Coming soon

The following features are coming in the next release of ThoughtSpot Cloud:

April 2026 26.3.0.cl

ThoughtSpot can now automatically create Liveboards based on your prompts, populating them with relevant visualizations and KPIs. In response to your prompt, ThoughtSpot evaluates available data sources, creates visualizations, and lays out the information for you. After creating a Liveboard, you can enter more prompts in SpotterViz at any time to revise and refine it.

We now support locale-specific translations for all written text on saved Answers and Liveboards, including downloaded and scheduled PDFs, and downloaded PNGs of visualizations. Translations are matched with the locale chosen in the user profile. Note that admins must enable AI auto-translation for your instance.

For more information, see Translations of Answers and Liveboards.

Variables are now generally available (GA). ThoughtSpot introduces variables to allow you to substitute values for specific properties of an object and enable dynamic data propagation across Orgs in ThoughtSpot. You can create variables of the following types: table mapping, connection property, and formula variable. In addition, you can delete, search, edit, and rename the variables. Note that you can create variables in ThoughtSpot, or by using APIs. If you are using APIs, you can also create connection property per principle variables. For more information, see Custom variables.

You can now use SpotterModel to generate comprehensive, optimized Models ready to use in your analysis. SpotterModel automates table selection, join creation, and column selection, while allowing you to review and customize each step.

Formula variables are now generally available (GA). Formula variables can be of any data type that ThoughtSpot supports, and can be associated with a single value or list of values. Formula variables created in the primary Org can be shared with other Orgs. A formula variable created in a non-primary Org is only visible in that Org. For more information, see Formula variables.

ThoughtSpot now allows you to choose whether you want to keep parts of a Model in sync with the underlying tables in the cloud data warehouse (CDW). Previously, as tables from the CDW were updated with new information from the CDW, it would cause the link for some column properties from the table to the Model to break. If there were multiple Models on the same table, for changes like renaming a column or updating a description, you would have to go into each Model and update the values individually. To enable this feature, contact ThoughtSpot support.

ThoughtSpot introduces Sets at the Answer level to answer query-on-query questions by generating and consuming Query Sets within a single Answer flow. Users with edit permissions on an Answer can view and modify the associated Sets.

To enable this feature, contact your administrator.

In this release, we add support for non-equi joins when modeling in the ThoughtSpot UI which reduces modeling time. Up to now, only equi (=) joins were supported in the UI, which limited modeling flexibility. Beginning in this release, you can use the following non-equi joins in the UI: <, >, <=, >=, and !=. To enable this feature, contact ThoughtSpot support.

The export of high-resolution PNG images of Liveboards is now generally available (GA). You can now export improved high resolution PNG images of Liveboards. Previously, PNG exports were PDF reports in a PNG format created by taking a rolling screenshot of the PDF report generated in a headless browser.

ThoughtSpot introduces support for exporting pivot tables in XLSX format. This feature ensures that pivot tables retain their full visual formatting and structural integrity when exported as XLSX files from a Liveboard. To enable this feature, contact your administrator. NOTE: This features does not apply to individual Answer downloads.

Customers can now hide the header logo in ThoughtSpot notification emails, eliminating the need to upload a placeholder image.

ThoughtSpot introduces aggregate-aware query execution. This is a transparent query optimization feature where Search Data automatically switches from detailed Models to associated aggregate Models. This provides a mechanism to manage and reduce cloud costs and improve query latency while maintaining insight accuracy.

Spotter asks for permission before using integrated tools like Slack or Confluence in a conversation, specifying exactly which action it will take and the tool it will use. You can choose to allow once, allow in the current conversation, or deny. This allows you full control of Spotter’s actions.

You can now select the Learn from Liveboard option on a data model to incorporate context from Liveboards built on the Model. Spotter fetches the learned context during conversations and produces responses that reflect this context.

You can now disable Connections directly within ThoughtSpot, suspending all activity on that Connection until it’s re-enabled.

ThoughtSpot now checks the health of each Connection when your environment upgrades.

Aliasing dimension and fact tables is now generally available (GA). Previously, table aliases were only supported for role-playing dimensions, or when a single physical dimension is referenced multiple times in a fact table, with each reference linking to a logically distinct role for the dimension. Now, we support table aliases for chasm trap models, and aliasing the primary fact table. Note that data security (row-level and column-level security) applies to the aliased table as well as the original table. We also now support progressive filters for table aliases.

For more information, see Role-playing dimensions.

Linking parameters across Models is now generally available (GA), and enabled by default for all users. Liveboard creators can link parameters of the same type across multiple different Models. This reduces the repetitive effort required for Liveboard users to individually set identical parameters for each Model.

For more information, see Link parameters across Models.

Suggestion settings are now generally available (GA). We have added the Suggestion Settings column to Models. This is used in conjunction with row-level security to ensure that values displayed in sample values and Spotter flows are secured.

Where columns require that strict row-level security is applied to sample values and Spotter, then set the Suggestion Settings column as Secure Suggestions.

For more information, see Suggestion settings.

Admin users can now specify to skip indexing for tables of a certain size. By default, ThoughtSpot does not set a limit on the size of tables it indexes. You can choose to Index all tables, Disable indexing for tables with ≥ 100M rows, Disable indexing for tables with ≥ 250M rows, or Disable indexing for tables it ≥ 1B rows.

You can now adjust multiple Liveboard filters together in a single, centralized filter window. Instead of changing and applying each filter separately and reloading the page, you can combine your changes and apply them together.

ThoughtSpot introduces support for dbt MetricFlow that allows you to import semantic definitions like formulas, measures, dimensions, and metrics from dbt. This eliminates the need to maintain separate models in dbt and ThoughtSpot.

You can now set variables in row-level security as TS_WILDCARD_ALL, which grants access to the data, and TS_WILDCARD_NONE, which denies access. Previously, ThoughtSpot only supported the TS_WILDCARD_ALL designation, and you had to set the variable as a value or list of values to control access.

You can now include the current period within a date filter chip, when selecting a rolling date filter. This allows you to define a date range such as "last two months, including this month." This rolling date filters enhancement applies to both Liveboard and Answer filters.

You can now add objects to Collections directly from your Answers, Liveboards, and Data list pages without navigating to the Collections page.

For new features and enhancements introduced in this release of ThoughtSpot Embedded, see ThoughtSpot Developer Documentation.