AI combined with business intelligence (BI) could be a game-changer for analysts and consultants alike. Analysts now have more tools at their disposal enabling them to scan large datasets for patterns and anomalies using NLP. Now, with autonomous AI agents taking over analytics, companies already leveraging AI are stepping into the next phase.
California-headquartered BI and analytics company ThoughtSpot, with a significant presence in Bengaluru, recently launched its latest innovation: Spotter, the agentic AI tool that functions as a virtual analyst to assist businesses.
“When we launched Spotter, one of the first things we did was open it up so anyone can start using it right away. It’s one of the coolest AI agents for analytics out there,” said Ketan Karkhanis, the new CEO of ThoughtSpot, in an exclusive interaction with AIM.
Conversational AI for BI
From its inception, ThoughtSpot has aimed to demystify data for users. “We have been on a mission to let people really understand what’s happening in their data,” Karkhanis said, emphasising the company’s effort to redefine self-service BI. However, he believes that self-service BI, as traditionally implemented, is a flawed concept.
“Self-service is the biggest hoax in the industry. When they say self-service, they essentially mean ‘Go build your own dashboards’,” he said, adding, “Do you wake up in the morning and say, ‘I want to build some dashboards today’? No, you just want to run your business.”
With Spotter, ThoughtSpot aims to address these challenges by enabling conversational interactions with data.
Karkhanis highlighted that since humans do not communicate in the language of data, and nor does data communicate on human terms, there’s a need for an interface that caters to both. Spotter addresses this challenge by allowing users to pose complex, multi-step questions in natural language and obtain accurate, contextual responses.
“Across the world, customers are looking at their BI stack and realise that this is not meant to solve problems in the future. It’s hard for them to bolt AI to that stack; they can’t simply put AI on it. They need an analytics platform that’s built ground up for AI – and that’s driving ThoughtSpot’s growth,” said Karkhanis.
Interestingly, Spotter integrates seamlessly with all leading cloud platforms and LLMs, including the GPT-series models and Google Gemini.
Copilot and Agents
Thoughtspot was co-founded in 2012 by Ajeet Singh, who had also co-founded cloud computing company Nutanix.
With LLMs emerging as powerful tools for data interpretation and the rise of copilot-enabled solutions, analysts now have an abundance of options at their fingertips. Moreover, there are now plenty of AI agents introduced in the enterprise suite of products by big tech companies, such as Oracle, Microsoft, Salesforce and others. However, not all may be suited for analysis.
“Many customers have told us that they spent six months trying to make it work on ChatGPT, but it doesn’t. ChatGPT never claimed that it would work. It’s not interested in this since it’s not its primary business,” co-founder Singh told AIM.
Interestingly, the definition of agents is also being debated. Recently, Salesforce co-founder and CEO Marc Benioff criticised Microsoft’s Copilot as disappointing and not delivering accuracy.
“I have yet to find anyone who’s had a transformational experience with Microsoft Copilot or the pursuit of training and retraining custom LLMs. Copilot is more like Clippy 2.0,” he said.
The discussion on copilot and agents also brings into focus the definition of what each one is. Karkhanis highlights the clear distinction between assistive tools and true autonomous agents. He explains that while many systems today, such as Microsoft’s Copilot, operate on single-turn Q&A, answering one question at a time, they lack reasoning, adaptability, and ability to learn a user’s business to be called autonomous.
“There are a lot of nuances to this. If you can’t coach it, then it’s not an agent. I don’t think you can coach a copilot,” said Karkhanis. “You can write custom prompts [but] that’s not coaching.”
ThoughtSpot’s vision has been to enable users to understand data through a conversation with their data. “The company has always approached relational search as being the ‘Google for data.’ Now, it positions itself as the ‘Google plus ChatGPT for data,’” concluded Kharkanis.
