Coming soon

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

September 2026 26.9.0.cl

We have redesigned the experience in Spotter 3 to look more streamlined and conversational. No functional changes have been made to how Spotter defines its reasoning, the "thinking" behind the answer. The search bar options have been simplified, but no functionality has been removed.

You can now select multiple Models to search on in a single Spotter conversation. For example, you could select a Revenue Model and a Finance Model and ask a question that spans data from both. To enable this feature, contact ThoughtSpot support.

You can now pin chats to the top of the view in the Spotter side menu, so they’ll be easy to find. Pinned chats are organized by which chat has the most recent activity — most recently interacted with, or most recently pinned.

Spreadsheet puts a familiar, editable grid directly on your live data. Open any ThoughtSpot data model, table, or dataset, and every column and formula you add runs as a live query against the governed model, with no export and no stale copy. Formula columns persist as governed logic, while sorting, filtering, conditional formatting, and text and number formatting let you shape the grid on the fly. Save your work as an Answer that stays searchable, shareable, and governed like anything else in ThoughtSpot.

ThoughtSpot query cancellation for Search data is now generally available (GA). You can now stop in-progress queries on your cloud data warehouse (CDW) when you change or abandon a search. This reduces unnecessary CDW spend and lets you iterate faster.

Previously, queries made using Search data continued running until timeout or completion even after you modified the search, navigated away, or edited an Answer. This release introduces two cancellation mechanisms:

In this release, we improve load times for custom charts and charts in the new charts library. Previously, these chart types were only partially cached. ThoughtSpot now fully caches data for custom and Muze charts alongside other chart types, delivering faster and more consistent performance across all visualizations.

You can now encode a continuous measure, like profit margin or revenue, as a color gradient directly on new bar, scatter, and line charts. Instead of building a separate chart or table to surface a second variable, you can enable a gradient legend on any measure column and optionally set a "Vary by" measure to drive the color scale independently of the displayed value. This makes it faster to spot outliers, patterns, and directional trends at a glance. Hovering over a data point shows the exact value of the Vary by measure in the tooltip. This feature is compatible with the following charts in the new charts library:

You can now use data to drive the outline (border) color of chart marks such as bars, arcs, and points. Bind a field to the outline color encoding to automatically generate an ordinal color palette or a continuous gradient, complete with legend and tooltip support. Outline width and dash style remain configurable per data point. Existing outline configurations are fully backward compatible.

ThoughtSpot now renders a horizontal scrollbar across all chart types. For pivot tables, this removes the previous row-header column width cap, so attribute labels render at their full natural width instead of being truncated. The scrollbar behavior is consistent across chart types and viewport sizes.

ThoughtSpot adds two enhancements to Collections in this release.

Spotter memory from Liveboards, personal memory, and memory from conversations are now generally available (GA). Additionally, we have added a regenerate option to refresh your memory in the Memory sources page in order to stay in line with changes to underlying Liveboards.

We have replaced the previous + Add to memory button with a Remember this button. If you click Remember this below a Spotter answer, it triggers a command for Spotter to add the answer’s specifications directly to memory. You see a preview of the exact definitions it adds to conversational memory. Remember this is only supported for Spotter 3.

You can now prompt the creation of AI context from the Spotter optimization tab on a Model. Note that you must have edit access to a Model to create AI context. If you enable Spotter on a Model, it triggers AI context generation automatically.

Cluster admins can now view and manage Early Access flags that are not Org-aware directly from Admin > Feature Management > Early Access. Previously, these flags were only accessible via CLI and did not appear in the Feature Management UI.

Early Access flags that are not Org-aware appear as On/Off toggles. Toggling a flag applies the change cluster-wide, with no per-Org assignment required. Org-aware Early Access flags are unaffected and retain the existing per-Org assignment experience.

This feature is visible only to cluster admins. Org Admins do not see non-Org-aware flags in Feature Management.

You can now add a Scheduled Liveboards process type to a connection configuration, so scheduled Liveboard queries can use a separate warehouse, credentials, or non-OAuth authentication method instead of the parent connection’s defaults.

Look for the new Scheduled Liveboards checkbox under the Processes policy type in Connection Configuration, alongside the existing Row count stats option.

Administrators can now manage and remove parameterization on tables and connections directly from the ThoughtSpot UI. From the Data objects or Connections page, administrators can remove variable bindings from one or more objects or use the new Manage Parameters option to view, edit, add, or remove individual field-to-variable mappings without starting over.

Aggregate-aware query execution is now generally available (GA). Search data automatically switches from detailed Models to associated aggregate Models based on query granularity, reducing cloud costs and improving query latency without affecting insight accuracy. For more information, see Aggregate-aware Model switching.

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