This QuickStart lists all the new and public beta features released, as well as bugs fixed in July 2026.
It is summary in nature, and you should refer to the specific Sigma documentation links provided for more information.
Public beta features will carry the section text "Beta".
All other features are considered released (GA or generally available).
Sigma actually has feature and bug fix releases weekly, and high-priority bug fixes on demand. We felt it was best to keep these QuickStarts to a summary of the previous month for your convenience.
New first Friday features QuickStarts will be published on the first Friday of each month, and will include information for the previous month.
For those wanting to see what Sigma is doing on each week, release notes are now also available on the Sigma Community site. There, you can opt in to receive notifications about future release notes in order to stay on top of everything new happening at Sigma. You can also subscribe to automated updates in any Slack channel using the Sigma Community release notes RSS feed.
For more information on how to subscribe to release note notifications, see About the release notes
Sigma's help documentation site has been updated with a new homepage, a dedicated function reference section, and an enhanced API explorer with code representation views. The site also supports markdown export for any page.
For more information on Sigma's product release strategy, see Sigma product releases
If something is not working as you expect, here's how to contact Sigma support

Sigma now supports the Azure Australia region, with new deployments hosted in australiaeast and disaster recovery in australiasoutheast, reducing latency for customers in Australia.
For more information, see Supported regions, data platforms, and features
Email branding settings and SMTP configuration have moved to the Email customization tab under Administration > Scheduled exports & actions. Functionality is unchanged.
For more information, see Customize email branding
Users can now sign out of active connection sessions from their profile, giving individuals direct control over live connection credentials without administrator intervention.
For more information, see Delete a current connection session

Deploy a folder and all its contents — workbooks, reports, and data models — to tenant organizations in a single operation.
WHY IT MATTERS:
For teams managing content across multiple tenant organizations, folder-level deployment replaces a series of individual document deployments with one governed action — reducing the overhead of keeping tenant environments in sync as content evolves.
For more information, see Deploy content to tenant organizations
Reports can now be included in deployment policies, allowing them to be pushed to tenant organizations alongside workbooks and data models.
For more information, see Deploy content to tenant organizations
Embed users can now export directly to organization-configured Slack and Microsoft Teams integrations, without requiring access to a full Sigma account.
For more information, see Manage embed user settings (Beta)
A new setting lets administrators obscure which user scheduled or sent an export in the email body, giving organizations more control over how automated emails are presented to recipients.
For more information, see Customize your organization's email branding
The Locale section in account settings has been renamed to Localization, and Account locale has been renamed to Account language. Functionality is unchanged.
For more information, see Manage workbook localization
Parent organization administrators can now create and assign custom attributes for tenant organizations directly from the Administration portal, without switching into each tenant.
For more information, see Create and manage tenant organizations
Documents within an existing deployment policy can now be manually redeployed to tenant organizations without recreating the full policy.
For more information, see Create and manage tenant organizations

Sigma Tenants is now generally available, allowing organizations to set up a multitenant architecture with multiple Sigma organizations under a single account. Designed for enterprises and ISVs managing separate business units, regions, or customer deployments, Sigma Tenants is a premium feature.
WHY IT MATTERS:
Multitenancy has historically required separate Sigma instances or complex shared-infrastructure workarounds. Sigma Tenants gives enterprises and ISVs a governed, first-class path to isolate organizations, manage content deployment centrally, and audit activity per tenant — without the overhead of maintaining multiple accounts.
For more information, see Multitenancy at Sigma
Administrators can now export tenant audit logs to cloud storage, enabling centralized log management across tenant organizations.
For more information, see Configure tenant audit log storage integrations
A new TENANTS event category is now available in the Audit Logs connection, tracking deployment management, policy changes, and capability grants across tenant organizations.
For more information, see Audit log events and metadata
Recent queries can now be cached in external storage, reducing warehouse compute consumption and improving response times for repeated or similar queries.
For more information, see Manage universal result cache (Beta)
The universal result cache now supports connections using OAuth authentication, extending query caching to all warehouse authentication types.
Deployment policies now display which dependent documents will be deployed alongside the primary content, giving administrators visibility into the full deployment scope before pushing changes.
For more information, see Example: Software development lifecycle


