This QuickStart lists all the new and public beta features released, as well as bugs fixed in September 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.

Subscribe to What's New in Sigma

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

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

Additional account languages available (GA)

Sixteen new languages are now available for the organization account language setting, including additional Chinese, Dutch, French, German, Italian, Japanese, Korean, Polish, Portuguese, Russian, Spanish, Swedish, Thai, and Turkish variants.

WHY IT MATTERS:
Global rollouts often stall on a single unsupported language for one region's team. Broader language coverage removes that as a blocker for multinational deployments, letting more of an organization work in Sigma in the language they're most comfortable with.

For more information, see Set organization language and formatting region

Audit log events for AI conversations (GA)

The Sigma Audit Logs connection now includes an AI_CONVERSATIONS event category that records events related to AI interactions when chat history is enabled.

For more information, see Audit log events and metadata

Audit log events for AI settings (GA)

The Sigma Audit Logs connection now includes an AI_SETTINGS event category that records events related to AI provider and source configuration.

For more information, see Audit log events and metadata

Connect multiple Sigma organizations to one Slack workspace (GA)

Multiple Sigma organizations can now connect to the same Slack workspace.

For more information, see Manage Slack integration

Query variables support for Cortex Agents (Beta)

Row-level security from row access policies is now enforced when using Snowflake Cortex Agents, via immutable session attributes on configured query variables.

For more information, see Specify query variables for a Snowflake connection (Beta)

Restrict Can contribute access to specific version tags (GA)

Users granted Can contribute access to a folder can now be limited to specific version tags instead of all versions and documents.

For more information, see Restrict access to a folder using a version tag

Select Microsoft integration permission levels (Beta)

Administrators can now choose between write-only and read/write access when setting up a Microsoft integration.

For more information, see Manage Microsoft integration

View the code representation of a document (Beta)

The code representation of a workbook, data model, or report can now be viewed directly in the Sigma UI from the document menu's File > View code... option.

There is a QuickStart: Manage Sigma Workbooks as Code with Git and CI/CD

For more information, see Manage data models as code, Manage workbooks as code (Beta), and Manage reports as code (Beta)

Assistant in build mode: cohort retention tables (GA)

Sigma Assistant in build mode can now generate styled cohort retention pivot tables, with automatic heatmap styling to help identify retention patterns.

For more information, see Use Sigma Assistant to build dashboards and apps

Assistant in build mode: pivot table styling (GA)

Sigma Assistant in build mode now applies automatic heatmap and matrix styling to pivot tables based on the type of measure.

For more information, see Use Sigma Assistant to build dashboards and apps

Assistant in build mode: reuse data model metrics (GA)

Sigma Assistant in build mode now preserves data model sources and reuses their governed metrics instead of recreating underlying tables and calculations, so KPIs and charts it builds use existing metric definitions.

For more information, see Use Sigma Assistant to build dashboards and apps

Build workbooks and analyze data in ChatGPT using the Sigma plugin (GA)

The Sigma plugin for ChatGPT can now create and share Sigma workbooks directly from a natural-language analysis, in addition to searching, exploring, and analyzing data.

WHY IT MATTERS:
July's launch of the ChatGPT plugin let users query Sigma data from a conversation. This extends that same conversation into a governed, shareable workbook — the analysis doesn't have to stay locked in a chat transcript to be useful to the rest of a team.

For more information, see Use the Sigma plugin for AI assistants

Call Sigma agents from the Sigma MCP server (GA)

Two new tools available exclusively in the Sigma MCP server let users list the agents they can access and call an agent from any connected AI tool.

There is a QuickStart: Agents 06: Connect a Sigma Agent to GitHub with MCP Tools

For more information, see Use the Sigma MCP server

Chat history for Assistant and agents (GA)

Chat history is now available for conversations with Sigma Assistant, warehouse agents, and Sigma agents, letting users revisit previous conversations. Admins can configure storage and retention for organizations that meet certain conditions, including a choice between Sigma-owned bucket storage or a customer-owned bucket via external storage integrations.

WHY IT MATTERS:
Without persistence, every AI conversation starts from zero, limiting it to a single-turn interaction. Chat history now covers Assistant, warehouse agents, and Sigma agents alike, carrying context across sessions and giving compliance-minded organizations one consistent retention story instead of three separate ones — including the option to keep that data in storage they control.

There is a QuickStart: Agents 03: Giving Agents Memory

For more information, see Configure chat history and Chat with Sigma agents

Improved Sigma Assistant (GA)

Sigma Assistant has improved reasoning, faster performance, and higher-quality responses, including clarifying questions, more flexible data source access, visible reasoning, and the ability to export a conversation.

