10.14.0.cl Release Notes

You can now delegate Spotter coaching responsibilities without granting data model editing permissions. This feature is available for all data model editors and administrators. From the More menu on any Spotter-enabled data model, you can now grant the "Spotter Coaching Access" to other users or groups. Providing this access enables these users to manage all aspects of Spotter’s coaching, including the ability to view, modify and promote coaching from all users on that data model.

Adding Spotter context in reference questions is now on by default and available to all users. Rather than simply coaching Spotter to recognize search tokens in reference questions, you can now improve accuracy and generalize the coaching by adding context, natural language explanations of how to apply these tokens in similar queries. For example, you could tell Spotter to always add 'user name', 'user id', and 'user role' when asked about the list of users.

You can now teach Spotter core concepts, default behaviors, and specific data nuances by adding natural language instructions directly to a data model. These standalone directives act as global rules for Spotter to follow. For example, you can create a rule that makes Spotter automatically apply a "last 30 days" date filter if a user’s query doesn’t specify a time frame. To enable this feature, contact your administrator.

You can now add AI context to columns in any Model where you have Can Edit access. This context helps Spotter better interpret a column and its values, using this knowledge to differentiate between similar columns and answer your data questions more accurately. You can auto-generate this context automatically with a single click to get started.

In addition to selecting OpenAI Azure and Google Gemini as the LLM for your Spotter experience, you can now connect your own LLM by providing the LLM Key for one of the supported providers (Azure, Vertex) or via your own gateway. This option allows you to exercise full control over your AI stack, by connecting ThoughtSpot directly to your private LLM endpoints.

Tenant-based column aliasing introduces the capability to define a column name or description based upon the user’s membership in a ThoughtSpot Org or group. For example, Client 1 may refer to a column as Region, whilst Client 2 may refer to this as State. This feature is an extension of the language-based aliasing. Data analysts can define these aliases via CSV upload or with TML. This feature is particularly useful for embedded scenarios where the same underlying Model is leveraged across multiple clients.

For information about the new features and enhancements introduced in this release, refer to ThoughtSpot Developer Documentation.

For information about the new features and enhancements introduced in the latest release, refer to ThoughtSpot Mobile Documentation.