ThoughtSpot Sage: How to Use it for AI Insights

ThoughtSpot is a BI tool known for its AI-powered analytics. In the data analytics industry, this was one of the first vendors to put a major focus on generative AI. One of their offerings is something called ThoughtSpot Sage.

It promises to democratize data analytics by allowing anyone with basic computer skills to explore cloud data and get custom visualizations based on their needs.

But does it really work like that in practice? Let’s find out.

What is ThoughtSpot Sage?

ThoughtSpot’s analytics platform allows business users to create dashboards and visualizations based on data sets. For example, if you want to create a sales dashboard for your business, you would pool your data sources, connect them to ThoughtSpot and then a specialist analyzes that data and creates visualizations, to be seen by the end-user.

For ThoughtSpot, this sounded a bit complex, so they introduced Sage, a natural language search model that allows anyone to type in a query and get a response. Like all other major language learning models (LLMs), Sage is based on GPT-4 by OpenAI, which makes it fairly easy to use for most people.

ThoughtSpot connects to a wide variety of data sources, such as Snowflake, Databricks, Google BigQuery, 

It’s a type of human-in-the-loop model where AI tools and LLMs do the heavy lifting, but a human being influences the final outcome.

How does it work in theory?

In the past, data analytics and visualizations were only intended for data scientists, engineers, and analysts. ThoughtSpot Sage lets you upload data models and then ask Sage, the AI assistant, about the data inside.

‍

For example, say that you want to find out how your sales are doing. You connect your data sources, prepare a data model, and then anyone on your team can use artificial intelligence to get answers about that data.

For example, a marketer can ask “What is our top selling product in Q1 of 2024?” and Sage uses natural language processing to answer. They can quickly find out what’s new with their data, without involving data specialists.

The premise is that these AI-generated answers can bring business intelligence closer to the average user. You’ll no longer require large data teams to understand and visualize complex data or follow up with them many times to get answers to your questions.

The AI-powered search experience can be used internally, in your company, or you can embed it in your application, allowing the end users of your SaaS product to explore data without knowledge of SQL or data analytics. You don’t even need to know how to use Excel to be comfortable with Sage.

How does it work in practice?

There is certainly a use case for large language models in business intelligence and ThoughtSpot Sage fills that gap, sort of. Most of the reviews out there on Sage state that the tool is used in larger businesses. 

In other words, there is a team of data analysts that prepares the data, creates the data model, validates the data and as one user puts it, acts as the “methodology source of truth”.

According to this user, the GPT works fairly well as the company does not use the more advanced features in ThoughtSpot. The end-user can save time and type in their query to get a quick update on their programs.

In other words, the search technology works well, but under certain circumstances.

First, you need to have someone in-house to prepare, validate and model the data. Second, you need to have enough users who need to explore the data, and you need to get them familiar with the basics of data science and business intelligence.

Does ThoughtSpot Sage make sense in the modern business environment?

The promise of typing in a natural language query and getting a response with a granular overview of your data is enticing. The average business development manager, for example, can find a KPI on their own instead of ringing up the IT department every time they need a chart.

At the moment of writing this article, ThoughtSpot Sage is not there yet. The product itself is promising, with a decent API, mobile capabilities, solid entry-level pricing and functional liveboards.

The use case for Sage is very limited, though. It’s an interesting offer for:

• Companies with existing data science teams, engineers and analysts
• A large number of end-users who want quick access to that data
• Companies with end-users who know enough about data analytics to get the maximum out of self-service tools like Sage and its search data

In other words, if you have a small team and you want a BI tool with a great user experience that is easy to get started with, tough luck. ThoughtSpot Sage is not meant for beginners, even though it may seem like it.

If you don’t know a single thing about business intelligence and you’re looking for an easy entry into this world, Sage won’t give you that. To get usable data for end-users, you still need to fine-tune it, prepare the data model, validate it and get it ready for Sage.

The irony is that companies that have all the right conditions and a good number of data specialists don’t need tools like Sage. If a need arises for specific insights, the data analytics team can fetch them more quickly than it would take the end user to learn how to use Sage. These companies typically use something with more market presence, such as Microsoft’s Power BI or Tableau.

To sum up, the type of company that can benefit from a tool like ThoughtSpot Sage is not easy to find. Until ThoughtSpot Sage requires less work on data preparation and analysis, this tool will have a very narrow use case.

Wrapping up

ThoughtSpot Sage is a great tool in theory. In practice, it promises a lot but does not deliver. Or at the very least, it delivers to a very niche audience. For AI-powered analytics and GenAI to be widespread in BI tools, you’re going to need a tool more refined than Sage or employees with more than a basic understanding of data analytics.

Why not try a tool that delivers on its promises? At Luzmo, our AI-powered embedded analytics tool helps you create dashboards with just a few clicks and give your end-users analytics insights they can understand.

Want to learn more? Book a demo with our team for more information!