AI Doesn’t Need More Data. It Needs Better Context.

Can AI make good predictions even when historical data is weak - if it has strong contextual understanding?

Short answer: in many cases, Yes.

The Big Idea

Traditional predictive models relied heavily on large volumes of historical data. But in the real enterprise world, historical data is often incomplete, fragmented across systems, inconsistent after process changes and irrelevant after disruptions (COVID, geopolitics, demand shocks). So the assumption that history predicts the future often breaks.

When the world changes faster than the data that describes it, the assumption that history predicts the future begins to break.

What AI increasingly needs instead is context. And this is where knowledge graphs come in.

Historical data tells AI what happened. Knowledge graphs help AI understand why it happened. And that “why” is what makes better predictions possible.

For years we have all heard the phrase: “Data is the new oil.” I have said it myself countless times in presentations, client discussions and conversations. And to be fair, it made sense. Enterprises invested heavily in collecting, storing, and organizing data. Data lakes, warehouses, pipelines—entire architectures were built around the belief that if we had enough data, better decisions would follow. But enterprises don’t just run on data. They run on relationships.

Relationships between suppliers and components. Between products and customers. Between contracts, orders, logistics routes, and regions. Relationship intelligence is what allows AI to understand those connections.

But over the past year, I find myself asking a different question. Is data still the oil? Or is context the real king now?

Lot of the work we do today is no longer just about collecting or processing data. It is about engineering context - connecting entities, mapping relationships across systems, and giving AI the business understanding it needs to reason.

And the shift is striking. Many enterprises don’t suffer from lack of data anymore. If anything, they have too much of it- spread across ERP systems, CRM platforms, legacy and countless custom applications. The challenge is not always data availability. It is quality of context. Historical data tells us what happened. Context helps AI understand what could happen next.

For the past few years, organizations have chased data believing it was the path to their north star- better decisions and deeper insight. Today, the real competitive advantage may lie in something deeper - the ability to give AI the right context to reason about business.

Which brings me back to the original question: Is data still the new oil? Or is context the real king now?

My sense is that the future may not be about choosing one over the other. Data fuels AI. But context gives it intelligence.

So if I had to offer one piece of advice to technology leaders deciding where to invest right now, it would be this: The era of data accumulation is giving way to the era of context creation.