The missing layer in operational decision-making

First published on: 10/27/2026

By Andrea Stone, CEO at Zema Global

I have spent much of my career around businesses where information has a very short distance to travel before it becomes a decision.

That is particularly true in energy and commodities. A change in a dataset can flow into a curve, a valuation, a risk position or a commercial decision. The people closest to these processes understand those connections through a combination of experience and instinct. They know which signals matter, which dependencies to check and when something that looks technical could have a much wider business consequence.

Over the years, the industry has become considerably better at detecting those signals. We can monitor more data, processes and applications, and identify anomalies and failures earlier.

One thing has become increasingly clear to me: seeing a problem earlier does not necessarily mean understanding it sooner, anticipating what happens next or resolving it autonomously.

That gap is where operational intelligence becomes important.

Better visibility hasn’t solved the context problem

When something goes wrong in a complex operational environment, technology can usually tell us where the signal appeared.

The harder part comes afterwards.

“What relies on that data? Has another process already consumed it? Is a downstream calculation affected? Is this an isolated problem or part of something larger?”

I have seen some of the most important parts of that understanding reside with a remarkably small number of people.

They know the connections because they have worked with the systems for years. They know that one apparently minor feed is critical at a certain point in the day, or that several separate alerts usually trace back to the same dependency.

We tend to describe that as experience. I have come to think of it differently.

It is operational context, and operational context is a form of institutional knowledge.

If that context exists mainly in people's heads, the organization has knowledge that its technology cannot use. Every investigation requires some of it to be reconstructed manually. At increasing scale and speed, that becomes a constraint.

Operational intelligence connects the dots

Monitoring detects an event. Observability makes the health and behavior of systems, data and processes more visible. Both are essential.

Operational intelligence combines that visibility with the context needed to understand significance, dependencies and potential impact.

It connects what is happening now with what sits upstream and downstream. Lineage, dependency awareness and historical patterns help teams distinguish an isolated alert from a wider issue and determine where attention is needed first.

It can also become predictive. Rather than waiting for a failure and reconstructing what happened, organizations can identify patterns that suggest a problem is emerging and understand what could be affected before disruption spreads.

As systems learn from those patterns, the opportunity extends further: from detecting and understanding issues to anticipating them and, where appropriate, resolving them autonomously.

Every operational issue carries a potential business cost. The longer it takes to understand what is happening and what is affected, the greater the risk of disruption. Operational intelligence shortens that gap, helping organizations protect time, productivity and ultimately financial performance.

Context needs to become part of the infrastructure

For many years, organizations could rely on experienced people to supply the context that systems lacked.

That becomes harder as the number of datasets, external providers, models, applications and dependencies grows. Energy and commodity businesses are particularly exposed because so much decision-making relies on interconnected internal, proprietary and third-party information.

Through our work at Zema Global, I have come to see context as something we need to treat with the same discipline as data itself.

We need to understand where information came from, where it goes, what depends on it and what may be affected when something changes. We also need to preserve more of the relationships that experienced people carry in their heads and make that knowledge available across the organization.

Context is becoming infrastructure.

For me, that is the foundation of operational intelligence: a connected understanding of what is happening and what it means.

AI depends on context too

There is a great deal of focus on what AI can do with operational information. I am more interested in what it needs to know before we rely on what it tells us.

An isolated alert gives AI something to describe. Trusted data, lineage, dependencies and operational history give it something to reason over.

That creates more useful possibilities. Patterns from previous incidents can inform the next one. Emerging issues can be identified earlier. Repetitive responses can increasingly be automated.

The question for leaders is not simply where AI can be applied. It is whether the organization has enough connected operational context to make its application useful.

A different measure of operational maturity

There is a simple way to see where that operational knowledge still sits.

“Imagine one of the people who understands your most critical data flows better than anyone else is unavailable tomorrow. What becomes harder?”

If investigations take longer, dependencies become less obvious or teams struggle to judge which issues matter most, you have identified knowledge that is important to the operation but has not yet become part of its infrastructure.

This is not about removing people from the equation. It is about allowing their expertise to travel further through the organization.

I think that will become an increasingly important measure of operational maturity: how quickly an organization can turn an operational signal into an understanding of business impact.

Where we are taking this at Zema Global

This thinking has shaped how we have approached Zema Sentinel, our operational intelligence platform.

We did not set out simply to create another monitoring tool or another stream of alerts. Zema Sentinel brings together monitoring, lineage, dependency awareness, root-cause analysis and predictive detection to help teams understand what is happening, how it connects and what may be at risk.

As Sentinel learns from operational patterns over time, the opportunity is to anticipate issues earlier and enable more autonomous responses.

We see this as part of Zema Global's broader Decisioning Infrastructure. Confidence in a business decision depends not only on the data and analytics in front of someone, but also on the data and processes behind them.

For years, the industry has worked to improve visibility.

Now we are turning that visibility into understanding, and that understanding into action.