Summary
Highlights
A quick recap confirms that conversational analytics and the knowledge graph are coming to cloud and server soon. M.K. introduces a 'vision demo' for an Agentic Analytics Command Center. This central hub would allow users to monitor the performance of all their AI agents (conversational, metric, alerting, data pipeline), assessing efficacy, adoption, trust, and data integrity. It would also provide 'to-do' recommendations, suggesting actions to improve agent performance or engagement, with a detailed lineage view of each agent's components.
The speakers encourage attendees to explore sessions, learn about AI, new features, and connect with the community. Mark Recker thanks everyone and announces that the Tableau Conference will return to San Diego in 2027.
The conference welcomes attendees to San Diego and celebrates the global Tableau community, the Datafam. Special recognition is given to visionaries, ambassadors, and user group leaders for their contributions and passion for sharing data skills. The speaker highlights the growth of the community, from 187 attendees at the first event to thousands globally, emphasizing its role in helping people understand and act on data.
The speaker acknowledges the community's influence on Tableau's product development, citing examples like Tableau Public and Tableau Free Desktop, which has seen 100,000 downloads since March. Features like rounded corners, custom color palettes, and new map functionalities are also attributed to user feedback. The discussion shifts to the growing importance of AI, noting the community's leadership in leveraging AI with Tableau and emphasizing that human trust in data remains paramount.
Mark Recker, the new GM for Tableau, expresses his excitement about the role and the conference. He shares his personal journey and passion for data, highlighting his belief that understanding data leads to better decisions. He introduces 'Agentic Analytics' as the next era of data, where insights will be consumed by both humans and AI agents. He emphasizes the need for 'knowledge'—trusted data and semantic models—to ensure AI accuracy.
Mark outlines four key aspects of Tableau's agentic analytics platform: a knowledge engine for transforming data into semantic models using natural language, a knowledge graph to understand an enterprise's entire knowledge base, a decisions engine for real-time insights and actions by humans and agents, and the ability to take trusted knowledge anywhere through headless analytics (e.g., Claude, ChatGPT, Slack, Microsoft Teams). He also stresses that the platform is secure, governed, composable, and extensible.
Mark discusses how the Datafam's role is evolving into 'agentic architects' or 'knowledge architects,' driving greater impact by composing their company's analytics and data strategy. He announces that the 'tell' part is over and the 'show' part will begin, demonstrating how Tableau will help architect knowledge, power decisions, and identify actions across all product experiences.
Chad Stroup, VP of Global Retail Analytics at Disney, joins the stage to discuss Disney's vision for moving from reporting to a decision layer. He highlights the need for Tableau to connect more directly to systems where decisions are made, enabling continuous real-time loops. Chad expresses excitement about agentic and conversational AI, seeing Tableau and AI as a 'personal data scientist' to answer follow-up questions beyond initial dashboards.
Michael Andrew, Chief Data Officer at Salesforce, shares his long history with Tableau and how Salesforce now runs its business on it. He demonstrates new Tableau features using an imaginary company, 'Bolt Bikes.' Key highlights include 'composable data sources' for joining data from different databases and 'agentic analytics' in Tableau Cloud and Server, allowing users to ask natural language questions for immediate insights. He introduces the 'analytical knowledge graph' that automatically understands data semantics, enabling trusted AI responses, and new AI-powered tools for building semantic models faster.
The demo extends to showing how agentic analytics integrates into Slack. Users can ask questions directly in Slack, and Tableau's trusted knowledge graph provides accurate answers and recommendations, bringing insights into the flow of work. The recap emphasizes composable data sources, conversational analytics in various platforms, the automatic analytics knowledge graph, and Slack integration, all grounded in trusted data with 95% accuracy.
A video presentation features Engine, a modern travel and spend management platform, discussing how Tableau and Salesforce are crucial to their operations. They highlight how Agentic Analytics and Tableau have unified their data, allowing them to leverage natural language to build semantic models and analyze data. The integration with Slack enables quick access to insights and faster iteration, calling it a 'revolutionary' shift from dumb to smart software.
Rekha discusses the challenges of decision-making in an era of abundant AI tools and reports, where different tools often yield inconsistent answers. She asserts that Tableau's core strength is its ability to build models based on trusted data, enabling composability and extensibility. This allows Tableau to power every surface, agent, and workflow with accurate, trusted knowledge, significantly boosting productivity and reimagining decision-making.
Rekha introduces the new Datafam AI & Analytics Showcase, a platform for celebrating community projects and breakthroughs. She then welcomes Will Sutton, a Tableau visionary and consultant, to the stage. Will shares his journey with Tableau since 2013, noting the rapid changes with AI like ChatGPT, but emphasizing the Datafam community remains his grounding force. He discusses how he built the Tableau MCP (Microsoft Co-Pilot) on LinkedIn, driven by curiosity to connect Tableau to AI.
Will Sutton demonstrates new AI capabilities. He shows how to use natural language (via Claude) to convert a sketch of a complex viz into a Tableau viz extension, highlighting significant time savings. He then demonstrates publishing this viz to a Tableau portal for retail partners using Claude to generate embed code. Finally, he showcases conversational analytics both in Claude (connecting to Tableau MCP for distributor queries) and with a voice assistant on the go, emphasizing the ability to get immediate, accurate insights anywhere.
M.K., President and CTO for Engineering at Salesforce, emphasizes Tableau's consistent role in guiding users through evolving analytical eras. He demonstrates 'identifying actions' in Tableau Nexo, starting with combining sales forecasts and warehouse inventory into one operational view. Crucially, he shows how unstructured data (like analyst reports) can now be integrated, allowing AI to perform root cause analysis. The system then enables direct action, like moving inventory, transforming passive dashboards into active operational systems. He highlights the power of 'auto mode' allowing agents to take automated actions based on semantic models.