time — along with visualizations for job volume, status distribution, jobs by group, a duration heatmap, and a queue time vs. runtime scatter plot.
Data Planes Insights tab showing job volume, status distribution, duration heatmap, and queue time vs. runtime charts
Week of February 23, 2026
PlatformWorkflowsNew FeatureImprovement

Update Sample Action

You can now request data samples directly from the UI on all supported data planes. Open a dataset, click the Actions dropdown, and select Update sample to trigger a sampling job.
Actions dropdown menu showing Update sample option under Maintenance
Once selected, the Update sample button appears in the top bar for quick access.
Update sample button in the dataset toolbar
A data sample includes two sets of columns:
  • Dataset fields — Raw columns inherited from the dataset schema. These are always present, even when all values are empty.
  • Rosetta Stone attributes — Normalized attribute properties promoted to top-level columns for convenience. For example, the longitude property of a location attribute (queried as _rosetta_stone.location.longitude in NQL) appears as a top-level longitude column in the sample viewer.
Columns identified as Rosetta Stone attributes are removed when all values are null. System properties (columns starting with _) are also excluded from the sample view.This release note focuses on the UI experience. For programmatic access, see the Sample Data API reference.

Data Plane Details Page

You can now click into any data plane from Settings > Data Planes to view its details. The details page includes an Overview tab showing configuration like display name, type, status, cloud provider, and region, along with a Compute Pools tab listing all compute pools assigned to that data plane.
Data plane details page showing the Overview tab with configuration fields including display name, type, status, cloud, and region

Workflow Data Passing

Workflows now support structured data passing between tasks. Instead of coordinating tasks purely by name-based dataset references, you can capture task output and inject it into downstream task parameters using a shared workflow context.
  • Task output — Each task produces structured JSON (dataset IDs, snapshot IDs, row counts) accessible to subsequent tasks
  • Export — Use export.as with jq expressions to merge task output into a workflow-wide $context that accumulates as tasks run
  • Variable expressions — Use ${…} syntax in task parameters to inject values from the previous task’s output or $context
do:
  - createView:
      call: CreateMaterializedViewIfNotExists
      with:
        nql: "SELECT user_id, email FROM company_data.users"
        datasetName: active_users
      export:
        as: '$context + { datasetId: .dataset_id }'
  - logResult:
      call: ExecuteDml
      with:
        nql: |
          ${"INSERT INTO company_data.log (dataset_id) VALUES (\($context.datasetId))"}
View documentation →

Enhancements and Bug Fixes

Secret Sharing UI: The Secret Sharing interface has been updated with a refreshed design that matches the look and feel of the rest of the platform. View documentation →
Week of February 17, 2026
PlatformRosetta StoneImprovement

Sample Data

Dataset sampling now includes Rosetta Stone normalized attributes alongside your raw data columns, giving you full visibility into how your data transforms through Rosetta Stone mappings directly in the sample viewer.
  • Hover over any dataset field with computed statistics to see detailed information including value distributions and null rates
  • View Rosetta Stone attribute values alongside your source columns in dataset samples
  • Dataset sampling on the Narrative AWS US-East-1 Shared data plane now runs as a background job, matching the Snowflake data plane experience—samples can be requested and gathered for any dataset
  • View documentation →
A dedicated action button to trigger sampling on AWS data planes from the UI is coming soon. Samples can currently be requested via the API and SDK.
Week of February 9, 2026
PlatformNew FeatureNQL

Context Selector

Set your execution context across four dimensions with the new context selector in the top navigation. This segmented pill button replaces the previous “Data Plane” dropdown and gives you control over where your queries execute and what data is visible.
  • Switch between data planes to change which objects are visible throughout the platform
  • Choose compute pools to control whether queries use dedicated, shared, or default resources
  • Set database and schema scope for your current session
  • Context is saved per-browser and persists across sessions
  • View documentation →

Enhancements and Bug Fixes

OBJECT_REMOVE_NULLS Cross-Platform Support: The OBJECT_REMOVE_NULLS function is now accepted on AWS hosted data planes as a no-op, so NQL queries that use it on Snowflake don’t need to be rewritten for AWS.
Week of February 2, 2026
WorkflowsNQLNew FeatureBug Fix