WHY IT MATTERS:
Conversation export and visible reasoning turn Assistant's output into something a reader can verify and hand off, rather than a black-box answer you either trust or don't. Paired with faster, higher-quality responses, this closes the gap between "an AI answer" and something you'd actually put in front of a stakeholder.

For more information, see Ask natural language queries with Sigma Assistant

MCP connectors for Sigma agents (GA)

Admins can add MCP servers as connectors, enabling Sigma agents to retrieve context, fetch data, and perform actions in third-party tools.

WHY IT MATTERS:
An agent is only as useful as what it can reach. MCP connectors let an agent pull in and act on whatever third-party tools an organization already runs, governed the same way as every other connector in Sigma, instead of being limited to data already sitting in the warehouse.

For more information, see Configure MCP connectors

Migrate to Sigma with an AI assistant (Beta)

Migration skills for Sigma let an AI assistant rebuild source content — the dashboards, reports, and data models in another BI tool — as a Sigma document.

WHY IT MATTERS:
This is the same AI-driven migration pattern behind the Migration QuickStarts family — Sigma now documents and supports it directly instead of leaving it as an unofficial pattern. It gives prospects and partners a sanctioned starting point for moving dashboards, reports, and data models off another BI tool without a full manual rebuild.

There are vendor-specific QuickStarts: Migration category

For more information, see Migrate to Sigma with an AI assistant

New models used for AI providers (GA)

LLM model updates across providers: OpenAI now uses GPT 5.6, Databricks uses Claude Sonnet 5, and Anthropic uses Claude Sonnet 5.

For more information, see Supported AI models

New skills for creating workbooks and reports with an AI assistant (GA)

The sigma-workbooks and sigma-reports agent skills give AI assistants reference materials and instructions for authoring the code representation of a workbook or report.

For more information, see Install skills for AI assistants

Sigma agents (GA)

Sigma agents are now generally available. Build an agent using instructions, data sources, actions, warehouse agents, search services, and MCP connectors, then chat with it, schedule automated runs, or call it via the REST API or MCP server. Chat history resumption, task approval, and embedding are all supported, and admins can review token usage, owners, data sources, access grants, feedback, and execution logs.

WHY IT MATTERS:
This is the full agent platform landing at once — build, govern, call, and embed, with the same admin oversight (token usage, access grants, execution logs) expected of anything else running in Sigma. Agents aren't a bolted-on chat feature anymore; they're an audited, first-class part of the platform, reachable from a workbook, the REST API, or an MCP server.

There are eight QuickStarts on Agents: Agent category

For more information, see Build agents, Chat with agents, Create actions that interact with agents, Call agents with the API, Embed an agent chat interface, Example use cases, Manage agents, Build evaluation suite, and About Sigma agents

Sigma Assistant in the workbook (GA)

Sigma Assistant in the workbook is now generally available. Natural-language prompts support data exploration, insight analysis, and dashboard or app creation, running on a code-first architecture with broader element support (maps, combo charts, navigation, drawers, single row containers), in-chat chart formatting, and improved design defaults.

WHY IT MATTERS:
A code-first architecture means Assistant's output in the workbook is something you can read and understand, not just trust blindly. Combined with broader element coverage and in-chat chart formatting, Assistant now builds and edits more of what a workbook actually needs, instead of a narrow slice of chart types.

For more information, see Use Sigma Assistant to explore and analyze workbook data and Use Sigma Assistant to build dashboards and apps

Use warehouse agents with Assistant and agents (GA)

Sigma Assistant and Sigma agents can now use Snowflake Cortex Agents or Databricks Genie Agents as tools, bringing warehouse-native agent capabilities directly into Sigma.

WHY IT MATTERS:
This isn't Sigma building a competing warehouse agent — it's Sigma's AI layer calling out to whatever agent ecosystem you've already invested in on Snowflake or Databricks, and folding the result back into a governed Sigma conversation. You get one consistent interface over both Sigma-native and warehouse-native AI, instead of switching tools depending on which agent has the answer.

For more information, see Use warehouse agents with Sigma

Databricks connection endpoint additions (GA)

Databricks connection endpoints now include enableHiveMetastore, usePython, and pythonComputeClusterId fields for programmatic configuration.

For more information, see Get connection details, List connections, Create a connection, and Update a connection

Impersonate a user in API calls (GA)

A user assigned the Admin account type can impersonate other users in API calls — for example, to retrieve and display the contents of their Documents folder in embedded content.

WHY IT MATTERS:
Embedded experiences often need to act on behalf of the end user rather than a shared service account. Admin-level impersonation makes that possible without provisioning and managing a credential for every embedded user.

For more information, see Impersonate users

Manage workbooks from a code representation (Beta)

The Sigma API can now retrieve, update, and create workbooks based on a JSON or YAML representation.

