Agentic AI was the new kid on the block at this year’s Gartner Data and Analytics Summit. The working definition seemed to be—a main dish of GenAI LLMs, with some APIs on the side. “A trenchcoat and 3 APIs” as some of us were joking.
There were a fair amount of GenAI and RAG components built into some of the more substantive Data & Analytics offerings this year. Along with a decent number of newish vendors trying to catch a set of GenAI waves that were a tad haphazard and just forming. The show seemed a little smaller than last year in the number of vendors and a little less starry in outside speakers. I had some thoughtful conversations with colleagues from competitors, analysts, and partners.
Amazon, MSFT Azure, and Google showed GenAI, RAG, AgenticAI, and ModelOps services architectures and components. It was notable that MSFT came as Azure, featuring MSFT Fabric and no stand-alone Power BI offering—as you must now purchase MSFT Fabric to access Power BI.
There wasn’t much open-source dev or San Francisco stuff. So while it was great to catch the vibe…it was also cool to get back to San Francisco, and my world of LangChain, Anthropic, Cohere, Perpexity, OpenAI, Google, NVIDIA, and the others.
Generative AI’s evolution to agentic AI
Rita Sallam at Gartner described how GenAI advances—in learning, reasoning, synthesizing, and adapting to data and situations—have enabled agentic AI systems to create alternative business options, and to propose and take actions by AI agents and humans.
In a data and analytics workflow scenario, Rita described how AI Agents can be used to perform what Gartner is calling “perceptive analytics.” Per the definition of perceptive, these analytics understand things quickly, especially things that are not obvious.
She described how “perceptive analytics” are a next step from descriptive, diagnostic, predictive, and prescriptive analytics. They are “always-on” and can ultimately become autonomous, providing real-time monitoring, actions, and interventions on critical business systems and outcomes.
Rita described how agentic AI for perceptive analytics, is like ML for predictive analytics. She outlined a generic architecture for data management and analytics in Figure 1 below.
Figure 1. Agentic Data and Analytics Hi-level Architecture. Gartner 2025.
Agentic AI is an evolution of RAG (Retrieval Augmented Generation), which extends the capabilities of LLMs (Large Language Models) to an organization’s internal knowledge base to provide relevant, accurate information to the user’s questions at hand, without retraining the LLM.
Agentic AI builds upon RAG (with multi-modal data) by prompting the LLM multiple times, refining each step’s task and output, and calling external systems to make decisions and execute tasks. While I don’t buy into all the hype presented at Gartner D&A, I do see some great agentic AI applications emerging in multi-dimensional problems with lots of data and potential actions.
I couldn’t help but notice how Rita and I looked like Agent 99 and Agent 86, from Get Smart… hence the title of this blog.
Spotfire Copilot and the AI revolution
It was fun to assess our Spotfire AI journey with the ABI, DSML, and cloud hyperscaler vendors at Gartner. Our goal with AI in Spotfire has always been to rapidly extract insights from data, suggest actions, shorten the learning curve for non-Spotfire users, and drive productivity for all Spotfire users.
Our initial AI work included a continuously-running AI-powered engine—the Spotfire® Recommendations Engine—which identifies multivariate relationships within the data and suggests visualizations that capture the insights and click into Spotfire. Spotfire Copilot built upon Spotfire Recommendations early in the GenAI cycle and now features advanced multi-modal RAG functionality. This includes Spotfire visual data science analysis, data interrogation, and interpretation from natural language conversation inside the Spotfire experience.
We are currently introducing an agentic AI component into the Spotfire CopilotTM architecture targeted at the more diverse and complex areas of the Spotfire platform. Recently released, Spotfire Copilot 2.0 is available to download from the Spotfire Community.
I’ll be diving deeper into the Spotfire Copilot in an upcoming blog.
AI, data, and analytics solutions at Gartner D&A Summit 2025
The BI Bakeoff featured Tableau, Oracle Analytics, and Strategy (formerly MicroStrategy). The data were from the Organisation for Economic Co-operation and Development (OECD) about the use of AI in economic and R&D functions with time, spatial, and demographic attributes.
All vendors were overly buzzword compliant and the demo jocks were smart and fun as they stepped through data management, analysis, content creation, and collaboration along with sharing findings and a cool new innovation.
The Tableau “AI Command Center” is a list of dashboards, visualizations, semantic models (data flows), and data sources. A “Tableau Agent” (not an AI Agent) suggests calculations in Tableau data prep. Tableau Pulse is available in Tableau Cloud—it summarizes data from user definitions / filters with insights, metrics and trends; it does not discover insights like Spotfire Recommendations. The new innovation featured Tableau Next that will be available in Tableau Plus, a premium version of Tableau Cloud inside SFDC. It uses the Salesforce LLM Agentforce for creating Tableau applications from within SFDC. It appears to have some similarities to the Spotfire Copilot in the visual constructs. I’ll dig into this in my upcoming blog on the Spotfire Copilot.
I found the Oracle and Strategy analyses had an ok workflow but didn’t really grab my attention. Oracle’s generated voice, sounding like the Gartner presenter Aura Pope, was quite shocking last year, but felt pretty much the same when repeated this year—quirky but not much value.
The other breakout sessions on GenAI from analytics, BI, and data science vendors felt fairly thin. Folks seemed confused about what to say on agentic AI beyond the basic architecture, and the customer references were stretched to sound more substantive. There were some ok deployments like Thoughtspot’s Spotter which has a clean UX on top of a RAG style conversational analysis.
I enjoyed the Qlik fireside chat with Drew Clarke and Brendan Grady, heads of Qlik data management and analytics respectively, and their AI advisor Rumman Chowdhury. I found this conversation more real than other vendors’ claims that they’d already implemented functionality that didn’t ring true.
Stay tuned for more
The Gartner Data and Analytics Summit is a perennial favorite for data management and analytics professionals. It’s helpful to have a home for discussing the many updates in technologies and products, especially now in the age of GenAI.
It was also fun to have some face-to-face meetings with our customers and partners—in energy, manufacturing, finance, and CPGR. Our annual Spotfire Energy Forum has been recently announced, along with Spotfire presence at other data and analytics conferences around the world. The Spotfire Energy Forum is an invite-only event, May 20 in Houston. We will have some updates on our AI in Spotfire offerings on display there.
Great to catch up with the many Spotfire alumni at the Gartner conference— now at Alteryx, Dataiku, Google, Informatica, and other successful companies. Along with the Qlik folks, I met some from Cognos and Business Objects. We all reminisced on when it was just Spotfire—and then Spotfire, Qlik, and Tableau, disrupting the BI reporting space—the dawn of Data Discovery…the OG’s of AI in BI.
It feels like we are in a similar early stage with GenAI, RAGs, and Agentic AI right now. We started Spotfire to rapidly discover insights that drive actions and extreme value, and our mission to Spot the Fire in the data has not changed—we’re just Getting Smart-er, and making it simpler to do more comprehensive data discovery, and faster to take actions that generate extreme value.
Finally, I enjoyed the Exhibit Hall and collecting some socks from the vendors. I liked Snowflake, Cloudera, and Alation’s socks this year. I’m looking forward to catching up with the Spotfire Cult around the world this summer. Until then, good luck spotting the next fire in your data, and generating extreme value in your business.