Workflow Orchestration

Automate multi-step data pipelines that execute in a defined sequence based on task dependencies. Define your workflow once using the Serverless Workflow DSL (YAML format) and let the system handle orchestration—no manual triggering required.
  • Create materialized views, refresh them, and run DML operations in automated sequences
  • Schedule workflows with cron expressions for recurring execution
  • Sequential task execution with fail-fast error handling
  • Three built-in tasks: CreateMaterializedViewIfNotExists, RefreshMaterializedView, ExecuteDml
  • View documentation →
Workflow Orchestration is currently API-only. A UI for creating and managing workflows is coming soon.

FORMAT_TIMESTAMP Function

Format timestamps into custom string representations with the new FORMAT_TIMESTAMP NQL function. Convert timestamp columns to date-only strings, time-only strings, or ISO 8601 formats with or without timezone information.
  • Supports standard date/time format patterns (YYYY, MM, DD, HH24, MI, SS, TZH, TZM)
  • Works across both Snowflake and AWS hosted data planes
  • View documentation →

Enhancements and Bug Fixes

File Uploads with Special Characters: You can now upload files with special characters in their names—such as square brackets, spaces, and other non-alphanumeric characters—during dataset creation. Previously, these file names would cause the upload to fail, requiring you to rename files before uploading.
Week of January 19, 2026
Rosetta StoneAIImprovement

Rosetta Stone Normalized Datasets Page

Evaluate and improve your data normalization with full visibility into Rosetta Stone’s AI-powered mapping confidence scores. The new Normalized Datasets view (Rosetta Stone → Normalized Datasets) surfaces how well your datasets map to Rosetta Stone attributes, helping you quickly identify which mappings need attention.
  • View all normalized datasets with color-coded confidence scores (high/medium/low)
  • See mapping type indicators showing AI-generated vs. manual mappings
  • Sort by columns, rows, confidence, or creation date
  • Filter by data plane to focus on specific environments
  • Search datasets by name

Enhancements and Bug Fixes

AI-Generated Dataset Descriptions: Click the Rosetta icon next to any dataset description to automatically generate a rich, contextual description. Rosetta AI analyzes your dataset’s schema and sample data to create descriptions that explain what the data is and where it came from—saving you from writing documentation from scratch.Simplified Refresh Schedule Editor: We removed the confusing “Advanced” mode from the refresh schedule editor based on user feedback. Need a custom cron schedule? Just describe it in plain language and Rosetta AI generates the correct cron expression for you.
Week of January 2, 2026
IntegrationsNew Feature

TikTok Audience Connector

Deliver first-party audiences directly to TikTok Ad Accounts within your Business Center through Narrative’s new TikTok Connector.
  • Secure OAuth 2.0 authentication with TikTok Business Center
  • Support for SHA256 hashed emails, phone numbers, and mobile advertising IDs (IDFA/GAID)
  • Create new audiences or update existing TikTok Custom Audiences
  • Distribute audiences across multiple Ad Accounts
  • Available in Audience Studio and My Data → Datasets → Connections tab
Week of December 15, 2025
Audience StudioNew Feature

Audience Studio

Build audiences with immediate feedback on segment size as you work, eliminating guesswork and speeding up audience creation. This streamlined, purpose-built workflow for marketers and ad operations specialists gives you complete visibility into your audience composition in real time. From audience construction through activation platform delivery, Audience Studio makes your workflow a breeze.
Week of December 9, 2025
EnterpriseSecurityNew Feature

Enterprise Single Sign-On (SSO)

Enterprise customers can now access Narrative using their organization’s existing identity provider. SSO integration allows your team to authenticate through your workforce identity system—including Okta, Azure AD, and OneLogin—with support for enforcing SSO-only authentication and just-in-time user provisioning. Centralize access management, strengthen security compliance, and give your team a seamless login experience.
Week of December 2, 2025
Audience StudioNew Feature

My Audiences

My Audiences gives you a central place to organize, find, and manage all your previously built audiences in one view. Quickly filter audiences by relevance to your business, see where each audience is already connected, and understand audience size at a glance—making it easy to reuse and activate your best segments across platforms.