For more information, see Manage workbooks as code (Beta) and Workbook representation example library

New account type permissions update endpoint (GA)

Account type permissions can now be updated via a new endpoint to control feature access, with the response including license type.

For more information, see Update account type permissions

New AI provider configuration endpoint (GA)

A new endpoint configures an organization's AI provider programmatically.

For more information, see Configure the organization's AI provider

New API credentials listing endpoint (GA)

A new endpoint lists API credentials, returning client ID, owner ID, and scopes, with the client secret excluded from the response.

For more information, see List API credentials

New API endpoints for managing AI chat history settings (Beta)

Two new endpoints get and update an organization's AI chat history settings.

For more information, see Get the chat history settings and Update the chat history settings

New API endpoints to manage audit logging for an organization (GA)

Two new endpoints support getting and updating an organization's audit logging setting.

For more information, see Get the audit logging setting for an organization and Enable or disable audit logging for an organization

New API endpoints to manage CSV upload settings (GA)

Two new endpoints support getting and updating an organization's CSV upload settings.

For more information, see Get the CSV upload settings and Update the CSV upload settings

New API endpoints to manage email branding (GA)

Three new endpoints get, update, and reset an organization's email branding settings.

For more information, see Get the email branding settings for an organization, Update the email branding settings for an organization, and Reset the email branding settings for an organization

New comment settings management endpoints (GA)

Two new endpoints get and update an organization's comment settings, controlling commenting and image annotation functionality.

For more information, see Get comment settings and Update comment settings

New option for the List datasets endpoint (GA)

The List datasets endpoint can now filter results to datasets owned by a specific user using ownerId.

For more information, see List datasets

New options for some workbook endpoints (GA)

The tags array returned by workbook endpoints now includes an isArchived option to indicate inactive workbook tags.

For more information, see List workbooks, Get a workbook, and Get tags for a workbook

New options for the Create a deployment policy endpoint (GA)

A new useDependenciesWorkspace option specifies a separate workspace for deploying dependent documents.

For more information, see Create a deployment policy and How dependencies are deployed

New options for the List workbooks endpoint (GA)

A new includeTaggedSourceUrlId query parameter identifies source documents for deployed version-tagged workbooks.

For more information, see List workbooks

New sample connection management endpoints (GA)

Two new endpoints get and update an organization's sample connection settings.

For more information, see Get sample connection settings and Update sample connection settings

New team admins management endpoint (GA)

A new endpoint updates the administrators assigned to a team.

For more information, see Update team admins

New timezone management endpoints (GA)

Two new endpoints get and update an organization's timezone configuration.

For more information, see Get timezone setting and Update timezone setting

New user attribute update endpoints (GA)

A user attribute's name, description, or default value can now be updated via a new endpoint.

For more information, see Update a user attribute

Represent CSV tables in the code representation of a data model (GA)

CSV source tables can now be represented when reading and updating a data model's code representation.

For more information, see Manage data models as code and Example: representing a data model with a CSV table

Request and grant access to API connectors (GA)

Users can now view all API connectors in their organization and request access to specific ones, with admins or users with appropriate permissions able to approve or deny requests.

For more information, see Manage API credential and connection access

Sending and scheduling exports using the API on behalf of another user (Deprecated)

The sendAsUser option on the send-export endpoint and the ownerId option for scheduling are deprecated in favor of token-based impersonation.

For more information, see Impersonate users to send and schedule exports

Tag endpoint enhancements (GA)

The Update a tag endpoint now supports updating version tag descriptions and colors.

For more information, see Update a tag

1: The Update a data model from a code representation endpoint now validates that the documentVersion field matches the most recent version before updating.

2: Users with the Manage users permission can now invite guest users.

3: Document deployment now respects existing team and user attributes for column-level security enforcement across tenant organizations.

4: Failed materializations blocked by in-progress operations now display as skipped rather than failed.

5: Improved performance when editing large input tables.

6: MCP tools used by Sigma agents have been renamed MCP connectors.

7: Templates shared with an entire organization using Share with everyone in your organization now appear in the Shared with you template gallery.

8: Users can now swap tables, schemas, and databases/catalogs when tagging a document version without Can use access to the entire connection.

9: Fixed an error that prevented setting up Azure OpenAI as an AI provider with certain temperature values.

10: Version-tagging or swapping sources on workbooks with broken connections now returns clear error messaging instead of a generic failure.

11: Fixed the List workbooks for a tag endpoint returning no results for non-admin users with access only to tagged versions.

12: Fixed a 400 error occurring in conversations with Claude Sonnet 5 as the reasoning model after a dozen or more messages.

13: Fixed agent calls failing during automated actions when using OpenAI GPT-5.4 or GPT-5.1 models.

