From Queries to Agents: ThoughtSpot’s Bold Leap into Agentic Analytics

 

Logo image courtesy of ThoughtSpot

Today, analytics is not about data, reports, and dashboards anymore. It is not even about insights. It seems ThoughtSpot is aiming for something bigger: agents that think, act, and close the loop, all on your data stack.

If the last few years in analytics were about self-service and natural language, 2025 is clearly about something new: AI agents. Not just tools that explain business, but ones that act on its behalf, and ThoughtSpot, the company that once positioned itself as the Google of BI, is now doubling down on that vision with the launch of its Agentic Analytics Platform.

This is not just rebranding. It seems to be a strategic repositioning, backed by AI agents, real-time cloud-native architectures, and native integrations, especially with Databricks and Snowflake. It is also a signal that ThoughtSpot wants to move way beyond charts and dashboards and instead become a living, breathing part of the new enterprise decision loop.

But as always, hype is easy, so let’s break down what ThoughtSpot actually announced, what it means in practice, and what enterprise leaders should really be watching.

What was Announced?

Let us start with the topline: ThoughtSpot launched its new Agentic Analytics Platform, a combination of its core analytics engine with new AI-powered capabilities, embedded agents, and full support for Snowflake and Databricks ecosystems. The highlights of the announcement include:

• DataSpot. A new AI-powered assistant tailored for Databricks environments.
• SpotIQ Agent. A generative agent that identifies, explains, and recommends actions based on analytics.
• Natural Language Commands, not just for querying but for triggering actions
• Embedded actions within the analytics experience itself, allowing users to not just explore but also initiate workflows.
• Native integrations with Databricks Unity Catalog and Snowflake Cortex AI.

ThoughtSpot describes this as moving from "insights to actions" and even to "auto-actioning", meaning agents that do not wait for user prompts but proactively monitor, detect, and act. In other words, analytics that no longer need an analyst.

A Big Claim: Let Us Unpack It. 

Why It Matters?

At the heart of this shift there is something subtle but important: AI agents.

In the classic BI world, tools answer questions. But with agentic analytics, tools are expected to ask questions, decide what matters, and even act with minimal or no human prompting.

This is a much bigger leap than what we have seen with natural language analytics. In fact, if we trace the evolution of ThoughtSpot, the trajectory becomes clear:

1. Search-based BI (ask a question, get a chart.)
2. AI-assisted analytics (auto-insights, anomaly detection)
3. Agentic analytics (systems that reason, recommend, and respond)

And it is not just marketing speak; the integration of DataSpot with Databricks enables ThoughtSpot to connect directly to Unity Catalog, understand lineage, and navigate data access control, all crucial for enterprise-grade agent behavior.

These agents are not just parroting outputs from a model. They are operating within governance frameworks, aware of context, permissions, and business rules.

DataSpot: Agentic Analytics Meets the Lakehouse

Let us talk about DataSpot, the Databricks-specific agent ThoughtSpot just launched.

Built for customers on the Databricks Lakehouse Platform, DataSpot is designed to live inside the data ecosystem, leverage Unity Catalog for semantic understanding, and use AI to surface both exploratory insights and recommended actions.

What makes DataSpot different from other “chat with your data” assistants? Well:

• Contextual intelligence: It understands data definitions, security controls, and relationships defined in Unity Catalog.
• Actionable insights: It not only tells you what changed (e.g., revenue dropped) but also why (e.g., a specific segment churned) and what to do next (e.g., adjust pricing or reroute campaigns).
• Agent loop: DataSpot does not just wait for users to query. It can proactively surface anomalies or opportunities.

According to the company, this is not just GenAI on top of data. It is about injecting autonomous behavior, what ThoughtSpot calls “agentic intelligence,” into the entire business data stack.

As Ketan Karkhanis, Chief Executive Officer at ThoughtSpot, mentions:

“With DataSpot, we’re bridging the gap between centralized data operations and decentralized business insight. By combining Databricks’ powerful storage, compute, and governance with ThoughtSpot’s trusted, AI-powered analytics, we’re enabling hundreds of customers like Sephora, Chevron and Unilever to harness the full potential of their data estates—faster and more intelligently than ever before. Our agentic platform is designed to be a true thought partner, bringing perception and reasoning to every business user, so they can confidently drive their business on data and AI.”

Snowflake, Meet Cortex AI + ThoughtSpot Agents

ThoughtSpot did not stop with Databricks; its Snowflake partnership got a boost too, with tighter integration into Snowflake Cortex AI, the LLM and ML model suite designed to bring AI closer to Snowflake’s compute fabric.

Through this integration, ThoughtSpot agents can now tap directly into Snowflake’s ML functions, language models, and vector search, all without data leaving the Snowflake environment.

The implications are big:

• For security-conscious organizations, this means analytics and AI do not need to move data around, reducing risk and improving latency.
• It also means real-time analysis at scale becomes much more feasible—since ThoughtSpot can call Cortex AI features natively from within the warehouse.

For users, this could look like an agent that detects a drop in customer engagement, calls a sentiment analysis model within Cortex, and recommends changes to a campaign, all without a single query being entered manually.

ThoughtSpot Agents: From Chat to Closed Loop

Now, let us focus on the concept of agents.

Everyone has gotten a chatbot now, but ThoughtSpot’s take on agents moves beyond Q&A.

For instance, the SpotIQ Agent works in background mode, scanning datasets, triggering anomaly detection, and surfacing insights autonomously.

But more importantly, it can trigger actions, whether it’s alerting a team in Slack, updating a field in Salesforce, or kicking off a campaign in HubSpot.

In short, agents are becoming the glue between data and action, and ThoughtSpot is positioning itself as the layer that operationalizes these agents across cloud platforms.

That is why this is not just a feature release; it is a reframing of what analytics do.

Embedded Analytics Gets Embedded Actions

One of the most practical parts of this release is the inclusion of embeddable action elements. ThoughtSpot already allowed analytics to be embedded inside apps, but now, it lets users embed actions as part of that experience, so, for example:

When the agent detects churn risk in a dashboard, the user does not just get notified; they get an action button to initiate a retention campaign.

This is about closing the loop, turning insight into a real-world result, and ThoughtSpot is betting that this "last mile" of analytics—operationalizing insights—is where agentic analytics will make its biggest impact.

Governance, Guardrails, and Real-World Readiness

Now let us talk about reality. For all the buzz about agents, the biggest risk is governance. Enterprises will ask:

• Can these agents be trusted?
• Are actions traceable?
• Are models and logic transparent?

ThoughtSpot seems aware of this. The use of Unity Catalog and Snowflake Cortex is not just for performance; it is also about ensuring metadata, permissions, and auditability are enforced across the stack.

With that said, success will depend on execution. Can ThoughtSpot offer no-code or low-code agent design for analysts while still offering dev-level control for IT and governance teams?

That is the balance that will define adoption.

A Platform Redefined

With this launch, ThoughtSpot is no longer just a search-first BI tool. It is aiming to become the agentic layer for the modern data stack, a control plane where AI does not just assist humans but acts with them, and sometimes without them.

Sort of what Salesforce is to CRM and HubSpot is to marketing automation, ThoughtSpot wants to be, or seems to be heading to be, for agentic analytics:

• Fully integrated.
• Operationally embedded.
• Built for action, not just exploration.

It is a bold move and one that reflects where enterprise analytics is heading. The age of agents is not theoretical anymore; ThoughtSpot just put it into production.