Reflecting on the Gartner Data & Analytics Summit: Key Takeaways

As we mentioned in our post-show recap, the Rocket team came away from the Gartner Data & Analytics (D&A) Summit in Orlando with great insights and ideas from the many sessions and discussions. For those who couldn’t attend, the Summit brought together IT business, and data leaders from across industries to discuss, learn, and share insights around artificial intelligence (AI), data management, architecture, governance, and more. The attendees’ top-of-mind concerns  included how to use AI to its fullest potential, how data architecture can help, and what data professionals should do to prepare for the future. 

With the team experiencing the event from varied perspectives, we regrouped afterward and shared our individual takeaways. As it turned out, there were two major themes: The future of AI and data architecture. To help guide you on these important topics, I want to share our takeaways and what they mean for your roles, now and in the future. 

AI: The most used (and still the most confusing) 2025 acronym  

The bottom line is that companies are still working to figure out what their AI initiatives should be. The technology is evolving so rapidly that it’s hard to put a stake in the ground. (a recent study said AI can get “anxious”. Wait, WHAT?) But it’s clear that AI is the future of data management and corporate knowledge. The key is to fully define what you want from applying AI to your data collection and interpretation initiatives before you start.  

Here are our thoughts: 

• The level of investment in the data space, now driven by AI, continues to accelerate. But the business cases and results from AI initiatives are still evolving. While it’s critical to tie business outcome to AI investments, the untapped data trapped in the mainframes remain a major obstacle to building that strong business case.
• There was an interesting point made in one of the sessions: "Only one in five organizations have managed to use their data effectively for AI use cases.” With the predominance of AI-focused slogans, and companies claiming to have a good handle on anything AI-related, this quote struck a chord. It doesn’t quite fall in line — and we couldn’t agree more! The AI boom is not as simple as some people might think it is.  For instance, we know that the vast majority of mainframe data is still not being leveraged (72% according to our research). Leaving this critical data out of analytics and AI initiatives means that companies are making decisions about their futures with incomplete or even inaccurate data.
• Data management, governance, and change management are essential. They should be a part of every AI project’s cost consideration. But companies should invest in the foundation first to get a handle on managing and integrating their data. AI governance should mirror the existing data governance framework, not just focus on surface level practices, and ask the question: “What could go wrong?” Metadata is the key. It’s the one component that spans all data types and formats. But to work, the core metadata needs to be clean and accurate, well organized, and easy for anyone in the organization to use.
• AI is everywhere. It was mentioned at almost every exhibitor booth and in almost every session. While there are some clearer AI use cases than even a year ago, the exact benefits remain ambiguous. We didn’t hear anyone say "10x cost savings with AI for xyz use case." Attendees were looking for concrete answers to their specific problems. "Ready your data for AI" messaging was everywhere. But what does that mean? The answers were almost always vendor-specific. An overriding definition and thought leadership are needed to help companies understand what it means for them.
• Data security and how to safely use enterprise content with GenAI was a major concern among attendees. Guidance about how to accomplish that security resonated with them, especially around mainframe data. Data and IT professionals understand that mainframes hold vast amounts of mission-critical data and that leveraging it is too often a roadblock they don’t know how to get past. Many don’t even know if they can get past it.  
   

DATA MESH vs FABRIC vs WAREHOUSE vs DATALAKE vs … 

Data architecture was another huge area of focus, and for good reason. There are so many options and the most common response to “Which should I use?” was, “It depends.” We agree on this point. There’s no one perfect answer. So, working with strong partners who understand all the options and possible combinations — and is willing to deeply understand your situation and goals — is the most efficient and effective way to move forward. The alternative, an “earn while you learn” approach, is simply untenable.  

Here’s what we took away:  

• It’s interesting to see that data fabric has gone from a somewhat vague concept last year to a tangible strategy being implemented at large enterprises. The kicker from sessions on this topic is that no two strategies will look alike. There will be amalgamations of lakehouse, mesh, fabric, warehouse, and more, depending on the situation and need. And it’s not something organizations can just drop into place. They have to build and adjust as they go, so the results may end up a completely different shape than what they envisioned. All this underscores the importance of good metadata across the organization. Metadata-driven integration is the only enterprise integration approach that can support all these variations.
• The data mesh and data fabric concepts have been out for a few years now. Gartner was clear on how they differentiate: Data mesh decentralizes data and data fabric centrally controls it. Yet, they were still being used interchangeably by attendees. Maybe the confusion stems from the fact that neither is a “product” companies can buy. And, to the point above, their application is highly variable and situation-specific. Fabric is definitely gaining traction, but attendees wanted to know what’s best for their businesses in their specific industries. “If I’m in XYZ industry and I use Databricks, what data fabric strategy should I use?” It’s not that cut and dried.
• Lakehouses continue to be the growing trend. They combine the benefits of data lakes (cost-effective storage of raw data) and data warehouses (structured data management and governance) into a single platform. Since both structured and unstructured data support analytics, lakehouses that combine the two provide the best of both worlds.
• And, of course, effectively tapping into mainframe data is still a problem. My session on Harnessing Mainframe Data for AI-driven Enterprise Analytics seemed to drive a lot of traffic to the Rocket booth. It was well-attended, resonated with the audience, and we talked to several people afterward about the untapped data trapped in their mainframes that decision makers aren’t leveraging. That revelation definitely caught the attention of the data, analytics, and BI folks. It was surprising how few attendees knew that it was even possible! As we said, mainframes hold vast amounts of mission-critical data and the owners of those systems zealously protect it. Looking at this data from a business-value perspective is key. Data, AI, and analytics professionals need to build the business case from the end-goals back to optimize data’s tremendous impact on business strategy, direction, and outcomes — in other words, its tremendous impact on your bottom line and your company’s future. 

A final note 

As data and analytics professionals, your company’s future is literally in your hands. Providing leadership with complete and trusted data will enable the best strategic decisions. As you move forward with initiatives, it’s important to fully understand where you are today, where you want to go, and map out a manageable strategy to get there. This means having clearly defined goals, data management and architecture roadmaps, and the strategic partners you need to be successful. You can design your future with certainty. 

Learn more about leveraging AI and optimizing your entire data environment for the future.