The 2025 Gartner Summit confirmed what many of us already feel—AI is everywhere. But while artificial intelligence dominated the agenda, a quieter but more urgent message came through: most organizations still lack the foundational data capabilities needed for meaningful AI success.

At UD4D, we attended with a clear purpose: to look beyond the buzzwords and extract insights that truly matter to companies today. Here's our take on what you need to know to cut through the noise and focus on what will move your organization forward.

‍

1. AI is About to Influence Half of Business Decisions

By 2027, 50% of business decisions will be augmented or automated by AI agents.

This bold forecast sets a clear direction. But while the technology is advancing rapidly, it also presents a major leadership challenge. If AI is to transform how we make decisions, top management must step up and learn to work with AI—not just from a technology standpoint, but as the driving force behind adoption and cultural transformation.

But here's the reality: Only 4% of companies currently say their data is fully “AI-ready”. And 55% openly admit they're struggling to get there. The desire exists—29% of organizations are already investing in AI-ready initiatives—but most are running into the same roadblocks: poor data quality and governance, incomplete or fragmented data integration, a lack of clear metadata strategy, and siloed, outdated infrastructure.

‍

2. The Rise of Smaller, Specialised AI Models

By 2027, organizations will use task-specific AI models three times more often than general-purpose ones.

This marks a major shift. While general-purpose models like ChatGPT grab the headlines, Gartner's data points to a more pragmatic future—focused, efficient, domain-specific AI that solves real business problems. These models are easier to train, faster to deploy, and more directly tied to operational outcomes. That’s where democratized innovation will come from.

We believe these smaller models will play a crucial role in unlocking practical AI adoption, especially in industries where accuracy and domain context matter more than flashy tech.

‍

3. Metadata Is Everything

By 2027, organizations that prioritize AI-ready data will increase GenAI model accuracy by up to 80% and reduce costs by 60%.

From our experience, many companies still treat metadata as a documentation layer—something nice to have, not strategic. But to succeed with AI, metadata must become a foundational element of your data platform. It enables trust, context, lineage, and automation.

One major takeaway? Metadata is no longer optional. Gartner forecasts that organizations prioritizing semantics and metadata will dramatically improve AI model accuracy and cut costs in the process. No wonder the word "metadata" was mentioned more than 35 times in the opening keynote alone.

“No metadata, no AI.” It's that simple.

‍

4. Natural Language Interfaces Are the Future

By 2026, natural language will become the dominant interface for data querying, improving data consumption tenfold.

The promise of natural language as the dominant way to query data is exciting—finally, more people can engage with data without being technical experts. But there's a catch: the biggest barrier to progress isn’t the technology—it’s culture.

We’ve seen both ends of the spectrum. In some organizations, such as banks with strong finance leadership, a “no data, no discussion” mindset drove transformation quickly. But many companies still operate on gut feelings and instincts, and enforcement is weak. As a result, data is often debated instead of trusted.

According to Gartner, 70% of companies regularly argue over which data is “correct”, often without resolution. Whoever has the louder voice, not the better data, wins the argument. Until this changes, even the best AI tools won’t fix poor decision-making.

‍

Our Take: Start With Trust, Not Perfection

The Gartner Summit confirmed what we see in practice: AI is as much a data challenge as it is a cultural one. The tools are evolving rapidly, but without trustworthy data and leadership that believes in it, transformation will stall.

So, if your organization is serious about AI, don’t wait for a perfect data landscape. Instead:

• Prioritize the most critical areas
• Focus on one use case at a time
• Define clear quality requirements
• Implement context-aware trust models
• Iterate, measure, and scale gradually

In the end, the question isn’t whether AI will transform your business, but whether your data, your culture, and your leadership are ready to lead that transformation.

‍