whyframeshot - stock.adobe.com

Tableau Einstein is a new BI platform with AI at its core

The new platform marks an evolution for the longtime analytics vendor, making AI the focus with capabilities such as agent-based insight generation and real-time analysis.

Tableau Einstein represents a new era for Tableau.

The new AI-based analytics platform, which combines existing capabilities from Tableau and Salesforce with new features, shifts analytics away from reports and the dashboards that have been Tableau's historical focus to insight delivery within the flow of a user's work.

The platform is built with four categories of capabilities. They include those that use AI to deliver autonomous insights, enable the development of a semantic layer aided by AI, provide access to real-time insights, and create a marketplace for data and AI products to enable customers to compose their own instances of Tableau.

Currently available are capabilities such as Tableau Pulse and Tableau Agent that deliver autonomous insights. Others such Tableau Semantics and the marketplace are scheduled for general availability in 2025.

Access to Tableau Einstein, meanwhile, is available through a Tableau+ subscription. Pricing for Tableau+ is not made public.

Based in Seattle, Tableau is a longtime analytics vendor that has been a subsidiary of Salesforce since being acquired by the CRM giant in June 2019. Over the past two years, like many other analytics vendors, Tableau has made AI, including generative AI, a focus of its product development.

Toward that end, the vendor released Tableau Pulse, a generative AI-powered tool that surfaces insights and delivers them to users in natural language, in February. In addition, Tableau Agent, an AI-powered assistant that enables users to prepare and analyze data and was formerly known as Einstein Copilot for Tableau, is being rolled out with Tableau Agent for Tableau Prep made generally available in July.

Now, Tableau is building an entirely new platform with AI as a foundation. Given its blending of existing Tableau and Salesforce capabilities to go with new ones under development, it represents a new level integration between Tableau and its parent company, according to David Menninger, an analyst at ISG's Ventana Research.

"These new capabilities are a real combination, maybe even the realization of true integration, between the Tableau stack and Salesforce," said.

Donald Farmer, founder and principal of TreeHive Strategy, likewise said that Tableau Einstein is a significant integration of capabilities.

It remains to be seen whether Tableau Einstein's AI capabilities are as robust as those being developed and delivered by Tableau's many competitors, he noted. But whether better, as good or merely good enough, Tableau has put together a new platform that is more than just add-ons.

"It's an impressive release," Farmer said.

The platform

One aspect of Tableau Einstein is that it is agentic in nature, meaning that AI-powered capabilities -- agents -- provide insights without being prompted.

Traditionally, analytics users come up with a business question, query their organization's data, look at dashboards and reports developed based on the query, and come up with insights that lead to decisions. Tableau Einstein aims to change that by proactively delivering insights to users within the flow of their work, which reduces the need for data consumers to come up with their own questions and instead provides them with information they might not have known they need.

That's the concept of autonomous insights and includes features such as Tableau Pulse and Agentforce for Tableau, which is a tool that enables customers to develop autonomous agents.

Meanwhile, by developing a more agentic platform that reduces the need to query data and looks at dashboards and reports for insights, Tableau is providing customers with capabilities that are in line with where analytics is headed, according to Menninger.

"The market is moving toward agentic AI and agentic analytics. So the notion of building agents rather than relying on dashboards represents a paradigm shift in how organizations utilize data," he said. "Rather than data appearing in dashboards with data interpretation left to the viewer, agents can initiate actions based on the data."

Beyond providing a more agentic experience, Tableau Einstein's autonomous insight generation capabilities are put together in such a way that they stand to be of significant benefit to users, according to Farmer.

He noted that the platform repackages certain pre-existing capabilities. But by combining them with new ones, it provides an improved analytics experience.

"The … application of Tableau Pulse has been elegantly conceived. The experience is intuitive and powerful, and the Agentforce AI is [included] in a way that feels deeply integrated rather than bolted on," Farmer said.

A key feature that will underpin generative AI-powered autonomous insight generation is Tableau Semantics, which is expected to be generally available in February 2025.

Tableau Semantics is a semantic layer aimed at improving the quality of AI models and application by enabling enterprises to define and describe data uniformly. That way, it can be easily discovered and used to train AI. In addition, it can be used to create and manage metrics, data dimensions, relationships between datasets and goals for the use of data.

Tableau Semantics will come with some prebuilt metrics for data that originates from Salesforce applications and some of Tableau's partners. It will also be infused with AI capabilities designed to help organizations build their semantic layer and manage it.