Three new workbook templates — Claude, OpenAI, and Snowflake — visualize the costs associated with AI tool usage across your organization. Use them as a starting point for internal AI cost tracking and governance reporting:

WHY IT MATTERS:
AI spend is increasingly showing up in IT and finance conversations — and right now most teams have no easy way to see where it's going. These templates give you a working cost dashboard on day one, built on live data, without starting from scratch.

A new dashboard tracks token consumption, conversations, engagement, and model usage across your organization. Includes an AI analyst agent for querying usage data in natural language:

WHY IT MATTERS:
As AI usage scales across teams, cost visibility becomes a governance requirement. The AI usage dashboard gives administrators a live view of consumption by model, user, and feature — the audit trail enterprises need before expanding AI access broadly.
For more information, see AI usage dashboard
Amazon Bedrock is now a supported AI provider, allowing Sigma's AI-powered features to run on Anthropic foundation models through your existing Bedrock configuration.
WHY IT MATTERS:
Enterprises with AWS infrastructure can now route Sigma's AI requests through their own Bedrock account — keeping AI workloads inside their AWS boundary and satisfying data residency or compliance requirements.
For more information, see Add Amazon Bedrock as an AI provider
Add an Anthropic API key to connect Sigma's AI features directly to Anthropic's models, without routing through Amazon Bedrock.
For more information, see Manage external AI integrations
AI columns on Databricks connections now support configurable response formats and include usage visibility for monitoring token consumption.
For more information, see Create AI columns (Beta)
Tag Sigma MCP server queries with "kind":"mcp" to track costs and usage in the AI usage dashboard, giving administrators visibility into LLM consumption from MCP-based workflows.
For more information, see Use the Sigma MCP server
The Users and Document Activity dashboards now support AI agent queries, letting administrators ask questions about usage patterns in plain language. Requires a configured AI provider:

A one-click installation connects Sigma's MCP server to ChatGPT, letting users query Sigma workbooks, data models, and live data directly from their ChatGPT workflow.
WHY IT MATTERS:
Analysts who live in ChatGPT can now access live Sigma data without switching context — with access governed by Sigma's existing permissions and audit logging intact.
For more information, see Set up the Sigma MCP server in ChatGPT

Import an API connector configuration directly from a cURL command, reducing manual setup when converting existing API calls into Sigma connectors.
For more information, see Create an API connector based on a cURL request
Connections can now be configured with independent OAuth settings — client ID, client secret, and scopes — per connection, enabling different OAuth configurations across environments.
For more information, see Manage connections
List lineage endpoints now include uploaded CSV files as data sources. CSV entries display with a csv-upload type and include a csvId field.
For more information, see List lineage tree and List data model lineage tree
A new POST /v2/datasets/:datasetId/migrate endpoint supports programmatic migration of datasets to data models, including a dry-run option to validate the migration before committing.
For more information, see Migrate a dataset to a data model
Connection API responses now include write destination information for OAuth connections, giving programmatic workflows access to writeback configuration details.
For more information, see List connections
The connection creation and update endpoints now support additional options including queue size, writeback descriptions, dynamic table configuration, and Python warehouse settings.
For more information, see Create a connection
The workbook and report send/schedule endpoints now support Cc and Bcc recipients, adding flexibility to automated report distribution.
For more information, see Send workbook and Send report
The data export endpoint now accepts a pageId parameter, enabling programmatic export of specific report pages.
For more information, see Export data
The connection paths endpoint now supports filtering by connectionId, making it easier to scope path queries to a specific connection.
For more information, see List connection paths
Nineteen new report endpoints add programmatic control over version history, duplication, source swapping, SQL queries, controls, lineage, columns, grants, pages, and tags.
For more information, see the API reference
A new POST /v2/members/:memberId/revoke endpoint allows administrators to revoke a member's OIDC and warehouse connection tokens without interrupting active sessions.
For more information, see Revoke member tokens
Snowflake semantic views now appear in lineage endpoint responses, enabling complete lineage tracing for workbooks sourcing data from semantic layer definitions.
For more information, see List lineage tree
The connection creation endpoint now supports Starburst and Azure SQL DB as connection types.
For more information, see Create a connection
A new POST /v2/workbooks/{workbookId}/convertToReport endpoint enables programmatic conversion of workbooks to reports.
For more information, see Convert a workbook to a report
A new POST /v2/workbooks/:workbookId/restoreVersion endpoint enables restoring a workbook to a previous version programmatically.
For more information, see Restore workbook version

