Seeing the Business Before It Breaks.

Finding the Signals You Did Not Know Were There.

When my wife’s car got stuck in that gym carpark, the problem was not that something had failed, but by the time we knew enough to act, the disruption had already arrived.

That is the part leaders will recognise.

Not because cars and businesses are the same. They are not. But the feeling is familiar. Something has stopped working as intended. The thing that should have been moving now cannot move. The rhythm changes. Time gets lost. Money gets spent. People have to work around the problem.

And later, once the immediate issue is being dealt with, the question starts to form.

Was there a signal we missed?

That is often how it feels in business too. A leadership team sits around a table with the monthly pack open. The numbers are mostly green. The dashboard says things are stable. