Business intelligence platform provider ThoughtSpot Inc. today is broadly enhancing its core application and launching a new mobile client.

New capabilities include expanded search options, additional connections to data sources and platforms, new ways to collaborate on analytics and options to deliver information without requiring a user login.

ThoughtSpot, which has raised more than $663 million in funding, has a platform that enables business users to search for useful patterns in operational data, such as how a new ad campaign is influencing customer demand. Coincident with its Beyond conference that begins today, the company is launching a new search engine in a private preview that combines foundational language models and its existing search technology.

With the recent addition of generative pre-trained transformer technology of the type used in the popular ChatGPT question-answering application, ThoughtSpot Sage makes it possible for nontechnical users to query data using natural language. Results are delivered within seconds and the embedded artificial intelligence capabilities provide recommendations for additional questions to ask. Users can train machine learning models by correcting keyword tokens.

The use of large language models also makes data modeling easier and faster by automatically generating synonyms for all data columns, the company said.

“When you ask a question, ThoughtSpot will understand the context and start a conversation with that answer that leads to the next question and the next question after that,” said Chief Executive Sudheesh Nair (pictured). “Every step of the way, we will tell you how we interpreted your intent using what we call ThoughtSpot tokens. Underneath that, we will show you the SQL we created.”

ThoughtSpot said Sage will be available on a free trial basis to all current and new free trial and team edition users.

Expanded integrations

The company today is also introducing integrations with common workplace productivity tools and office suites, including Salesforce Inc.’s Slack, Microsoft Corp.’s Excel and Google LLC’s Workspace. This allows users to share links from Liveboards, which are the company’s version of dashboards, in Slack and automatically generate visualizations. Slack users can then jump to the source of the visualization in ThoughtSpot for further exploration.

“We have a thesis that more analytics should happen outside of analytics tools in places like Google, Slack and Excel,” Nair said. “That is where data people are hanging out.”

Embeddable analytics

In keeping with that philosophy, the company is also introducing ThoughtSpot Everywhere, a toolset that enables developers to embed the company’s analytics into other applications.

A new assistant called Spot can answer questions posed in natural language from within Slack, querying back-end sources that include Excel, comma-separated values data and cloud data platforms such as Snowflake Inc.’s data warehouse and Google’s BigQuery. The capability will be available in the second half of this year.

Enhancements to Liveboard include improved interactivity with note tiles and custom sizing that can be used to add branding, explain terminology and provide context to data, the company said. Cross-filters enable users to apply a filter in one tile to the rest of the dashboard and a new in-app commenting system lets them leave feedback and collaborate without leaving ThoughtSpot.

Other additions to Liveboard include a styling framework that supports cascading style sheets and compatibility with open-source version control systems like Git.

The core platform is also getting a new data modeling studio within the ThoughtSpot Data Workspace that works with other platforms.

“As data is centralized in warehouses the definition of the semantic models and the metadata become extremely valuable,” Nair said. “We took a philosophical approach that ThoughtSpot will not be the only tool you will use, which means that we cannot make any proprietary artifacts. We have made everything available through [application program interfaces and software development kits] so if customers are modeling in [Google’s] Looker they can integrate with one click.”

Mobile smarts

A new version of the company’s mobile app, called Monitor for Mobile, enables users to monitor key performance indicators with proactive alerts that deliver insights without users having to ask for them, the company said. Once data teams set up the right data models and infrastructure, Monitor operates in the background and pushes notifications pushed when metrics change. The app learns from user behavior, automatically adjusting the information it delivers based on user ratings. It also explains why it made the changes it did.

“This is not about generating a dashboard; it’s about creating a continuum from data to insight and knowledge to action,” Nair said. “For an action to happen the right insight has to come together at the right time in front of the right person and often these are not aligned. Monitor will understand the intent, so if the system knows you’re responsible for customer churn in the northeast it sends you an alert saying your customer churn in some stores has gone up.”

Monitor for Mobile is currently in preview and will be available on mobile app stores in the coming months, ThoughtSpot said without offering specifics.

Photo: SiliconANGLE