Data model tables can now be used as sources for Sigma agents, giving builders access to governed, semantic-layer data when configuring agent capabilities:

For more information, see Build Sigma agents
Sigma Assistant in build mode now supports guided balance sheet creation, with adaptive data structure support and snapshot and comparative layouts for financial analysis.
Sigma Assistant in build mode can now join tables on formulas and expressions, not just raw columns, enabling more flexible data assembly during workbook construction.
Sigma Assistant in build mode can now offer users predefined response options, guiding them through structured workflows and capturing consistent input.
Sigma Assistant in build mode now detects when a user may be dissatisfied and proactively suggests sending feedback. Users can review, edit, or dismiss the proposal before it's sent.
Builders can now set a fixed greeting message for agents rather than relying on an AI-generated response. Configure agents to open every conversation with a specific, controlled message — such as "Welcome to Sigma!"

For more information, see Build agents
API connectors now support GraphQL endpoints, allowing agents and actions to execute queries and mutations against GraphQL APIs in addition to REST.
A new action type allows agents and actions to insert one or more rows into an input table. Rows can contain new defined values or data pulled from existing workbook sources.
For more information, see Create actions that modify input table data
API connector call actions can now insert multiple rows by parsing array structures from API responses, turning list and paginated responses into table data.
Sigma Assistant now supports plan and build modes in the workbook, enabling natural language construction of dashboards, charts, tables, KPIs, forms, and AI-powered apps.
WHY IT MATTERS:
Plan and build modes shift Sigma Assistant from answering questions to constructing workbooks. Describe what you want to build, review the plan before a single element is created, and end up with a governed, publishable workbook — live on real data, with version history and permissions intact from the start.

For more information, see Build Dashboards and Apps with Sigma Assistant

A growing library of pre-built app templates is now available, covering common business scenarios including Project Management, Revenue Forecasting, Demand Planning, and seven additional use cases. Use templates as a starting point and customize them to your data and workflows:


1: Restored documentation MCP server functionality.
2: Fixed inactive version tag timestamps to show correct dates.
3: Preserved element layout positions when copying/pasting multiple elements.
4: Resolved Snowflake Cortex Agent warehouse selection issues.
5: Improved tooltip theme setting compliance.
6: Preserved code representation id values in data model creation endpoint.
7: Linked input table creation now validates write-back schema compatibility before proceeding.
8: Improved document deployment with custom page visibility for tenant organizations.
9: Resolved deployment of tagged documents in policy folders.
10: Assistant now preserves Top-N SQL results in charts.
11: API credentials dropdown is now searchable.
12: Source swap policy lists now display completely.
13: Conditional formatting can now remove bold from totals and subtotals.
14: Sigma agents now support Snowflake Cortex Agent names that contain spaces.
15: Fixed an error when granting stored procedure access.
16: MCP server details have moved from the Profile screen to the Integrations screen.
17: AI usage dashboard no longer fails when archival errors are present.
18: Custom SQL elements now deploy successfully with source swap policies and specific connection paths.
19: Improved KPI chart loading performance on initial workbook open.
20: User name and email are now shared with warehouse agents when using Sigma Assistant.

