Big Data London Reflections: Why I Left Thinking Data Governance Is Becoming a Trap

After spending time at Big Data London, attending presentations, speaking with vendors, and listening to countless discussions around data platforms, catalogs, governance frameworks, lineage, and AI, I left with a growing conviction: our industry is investing an extraordinary amount of energy solving yesterday’s problem.

The dominant message across the exhibition floor was familiar. Better governance. Better catalogs. Better metadata. Better lineage. More stewardship. More policies. More controls. All with a sprinkle of AI.

Almost every vendor, regardless of how they positioned themselves, seemed to end up telling a variation of the same story: if only organisations governed their data better, everything else would follow.

Yet the more conversations I had, the more I felt there was a disconnect between what the market is selling and what organizations actually need.

As enterprises are increasingly driven by AI, I am not convinced that better data governance is the answer.

I believe what we need instead is something different.

The Industry Is Mistaking the Means for the End

Let me be clear: governance matters.

Organisations need to know where data comes from, who owns it, how it is secured, and whether it complies with regulatory requirements. Data quality, lineage, stewardship, and ownership are all important capabilities.

But somewhere along the way, governance stopped being a means to an end and became the end itself.

Many organisations now measure the success of their data strategy by the number of assets cataloged, the percentage of metadata completed, the amount of lineage documented, or the number of governance processes implemented.

The problem is that none of these things, by themselves, create business value.

A company does not become more competitive because it has documented 10,000 datasets.

A CFO does not make better decisions because a glossary contains 500 approved business definitions.

An operational manager does not improve performance because data ownership has been assigned across the organization.

These things can help. They create order. They create consistency. They reduce risk.

But they do not automatically improve decisions.

And decisions are ultimately what matter.

AI Is Changing the Nature of the Problem

What struck me most at Big Data London was how frequently AI appeared in conversations, yet how often it was treated as an extension of existing governance thinking.

The typical narrative looked something like this:

“AI needs trusted data. Therefore we need more governance.”

While there is truth in that statement, I increasingly believe it misses the bigger picture.

The real challenge presented by AI is not simply that AI needs governed data.

The challenge is that AI dramatically accelerates the pace of decision-making.

AI creates a world where information is consumed continuously, recommendations are generated in real time, and decisions can increasingly be automated.

In this environment, the key question is no longer:

“Do we know this data asset?”

The key question becomes:

“Do we understand the business consequences of acting upon it?”

That is a fundamentally different problem.

And governance alone does not solve it.

The Catalog Is Not the Cockpit

As I walked around the exhibition floor, I saw countless impressive demonstrations.

• Beautiful catalogs.
• Lineage visualizations.
• Data maps.
• Business glossaries.
• Metadata repositories.
• Compliance dashboards.

The technology has become remarkably sophisticated.

But I kept asking myself the same question:

If I were a CEO, COO, plant manager, supply chain director, or customer service leader, would this actually help me operate my business tomorrow morning?

The answer was no.

Most solutions explain the organisation.
None help steer it.

There is a profound difference between documenting a system and navigating a system.

An aircraft manual contains detailed information about every component of the aircraft. A cockpit helps a pilot fly it.

Many governance platforms are becoming excellent manuals.

What organizations increasingly need are cockpits.

We Have Become Obsessed with Data and Forgotten Decisions

One of the biggest paradoxes in modern data management is that many organizations know more about their data than they do about their decision-making processes.

They can tell you:

• Which applications exist
• Which databases are used
• Which tables are connected
• Which reports consume which datasets

Yet they often struggle to answer:

• Which decisions actually drive performance?
• Which processes create value?
• Which operational risks matter most?
• Which actions should be taken when a KPI deteriorates?
• Which systems have the highest business impact?

The center of gravity has shifted toward documenting assets rather than understanding operations.

AI is exposing this weakness.

AI does not just need structured data.

• It needs context.
• It needs understanding.
• It needs to know not only what a customer record is, but why it matters.

Not only where a metric comes from, but what actions should follow from it.

Not only how data flows through systems, but how decisions flow through the business.

Governance Is About Control. Piloting Is About Action.

This realization led me to what I increasingly believe is the next evolution of the discipline.

Data Governance focuses on control.

Data Piloting focuses on action.

Governance asks:

• What data do we have?
• Who owns it?
• Is it compliant?
• Is lineage documented?
• Is quality acceptable?

These are valuable questions.

But Data Piloting asks different questions:

• What decisions depend on this information?
• What business outcome is at risk?
• What is changing?
• What should we do next?
• What happens if we modify this process?
• Where should management focus attention?

Governance manages information assets.
Piloting manages business outcomes.

The distinction may seem subtle, but it is transformational.

AI Requires a Digital Representation of the Enterprise

The winners of the AI era will not simply have the best catalog of data assets.

They will have the best understanding of how their organisations actually function.

To make intelligent recommendations, AI must be able to connect:

• Data
• Processes
• Systems
• Controls
• Risks
• Decisions
• Objectives
• Outcomes

In other words, it must understand the enterprise as a living system.

• A catalog alone cannot provide that understanding.
• Lineage alone cannot provide that understanding.
• Metadata alone cannot provide that understanding.

These are important building blocks, but they are not the destination.

The destination is a dynamic model of the business that allows both humans and AI to understand how things work, why they work, and what happens when they change.

That is the foundation of Data Piloting.

From Governance Platforms to Piloting Platforms

As I reflected on the conversations at Big Data London, I began to wonder whether the market is approaching an inflection point.

Today, vendors compete on:

• Catalog capabilities
• Metadata management
• Data quality functionality
• Stewardship workflows
• Lineage depth
• Governance automation

Tomorrow, I suspect organisations will care more about:

• Impact analysis
• Process intelligence
• Decision intelligence
• Business observability
• Scenario simulation
• AI-assisted recommendations
• Operational navigation

The question will no longer be:

“Can you document my business?”

The question will become:

“Can you help me run my business?”

That is a very different market. And it requires a very different philosophy.

Governance Will Not Disappear

None of this means governance becomes irrelevant.

Quite the opposite.

Governance remains essential.

Security remains essential.

Compliance remains essential.

Data quality remains essential.

Ownership remains essential.

But governance is becoming infrastructure.

It is the legacy foundation upon which something more valuable is built.

Just as nobody buys an aircraft because it has a strong frame, organisations should not invest in data programs simply to achieve governance maturity.

• The goal is not governance.
• The goal is performance.
• The goal is agility.
• The goal is better decisions.
• The goal is better outcomes.

Governance supports those goals.

It is not the goal itself.

The Future Is Data Piloting

My biggest takeaway from Big Data London is therefore not about technology.

It is about mindset.

For years, the industry has focused on helping organisations manage data.

In the age of AI, the real opportunity is helping organisations navigate complexity.

The leaders of tomorrow will not be the companies with the most complete catalogs, the most detailed glossaries, or the largest stewardship organisations.

They will be the companies that can understand their business in real time, anticipate impacts before they occur, evaluate scenarios instantly, and guide both humans and AI toward better decisions.

That is not governance.

That is piloting.

And I increasingly believe that Data Piloting, not Data Governance, will become the defining discipline of the AI era.