7 Takeaways from the Gartner Data & Analytics Summit 2025

Two days. Dozens of sessions. And one resounding theme: data and analytics are no longer support functions; they’ve become the operational core of modern enterprises. At the 2025 Gartner Data & Analytics Summit, conversations centered on deploying GenAI agents in live workflows, redefining governance, and rethinking the role of dashboards. Leaders came in ready to discuss action. 

Here are the top takeaways from Gartner Data & Analytics Summit 2025: 

1. AI Agents Are Moving From Theory to Execution 

Agentic AI came up early and often. Instead of just describing the past, enterprises are now building agents that can interpret, decide, and act in real-time. Sessions dived into what it looks like to embed agents into workflows for tasks like root cause analysis, alerting, and coordination. 

Some early applications of agentic AI showed promise, including how manufacturing teams are using agents to detect anomalies and escalate issues automatically. While results may be different across industries, the focus has shifted from dashboards to real-time action systems that integrate directly into operations. 

Try this: Identify one manual decision process that takes too long, repeats often, and has clear criteria. Pilot an AI agent there with tight feedback loops. 

At the Lumenore booth, many visitors from manufacturing and BFSI were exploring this exact need – mobile-deployable, domain-specific agents that could surface contextual answers in plain language. Embedded governance was often non-negotiable. 

2. Data as a Product Needs Product Thinking 

The Fidelity session made a strong case for applying product thinking to measurement. Not just owning the data, but owning the experience around it. Their team walked through what happens when metrics lack intent: people stop trusting them, or worse, ignore them completely. 

Start treating metrics like user-facing products: designed for a purpose, built for an audience, and evolving with feedback. 

Takeaway: Give your metrics a product owner. Someone accountable for how it’s used, not just how it’s built. 

We’ve seen greater adoption when teams use custom apps or lightweight visualizations for specific roles, rather than universal dashboards for “everyone.” It turns out that a well-scoped data product, even if it’s simple, is better than a complex one that is unused. 

3. GenAI Is Expensive and the Pressure to Show Results Is Growing 

According to Gartner, at least 30% of GenAI projects will be abandoned after proof of concept by the end of 2025, often due to poor data quality, inadequate risk controls, escalating costs, or unclear business value. 

Technical readiness isn’t the problem, strategic clarity is. Several sessions emphasized the importance of narrowing the scope and tying GenAI pilots directly to operational pain points. Projects that started with a clear use case and measurable outcome, like reducing analyst ticket volumes or generating automated summaries were more likely to move beyond POC. 

Practical next step: Find one question your business teams repeatedly ask, then test if a GenAI assistant can reliably answer it. 

We had several conversations about exactly this. Healthcare and retail teams were curious how our GenAI summaries could help non-technical users get quick, governed answers, without waiting for analysts or risking missteps. 

4. Ecosystem Thinking Is Replacing Platform Loyalty 

The days of going all-in on a single BI or data platform are over. Nearly every session on architecture emphasized interoperability: lakehouses that connect to external tools, data fabrics that stretch across business units, and federated governance models. 

What this means: Your data strategy doesn’t need a rip-and-replace. Instead, look for tools that plug into what you already have – cloud or on-prem, and scale according to your needs. 

That’s something we heard at the Lumenore booth too. Attendees were interested in solutions like Ask Me or Data Magnet that could be embedded individually into their existing environments, without a whole redo. 

5. Governance Is No Longer Optional, It’s Foundational 

Gartner’s message was consistent: you can’t scale AI without scaling trust. The focus has shifted to actively designing governance into data experiences through lineage, alerts, and explainability. 

To act on this: Ask yourself, can a business user trace a number back to its source? Can they see when and how it was last updated? 

If governance only lives in IT, it’s not working. Platforms that embed governance into the user experience, like showing trust scores or lineage paths are making a stronger impact than backend-only controls. It’s something we’ve prioritized at Lumenore with usage monitoring and agent-level access controls. 

6. Storytelling Must Evolve Beyond Dashboards 

Dashboards aren’t going away, but they’re no longer enough. Many businesses still rely heavily on analysts to translate dashboards and extract insights. That translation layer slows insights, decisions, and everything else. 

Next step: Look at your last few reports. If someone had to explain them before the action happened, it’s time to rethink this. 

Tools that layer narratives, summaries, or guided exploration on top of charts are gaining traction. We saw strong interest in Lumenore’s “Ask Me” for its natural language outputs and its ability to reduce dependence on analysts for recurring questions. 

7. Industry Use Cases Drive the Most Urgency 

Across sessions and booth visits, we noticed interests accelerate when the use case is rooted in real-world, domain-specific problems. Retail leaders asked about demand forecasting. Healthcare execs discussed wait-time reduction. BFSI teams wanted intelligent alerting around operational risk. 

If you’re prioritizing data and analytics projects: Anchor each initiative in a business outcome or pinpoint, and ideally, an industry-specific one. Avoid generic dashboards. Prioritize domain understanding. 

At the Lumenore booth, we showed sample use cases for almost every industry. Visitors were appreciative of how Lumenore’s AI-powered visual, conversational tools could be tailored to their organizational needs and the ability to interact with data in the way their teams work. 

Final Thoughts 

This year’s summit reinforced that analytics is no longer an internal toolset, it’s an operational layer. Leaders aren’t just asking for insights. They want systems that think, act, and guide decisions. 

People from around the world, across industries spoke about the current pressure to operationalize GenAI, embed trust, get answers and prove value fast. It was honest, practical, and refreshingly executable.  

A big thank you to Gartner! We’re so grateful to have been part of this conversation.