Databricks Unity Catalog metric views are now browsable in Sigma's data catalog. Use metric views as sources for tables, pivots, and charts, and write custom SQL queries against them.
For more information, see Sigma integration with Databricks Unity Catalog and Unity Catalog metric views


The new BusinessDays function calculates the number of weekdays between two dates, including both the start and end dates and excluding weekends.
WHY IT MATTERS:
Date calculations in business contexts — SLA tracking, payment terms, project timelines — routinely need to exclude weekends. A dedicated function makes these calculations accurate and readable without workarounds.

For more information, see BusinessDays

This QuickStart shows how to install and use the Sigma CLI (sigcli), a typed command-line wrapper over Sigma's REST API, to turn repetitive administrative work into repeatable, auditable automation.
It walks through how to:
sigcli with a named authentication profilejqWHY IT MATTERS:
Anything you can do in the Sigma REST API, you can now do from a script — turning one-off administrative clicks into repeatable automation for inventory, provisioning, and deployment. Because every command runs through the same API, permissions, and audit logging as the rest of Sigma, that automation stays governed.
This QuickStart shows how to use Sigma Assistant's Plan and Build modes to design and construct dashboards and AI apps from scratch, then refine them conversationally — all live on your data.
It walks through how to:
Plan and Build modes, and switch between themWHY IT MATTERS:
This moves Assistant from analysis to construction — describe what you want in plain language and get a governed, publishable workbook rather than a throwaway mockup. Builders design and iterate faster while the result stays live on real data, with Sigma's permissions and version history intact.
Sigma Starter Apps are fully built, production-ready applications covering common business workflows across finance, sales, operations, and support. Each app is built entirely on Sigma's native capabilities — input tables, AI agents, actions, and live warehouse data — and comes with a QuickStart that walks through the design patterns so you can adapt it to your own data.

Pivot tables now support one-off calculated columns for analysis without modifying the underlying dataset.
The new Unnest feature creates table rows from array data, expanding nested arrays into a flat, analyzable table structure.
For more information, see Create a table from an array
Button and navigation elements now support icons alongside labels, or as a standalone replacement for text, giving interactive controls and navigation a more visual and compact appearance:
For more information, see Button elements
Users can now request cancellation of in-progress element materializations, giving more control over long-running operations without waiting for them to complete.
A new data loading configuration option lets builders manage prefetch queries for specific elements, optimizing performance for pages with multiple or complex data sources.

Existing workbooks can now be converted to reports, enabling pixel-level formatting control and more reliable export behavior:

WHY IT MATTERS:
Builders who already have a workbook no longer need to rebuild it as a separate report asset to get pixel-perfect formatting and reliable exports. Convert in place and keep everything you've already built.
For more information, see Convert workbooks to reports
Excel and Google Sheets exports now preserve cell background color, text color, row banding, and font properties from the source workbook.
WHY IT MATTERS:
Export fidelity has long been a gap between what analysts see in Sigma and what stakeholders receive. Preserving formatting in Excel and Sheets exports means reports land polished and ready to share — without manual cleanup.

Find in table is now generally available, letting users search for specific values within tables and input tables directly in the workbook.
WHY IT MATTERS:
For analysts working with large datasets, locating a specific row or value without sorting or filtering reduces friction — especially in input tables where pinpointing a record before editing is a common step.
Append :lng= to a workbook URL to apply locale-specific number and date formatting without changing account-level localization settings.
Text elements now support mailto: links, allowing builders to add clickable email links directly in workbook text.
For more information, see Text elements
The Manage locales panel has been renamed to Manage translations. Functionality is unchanged:

The Sigma homepage now includes Popular activity and Your Work sections, surfacing frequently accessed and recently visited content for faster navigation.
Segmented controls with long display values now automatically convert to dropdown menus, preventing text truncation in compact layouts.
A new treemap element displays values as proportional nested boxes grouped by category, providing an alternative visualization for hierarchical or part-to-whole data.
For example:

For more information, see Build a treemap (Beta)

Additional Resource Links
Blog
Community
Help Center
QuickStarts
