Your AI is only as smart as the data fabric it sits on

AI has moved from lab experiments to daily operations faster than most of us expected. By the end of 2025, half of all companies were running AI in at least three business functions—finance, supply chains, HR, customer ops, you name it. Copilots, agents, predictive systems: they’re everywhere.

But here’s the thing nobody warns you about during the AI hype. The biggest headache isn’t model accuracy or GPU shortages. It’s the quality and context of the data feeding those systems. AI introduces a new requirement that most legacy data setups just weren’t built for: the system doesn’t just need to access data—it needs to understand what that data actually means in the context of the business.

Irfan Khan, president and chief product officer of SAP Data & Analytics, puts it bluntly: “AI is incredibly good at producing results. It moves fast, but without context it can’t exercise good judgment, and good judgment is what creates a return on investment for the business. Speed without judgment doesn’t help. It can actually hurt us.”

He’s right. And I’ve seen this play out myself—teams get excited about a model that spits out answers in milliseconds, only to realize those answers are technically correct but operationally useless. Sometimes worse than useless, because people trust the speed and act on bad decisions.

The context premium

Traditional data strategies have been about aggregation for the last two decades. Extract from operational systems, load into warehouses and lakes, build dashboards. That works fine for reporting. But along the way, most of the meaning attached to that data—policies, processes, real-world relationships—gets stripped away.

Consider two companies both using AI to manage supply-chain disruptions. One feeds in raw signals: inventory levels, lead times, supplier scores. The other adds context across business processes, policies, and metadata. Both systems analyze data quickly. But they’ll reach very different conclusions.

The second system knows which customers are strategic accounts, what tradeoffs are acceptable during shortages, and the status of extended supply chains. The first one just sees numbers. Khan calls this the “context premium”—the advantage you gain when your data foundation preserves context by design.

“Both systems move very quickly, but only one moves in the right direction,” he says.

This matters more now than it did five years ago because AI doesn’t just display information anymore—it acts on it. Agents make decisions, execute trades, reroute shipments. If the system doesn’t understand why something matters, it optimizes for the wrong outcome. Inventory numbers might be accurate, but they don’t tell you which customer must be prioritized or which contractual obligations apply.

Don’t consolidate, integrate

So what’s the fix? It’s not dumping everything into a bigger data lake. That’s actually part of the problem. The emerging solution is a data fabric—an abstraction layer that spans infrastructure, architecture, and logical organization without forcing everything into one place.

For agentic AI, this fabric becomes the primary interface. Agents interact with business knowledge rather than raw storage systems. Knowledge graphs play a central role here, letting agents query enterprise data using natural language and logical relationships instead of writing SQL joins against tables that don’t carry meaning.

The key insight is that you don’t need to consolidate data into a single repository. You need to connect information across applications, clouds, and operational systems while preserving the semantics that describe how the business actually works. That’s a fundamentally different approach from what most companies have been doing.

The maturity gap

This realization is forcing companies to rethink their AI readiness. And the numbers aren’t pretty. Only one in five organizations consider their approach to data to be highly mature. Only 9% feel fully prepared to integrate and interoperate with their data systems.

That’s a scary gap when you consider how fast AI is being deployed. We’re essentially building autonomous systems on foundations that most companies themselves admit aren’t ready. It’s like putting a Formula 1 engine in a car with bald tires and no steering wheel.

Khan and others at SAP are pushing data fabric as the answer, and honestly, it makes sense. But it’s not a product you buy off the shelf—it’s an architectural approach. It requires thinking about data differently, investing in metadata management, and building the context layer before you scale AI.

Most companies won’t do that. They’ll rush to deploy agents, hit the context wall, and wonder why their AI investments aren’t delivering. The ones that get it right will be the ones that understand that speed without judgment is just fast stupidity.