14: Any user with edit access to a dataset can now migrate it to a data model.

15: Resolved an error preventing Properties tab selection after switching editor panel tabs.

16: Corrected OpenAI external provider configuration to expect GPT 5.4 instead of GPT 4.0.

17: Agent automated actions now retry on incomplete or incorrect responses and preserve null value types for downstream action compatibility.

18: Fixed Azure OpenAI agent and Assistant chats terminating prematurely with response cutoff errors.

Choose related datasets when migrating a dataset to a data model (GA)

Migrating a dataset to a data model now lets you select which related datasets to combine, rather than automatically pulling in every linked, joined, or referenced dataset.

For more information, see Migrate a dataset to a data model

Revert migrated dataset references (GA)

After migrating datasets to data models, references can now be reverted to use the original datasets again, for troubleshooting or testing an alternative migration approach.

For more information, see Migrate a dataset to a data model

Embed Sigma Assistant (Deprecated)

Embedding Sigma Assistant as a standalone experience is deprecated and will reach end of support on March 16, 2027. Consider embedding a customized Sigma agent, or Sigma Assistant in the workbook, instead.

There is a QuickStart: Embedding 08: Embedding Sigma Assistant

For more information, see Embed Sigma Assistant

Also see: REST API Usage 12: Call Sigma Agents from Your Application

Match on null values when performing a lookup (GA)

Lookups now support a Match null values option so rows with a null join key on both sides can match, using the new LookupMatchNulls function.

For more information, see LookupMatchNulls and Add columns through lookup

TextJoin function (GA)

The TextJoin function joins multiple strings of text using a common delimiter.

For more information, see Textjoin

Agents

September also introduces a new Agents category — an 8-part series building up a Sigma agent from a single grounded answer through actions, memory, scheduling, custom tools, and warehouse-native AI.

• Agents 01: Building Your First Sigma Agent
• Agents 02: Agent-Driven Actions & Writeback
• Agents 03: Giving Agents Memory
• Agents 04: Scheduling Unattended Agent Runs
• Agents 05: Build Web Search for a Sigma Agent
• Agents 06: Connect a Sigma Agent to GitHub with MCP Tools
• Agents 07: Running Python from an Agent
• Agents 08: Give Your Sigma Agent a Snowflake Cortex Specialist

Manage Sigma Workbooks as Code with Git and CI/CD

This QuickStart shows how to define a Sigma workbook as a single YAML file and move it through the same git-based review, validation, and CI/CD pipeline as the rest of your application code.

REST API Usage 12: Call Sigma Agents from Your Application

This QuickStart shows how to call a Sigma agent directly over the REST API — from your own chat interface, a scheduled job, or another application — while inheriting the same governance and permissions configured in the workbook.

App templates (GA)

App templates let you start building an app in Sigma from an interactive preview before adding it to your organization as a workbook, with ten templates available including Project Management and Revenue Forecasting.

There are QuickStarts for Templates: Templates category

For more information, see Get started with templates

Apply translations in workbook or report export attachments (Beta)

After adding translations to a workbook or report, one of those translations can now be applied when exporting content.

For more information, see Apply translations in workbook or report export attachments (Beta)

Databricks support for stored procedure actions (GA)

Creating actions that call stored procedures is now supported for Databricks connections.

Date support for slider and range slider controls (GA)

Slider and range slider controls now support both Date and Number value types.

For more information, see Slider, Range slider, and Intro to control elements

Drawers (GA)

Drawers — side panels that slide in to overlay workbook content temporarily — are now generally available.

WHY IT MATTERS:
Builders have long had to choose between cluttering a page with detail or sending readers elsewhere to see it. Drawers give you a place to put that detail — instructions, drill-downs, forms — that slides in on demand and out of the way otherwise, without breaking up the main layout. Now that it's GA, it's ready for production workbooks rather than just evaluation.

For more information, see Use drawers to manage complex workflows

Export reports in PowerPoint format (GA)

Reports can now be exported and downloaded as PowerPoint (.pptx) files.

For more information, see Share and export reports

Export specific report pages (GA)

Users can now select one or more specific report pages to export, instead of exporting the entire report.

For more information, see Share and export reports

Manage reports from a code representation (Beta)

The Sigma API can now retrieve, update, and create reports based on a JSON or YAML representation.

For more information, see Manage reports as code (Beta) and Report representation example library

There is a QuickStart: Manage Sigma Workbooks as Code with Git and CI/CD

Manually trigger an AI column or cell run (Beta)

AI columns can now be configured with a Don't run automatically option, so they only run when manually prompted instead of on every change.

For more information, see Create AI columns (Beta)

Additional Resource Links

Blog
Community
Help Center
QuickStarts

Sigma Free Trial