Evolving Tableau touts tools to fuel AI-powered analytics
Proactive BI and a natural language-powered app development tool show progress toward an AI-powered platform, while tight integration with Salesforce provides differentiation.
The latest tools from Tableau show that the vendor is transforming its platform from a place for traditional BI to one for agentic AI-powered analytics.
Tableau's aim is to enable business users to observe, understand and act on data in real time, according to general manager Mark Recher, who spoke during the Tableau Keynote address at parent company Salesforce's annual Dreamforce user conference.
Toward that end, Tableau's bedrocks for enabling enterprises to build and benefit from agentic AI-fueled analytics capabilities include empowering each user to be an application developer, AI-powered analytics that surface and deliver insights to users in their workflows rather than forcing them to visit a BI environment, and enabling users to create trusted data foundations for their data and AI products.
New features designed to provide users with those bedrocks include Tableau Studio where analysts can vibe code via natural language to create applications, Proactive Intelligence to deliver insights within user workflows, and Tableau Knowledge to provide a context layer for AI and analytics to draw upon.
"They're a logical evolution of the platform," Mike Leone, an analyst at Moor Insights & Strategy, told TechTarget. "Back in May, Tableau introduced its knowledge layer, which keeps a company's business definitions in one place. [Application development] and Proactive Intelligence draw on those definitions."
However, rather than distinguish Tableau from competitors such as Microsoft Power BI, Qlik and ThoughtSpot, the new features show that Tableau is undergoing the same repositioning as other analytics specialists adjusting to the burgeoning era of AI.
"Tableau continues to be a leader, but the whole market is moving in this direction," he said. "They're building a lot of the same pieces, from semantic layers and agents to connections into AI assistants. One of Tableau's biggest advantages is Salesforce, which is so deeply embedded in businesses that Tableau can go from spotting a problem to acting on it without switching systems."
Evolving from BI to AI
With AI now enabling almost any business user to query and analyze their organization's data using natural language, Tableau and other traditional BI vendors have had to transform to remain relevant to their customers. All, in their own ways by repurposing existing capabilities and adding new ones, have prioritized providing users with a trusted foundation for analytics and AI as a means of staying essential.
Tableau continues to be a leader, but the whole market is moving in this direction. They're building a lot of the same pieces, from semantic layers and agents to connections into AI assistants.
Mike LeoneAnalyst, Moor Insights & Strategy
Tableau is doing so by making its longstanding semantic layering capabilities -- including the data products and business logic organizations have used semantic models to build -- a crucial part of Tableau Knowledge, which it first unveiled in May.
The additional new capabilities Tableau is touting, all of which are scheduled for general availability by the end of October, are designed to work with the knowledge layer to form a complete agentic analytics platform, according to Recher.
"You become an agentic analytics enterprise by taking probabilistic intelligence, and the speed and power of that, and marry that with your trusted data foundation," he said at Dreamforce.
Meanwhile, the new features show that the vendor is meeting the needs its customers as they change by enabling them to transform their analytics operations into an AI-powered workflow, according to William McKnight, president of McKnight Consulting.
In addition to Tableau Studio, Proactive Intelligence and Tableau Knowledge, Tableau introduced Data Apps, which extends Studio's application development capabilities to third-party environments such as ChatGPT and Claude.
"Tableau's role is expanding from end-user dashboarding into headless, embedded, and conversational analytics infrastructure," McKnight told TechTarget. "The new capabilities mark a significant addition by empowering non-technical users to build micro-applications through natural language while deploying an autonomous 'always-on analyst' that pushes proactive recommendations into operational tools."
Like Leone, he added that while Tableau is in line with its competition as BI vendors race to build agentic analytics platforms, its integration with Salesforce is one of the ways Tableau is differentiated.
"Tableau is responding to an industry-wide race toward agentic analytics here but brings some uniqueness with its Salesforce ecosystem integration, its installed user base, and allowing external AI models like Claude to safely access governed metrics," McKnight said.
Tableau general manager Mark Recher speaks during the Tableau Keynote at Salesforce's annual Dreamforce user conference.
Context as a critical element
While Tableau's aim is to "see, understand and act on data at the speed of thought," according to Recher, whether the capabilities highlighted at Dreamforce truly enable insight generation at that rate remains to be seen.
When integrated with other agentic AI capabilities from Salesforce, the new Tableau features do seem appropriately designed to help enterprises generate insights and take actions based on their data, according to David Menninger, an analyst at ISG.
"Coupled with other Salesforce agentic capabilities, these Tableau enhancements help enterprises achieve the goal of seeing, understanding and acting on the data," he told TechTarget. "I'll reserve judgment on the speed of thought comment … but for the information that is available, analyses and actions can certainly be accomplished more quickly and more consistently."
Toward that end, Tableau Knowledge is perhaps the most significant of the new features, Menninger continued.
"Context will be the most critical element to ensure that agents are making the right decisions and taking the right actions," he said.
Like Menninger, Leone noted that the features Tableau touted at Dreamforce advance Tableau's aim of enabling users better to see, understand and act on data. But Studio, which he called the most valuable of the new features, and Data Apps could lead to complications, he cautioned.
"The capabilities that Tableau talked about put them in a good spot to deliver on their whole speed-of-thought mission," Leone said. "The bigger issue is that when anyone can build an app in minutes, companies can end up with piles of near duplicates. And then folks are slowing down because they can't tell which app to trust."
Enabling system administrators to certify the data sources used to build applications in Studio and Data Apps could minimize the potential problem, he added.
One early adopter of Tableau's evolving platform is CrowdStrike, a cybersecurity provider based in Austin, Texas.
Tableau's platform has the technology CrowdStrike has used build the dashboards that enable it to oversee its accounts, according to Jacob Schlan, CrowdStrike's vice president of AI, data and analytics. Now, however, CrowdStrike's aim is to provide its employees with a common, trusted data foundation that they can use to not only analyze data but also build assets.
"The trusted data foundation is a non-negotiable," Schlan said. "If you're trying to use raw, uncontextualized data and pointing an LLM or your agent at that, you're never going to be happy with your result."
Using Tableau, CrowdStrike created curated datasets that include a context layer, he continued.
"That's been the true unlock for us in getting the correct analytics," Schlan said. "It's governed data, the same every time, that allows us to continue on the self-service analytics path."
"To succeed as an agentic analytics platform, Tableau must build native logic arbitration and collision detection to manage conflicting actions when multiple autonomous AI agents query its Knowledge Layer," he said. "It could also integrate real-time data quality and observability guards to prevent bad data from causing automated operational failures."
Menninger, meanwhile, advised Tableau to concentrate on improving the knowledge layer that he called the highlight new feature. Agents need knowledge layers, and the more context such capabilities can help deliver, the better agents will perform in production.
"The industry needs a comprehensive knowledge layer if agents are going to succeed broadly," Menninger said. "That requires robust metrics definitions that can be shared and exchanged among all the parts of an enterprise’s information architecture."
Eric Avidon is a senior news writer for Informa TechTarget and a journalist with more than three decades of experience. He covers analytics and data management.
