Organizations know the value of building their business on data if they want to beat the competition, live up to customer expectations, and create sustainable, lasting impacts. With that, large language models (LLMs) continue to create new opportunities regarding how companies turn data into insights that drive real action.
However, when it comes to creating new data insights and analytical content, this technology breaks down. GPT-3 (an LLM from OpenAI, for example) can’t handle complex data or data in multiple fact-tables, it lacks business logic, and it has issues working directly with data in a cloud data platform. This leads to inaccurate insights that can drive the wrong decisions, introduce risks, and erode trust from end users.
So, in comes ThoughtSpot, an AI-powered analytics company, and its new integration with GPT-3. The purpose of the integration is to make analytics accessible, trusted and scalable through AI. The integration includes the launch of ThoughtSpot Sage, a new search experience that combines LLMs like GPT-3 with ThoughtSpot’s patented search technology.
With this combination, all employees, no matter their skill level, can use AI-powered analytics on even the most complex data to ask and answer business questions, find accurate insights, and act.
Use cases of ThoughtSpot Sage include:
- Natural language search: Use LLMs to enhance the search-driven analytics capabilities in ThoughtSpot with true natural language. This means increased ease of use while preserving the accurate data discovery of the ThoughtSpot platform, along with improved search suggestions based on the search data model.
- AI-generated answers: Search existing analytical content or automatically create new insights, charts and visualizations based on natural language search. ThoughtSpot Sage then uses AI to create narratives explaining what insight is being surfaced, so users can easily understand a data visualization and act upon it.
- Data modeling: With AI-assisted data modeling that incorporates natural language, data professionals create search data models that allow AI to answer questions faster, while preserving the reliability and guardrails of the data model. Furthermore, analysts can automatically generate descriptions.
“LLMs promise to make interacting and conversing with data easier than it has ever been before, but doing so in a way that’s trustworthy and actionable will require the kind of analytics platform we’ve spent a decade building,” said Sumeet Arora, Chief Development Officer, ThoughtSpot. “We promise to make those conversations meaningful, accurate and actionable when it comes to analytics for our customers, partners and the cloud ecosystem at large.”
ThoughtSpot Sage, in preview, currently supports GPT-3 with support planned for additional LLMs.
Edited by Alex Passett