On Wednesday, Thoughtspot announced a new edition of Thoughtspot Embedded, adding agent AI capabilities that allow customers to embed into their applications so that users can benefit from agent-powered insights within their workflows.
Previous Edition of Thoughtspot Embedded was called thought pots everywhere before it was rebranded, allowing customers to embed AI capabilities into their user workflows. Now, customers can embed Spotter 3, an updated version of Thoughtspot's Agent AI-powered search engine, in addition to the structured data that has historically allowed thought spots to analyze users, to allow users to access unstructured data.
Additionally, the Thoughtspot built-in update allows users to embed AI-Aigmented Dashboards. It also includes new tools for developers to build applications powered by Thoughtspot that look like their organization's products.
There are some very important and widely sought-after improvements that make the embedded experience more complete in terms of analytical capabilities and more complete for developers. I think this will drive a significant advance in the embedded market.
Donald FarmerFounder and Principal of TreeHive Strategy
Given that the embedded Thoughtspot includes new agent AI capabilities and adds valuable tools for developers, this update is important for Thoughtspot users, especially independent software vendors who use Thoughtspot to power applications they deliver to their customers.
“There are some very important and widely sought-after improvements that make the embedded experience more complete in terms of analytical capabilities and more seamless for developers,” he said. “I think this will drive them significantly ahead in an embedded market, especially since we can talk good about both business intelligence and AI integration.”
Based in Mountain View, California, Thoughtspot has provided an AI-powered business intelligence (BI) platform since its launch in 2012, allowing customers to search and analyze data using natural language rather than code. However, before Openai enabled CHATGPT in November 2022, users still needed technical expertise and data literacy to use Thoughtspot.
AI injection
Before the Generator AI enabled true NLP, BI was limited to trained experts and self-service users with data literacy skills. As a result, for decades, only about a quarter of the employees in the organization used data to inform their work.
Embedded analyses are one way that has successfully enabled some organizations to benefit from BI, providing data-based insights within normal workflows rather than enforcing learning complex analytics platforms.
Currently, embedded analytics is expanding to include AI features such as freeform natural language queries directly within applications such as Salesforce and Workday.
Like Farmer, ISG Software Research analyst David Menninger is called the Thoughtspot Embedded Update “An Evolution” to advance the AI capabilities that users can embed.
“It's important that software providers make additional agent features available in built-in versions,” he said. “Other new features… you can easily brand your embedded features and embed them.”
According to Francois Lopitaux, Senior Vice President of Product Management at the vendor, Thoughtspot Embedded is based on four core pillars. They include:
Embedded intelligence refers to AI and BI injection in applications through Thoughtspot's Visual Embed SDK and REST API.
Native experiences that include an integrated developer playground and mobile SDK where developers use Thoughtspot features but can create web and mobile applications designed with the look and feel of the organization's tools.
Seamless Workflow covers Thoughtspot's Model Context Protocol Server, allowing users to extend Spotter's source data to AI platforms such as ChatGPT and Anthropic's Claude.
An open ecosystem that includes Thoughtspot's cloud-independent architecture, layered above the vendor's agent semantic layer and built-in governance capabilities, ensures that your data is reliable.
“The four pillars make sense,” Menninger said. “They are dealing with real issues that were obstacles to reaching more workforce in their analytics. They have made good progress in the embedding and native experience.”
Meanwhile, according to Lopitaux, determining what features to include in Thoughtspot Embedded was driven by customer feedback.
“We had a lot of customers who were already using the solution, but all of these customers wanted more,” Lopitaux said. “The Mobile SDK was one of the biggest requests I wanted to add. Also, allowing me to embed spotters in my application is definitely the top of the mountain as I want to provide an agent experience to my customers.”
One such client is Navan, a corporate travel and expense management company. With Thoughtspot built in, Navan added Thoughtspot's AI capabilities to its website, allowing customers to query data using natural language. For example, Navan users can ask which employees have traveled, where they traveled, and how many employees spent on the trip, gaining insight into how they use their travel budget.
“[Embedding AI] It helps to avoid spreading your dashboard because you don't know what your follow-up questions are. Lopito said. [IT] Update the visualization. We break that by allowing people to ask questions directly and get answers right away. ”
Spotlights in Spotter
Though Thoughtspot Embedded Update is based on four pillars, the latest version of Spotter allows agents to act as MCP hosts, allowing them to connect to unstructured data sources such as Slack, improving integration with the agent's large language model to increase intelligence, breaking complex requests into multi-step questions and verifying output.
“The integration of MCP stands out,” Farmer said. “I rather like the research mode where I laid out the multi-step inference process for analysis.”
However, Spotter 3 represents Thoughtspot's first foray into the field to provide Agent Teasoning capabilities, leaving room for improvement, he continued.
“This is in the early stages,” Farmer said. “I would like to see this multi-step process become more interactive with users who review their research plans and fine-tune them in a different direction after review.”
Menninger similarly highlighted the value of Spotter 3, which serves as a research mode for MCP hosts and agents.
“The MCP server connection allows Thoughtspot users to easily incorporate analytics into other generation and agent AI tools,” he said. “Research ability supports deeper analysis and reporting based on thought spot analysis.”
Together, improvements such as adding interactive components to Spotter's multi-step inference capabilities are needed, but according to Farmer, the advancements in Thoughtspot Embedded and Spotter 3 will help the needs of thinking spot users. Additionally, he said the update could make the vendor's built-in analytics capabilities more competitive with the capabilities of embedded BI specialists.
“Overall, this is a great release that addresses many weak points of the built-in storyline of thought spots and makes them more competitive,” Farmer said. “We're thinking about other vendors – Sisense in particular, but we need to be careful.”
Meanwhile, Menninger pointed out that while it's valuable to Thoughtspot users and to keep the vendor competitive, the update doesn't necessarily separate from vendors like Tableau, which also offers Thoughtspot's embedded analytics and agent AI capabilities.
“It's really hard for software providers to distinguish products in mature markets like business intelligence,” he said. “Almost every vendor races towards the same goal of making analytics easier to use and make it easier to integrate analytics into business processes. …Products like Tableau Next are similar to Thoughtspot Embedded and Spotter.”
Eric Avidon is a senior news writer at Informa TechTarget and a journalist with over 25 years of experience. He covers analytics and data management.