Tableau is not the first analytics vendor to include a semantic layer. Others include MicroStrategy and Looker. The infusion of AI, however, aims to make Tableau Semantic more agile than the semantic layers of the past that require significant manual labor, according to Southard Jones, Tableau's chief product officer.

The market is moving toward agentic AI and agentic analytics. So the notion of building agents rather than relying on dashboards represents a paradigm shift in how organizations utilize data. Rather than data appearing in dashboards with data interpretation left to the viewer, agents can initiate actions based on the data.
David MenningerAnalyst, ISG's Ventana Research

"The concept of a semantic layer is not new," he said. "But they've been hard to manage. They prevent agility. But if you can leverage AI to accelerate building the metrics as well as managing them, you provide data stewards with the single source of truth they want and analysts with a playground to do things super-fast."

Like Tableau Semantics, real-time data is key to providing users with autonomously generated insights.

Real-time data in Tableau Enterprise is enabled through an integration with Salesforce's Data Cloud and is now generally available. The integration provides Tableau Einstein users with access to hundreds of data sources -- not just Salesforce -- so they don't have to move their data. In addition, it enables users to combine data from disparate sources in a secure, scalable way.

Perhaps most significant, according to Jones, is that by using Data Cloud as a foundation for accessing real-time data, Tableau Einstein users will be able to use unstructured data in concert with structured data.

For decades, analytics platforms only supported structured data such as financial records and point-of-sale transactions. Now, however, it is estimated that over 80% of all data is unstructured, such as text, images and audio files. For an enterprise to get a full understanding of its operations, or for AI models and applications to be trained with as much relevant data as possible, unstructured data is essential.

"Data Cloud gives us access to unstructured data that we didn't previously have with Tableau and the ability to combine that with structured data in a simple way," Jones said.

Tableau users were able to combine structured and unstructured data before the advent of Tableau Einstein, he continued. But it took a lot of manual labor using tools from numerous different vendors.

"Now, this is something that is just part of the application," Jones said.

Finally, the marketplace will enable Tableau Einstein customers to develop a composable infrastructure, allowing them to pick and choose capabilities to personalize their instances through APIs. With general availability expected in June 2025, the marketplace aims to help customers to share data and AI assets across departments so they don't get isolated, fuel easier data discovery, and even make products available to other Tableau customers through Tableau Public.

Farmer noted that Tableau users have often struggled to collaborate with others outside their departments. As a result, the marketplace will be an important addition.

"Composable and reusable assets – such as models, visualizations and dashboards -- are significant advances," Farmer said. "Tableau users have often worked in departmental silos, and both IT managers and individual users have struggled with a lack of collaboration and knowledge sharing. The new API-driven reuse and internal marketplace features will be helpful here."

While helpful and representative of shifting focus for Tableau and a significant evolution for Tableau that provides important new tools for its users, Tableau Einstein does not markedly differentiate Tableau from its closest competitors, according to Menninger. It puts Tableau at the forefront, perhaps, but not in a way that will separate Tableau from peers such as Qlik, MicroStrategy and Microsoft Fabric over the long term.

"Tableau's current capabilities are competitive among the other BI platform vendors, but not entirely unique," Menninger said. "There are some components of the platform coming from Salesforce … that may give Tableau a leg up in the short run. But it is such a fast-changing market that no vendor will be the undisputed leader for any significant period of time."

Farmer, however, noted that even if Tableau Einstein includes features being offered by other vendors, its autonomous insight generation is an advancement beyond what most other vendors are doing.

"Tableau has done great work here in creating an agentic experience that feels, for the first time, like the real deal," he said.

In the future

With parts of Tableau Einstein generally available and others expected by the middle of next year, Tableau's roadmap for the new platform includes further focus on agentic AI, according to Jones.

Essentially, a goal is to provide each user with a personal analyst to comb through data and uncover insights.

"It's not just about productivity," he said. "It's about changing the value of what can be delivered."

Continuing to focus on providing users an even deeper agentic experience is wise, according to Menninger. He noted that although Tableau built its reputation on providing customers with dazzling dashboards, the vendor's movement away from dashboards is in line with where BI is headed.

"Dashboards are like data warehouses," Menninger said. "They don't solve any problems on their own. It's what you do with the information in dashboards or data warehouses that really matters. So a continued push down the path to agentic analytics and collaborative decision-making would be good for Tableau's customers and the market in general."

Eric Avidon is a senior news writer for TechTarget Editorial and a journalist with more than 25 years of experience. He covers analytics and data management.

Dig Deeper on Business intelligence technology

Data Management
SearchAWS
Content Management
SearchOracle
SearchSAP
Close