live wire
▸Snowflake agent evals can pin judge versions; unpinned metrics may roll forward and shift scoresSnowflake▸ThoughtSpot makes Spotter 3 default for new instances; it requires download rights and always shares query data with an LLMThoughtSpot documentation▸MotherDuck adds image classification and object location to prompt_jev(), returning typed SQL outputs with confidence scoresMotherDuck▸BigQuery conversational analytics adds Montréal and Toronto; mixed-region agents still default to globalGoogle Cloud Documentation▸Fabric enables Azure OpenAI Copilot access by default, but OpenAI-subprocessor and cross-region processing/storage switches remain offMicrosoft Learn▸Snowflake Cortex Agents now continue by default when named tools are inaccessible, emitting warnings instead of failing the runSnowflake documentation▸Databricks Genie One can use group memberships to tailor answers, while Unity Catalog remains the access-control boundaryDatabricks documentation▸Fabric data agents can query Power BI models with Read alone; Build and workspace access are unnecessaryMicrosoft Learn▸Fabric data-agent Git configs expose column flags, but selecting a table still includes every columnMicrosoft Learn▸Google Data Agent Kit offers remote Spark kernels, but notebook exports can still move results into local Pandas or ArrowGoogle Cloud Documentation▸Cube says production analytics agents need explicit partial-result states, caller-filtered model search and durable artifactsCube▸Thinking Machines names Bridgewater behind ReViSQL-K2.6, citing 92.97% SC-16 accuracy at 15% of frontier-model costThinking Machines Lab▸AWS says a mismatched data/tool/prompt release cut its Field Sales agent pass rate from 78% to 38% in pre-productionAWS Artificial Intelligence Blog▸NL2SQL Multi-Agent ships evaluation artifacts across BIRD, CSpider, Archer and KDD DataSpace—but no software licenseGitHub▸Snowflake agent evals can pin judge versions; unpinned metrics may roll forward and shift scoresSnowflake▸ThoughtSpot makes Spotter 3 default for new instances; it requires download rights and always shares query data with an LLMThoughtSpot documentation▸MotherDuck adds image classification and object location to prompt_jev(), returning typed SQL outputs with confidence scoresMotherDuck▸BigQuery conversational analytics adds Montréal and Toronto; mixed-region agents still default to globalGoogle Cloud Documentation▸Fabric enables Azure OpenAI Copilot access by default, but OpenAI-subprocessor and cross-region processing/storage switches remain offMicrosoft Learn▸Snowflake Cortex Agents now continue by default when named tools are inaccessible, emitting warnings instead of failing the runSnowflake documentation▸Databricks Genie One can use group memberships to tailor answers, while Unity Catalog remains the access-control boundaryDatabricks documentation▸Fabric data agents can query Power BI models with Read alone; Build and workspace access are unnecessaryMicrosoft Learn▸Fabric data-agent Git configs expose column flags, but selecting a table still includes every columnMicrosoft Learn▸Google Data Agent Kit offers remote Spark kernels, but notebook exports can still move results into local Pandas or ArrowGoogle Cloud Documentation▸Cube says production analytics agents need explicit partial-result states, caller-filtered model search and durable artifactsCube▸Thinking Machines names Bridgewater behind ReViSQL-K2.6, citing 92.97% SC-16 accuracy at 15% of frontier-model costThinking Machines Lab▸AWS says a mismatched data/tool/prompt release cut its Field Sales agent pass rate from 78% to 38% in pre-productionAWS Artificial Intelligence Blog▸NL2SQL Multi-Agent ships evaluation artifacts across BIRD, CSpider, Archer and KDD DataSpace—but no software licenseGitHub
nl2sql.ai
releaseRelease

ThoughtSpot puts its SpotterCode agent inside the embed prototyping loop

The new Playground panel turns natural-language requests into runnable Visual Embed SDK code, while leaving execution and conflict resolution with the developer.

AI-assisted embed coding in a developer playground.
Side by side: what changed
By The News Desk· Sep 12, 2026the quick take — two AI hosts go live when you do

ThoughtSpot has added its SpotterCode coding assistant to the Visual Embed Playground, putting prompt-to-code generation beside the editor and live preview developers already use to prototype embedded analytics. The company says the assistant is available now in the Playground for ThoughtSpot instances where the feature is enabled. ThoughtSpot’s announcement positions the addition as a faster route from an intended embed behavior to working SDK code.

What changed

Developers can choose an embed component, describe the desired result in natural language and have SpotterCode update the Playground’s code editor. The generated code targets ThoughtSpot’s Visual Embed SDK and REST APIs; developers then click Run to execute it in the preview pane and can continue with follow-up prompts or edit the code directly. The product documentation lists supported Playground contexts including Search Data, Liveboards, individual visualizations, the full application and the Spotter analytics agent.

Starter prompts illustrate the intended scope: applying runtime filters, keeping a data panel collapsed, adding custom actions, enabling Spotter in a full-app embed and changing interface labels. That makes this less a general-purpose coding agent than a narrowly scoped configuration assistant for ThoughtSpot’s embedding surface.

There is an important guardrail in the workflow. According to ThoughtSpot’s documentation, SpotterCode does not update the editor when it detects a conflict with existing code. The developer must review and clear the conflicting section before trying again. Generated output is also not run automatically; execution remains a separate click in the Playground.

Why it matters

SpotterCode was already designed to connect AI-native development environments such as Cursor and Claude Code to ThoughtSpot’s SDK and API knowledge through an MCP server. The Playground integration moves the same grounded-assistant idea into a disposable prototyping environment, where teams can test an embed configuration before carrying the resulting code into an application. ThoughtSpot distinguishes the IDE use case—working within a project’s structure and design system—from the Playground’s faster explore-and-preview loop.

For analytics teams, the practical significance is not autonomous software delivery. It is a shorter path through a configuration-heavy SDK: prompt for a component or customization, inspect the generated implementation, run it against the instance, and iterate. The feature remains bounded by the APIs and configuration options ThoughtSpot exposes, and its own conflict behavior reinforces that developers still own review and integration.

The release is another sign that analytics vendors are moving natural-language interfaces beyond asking questions of data and into building the surrounding analytics experience. Here, the agent is not writing a SQL query for an analyst; it is helping a developer assemble the product surface in which those queries and answers will appear.

sources

Filed by The News Desk. Corrections: desk@nl2sql.ai · Our standards →

comments · 0

    Comments are moderated before they appear. Your email is used once to confirm it is you — never shown, never sold. Corrections and questions get an answer from the desk when we have one.

    more from the desk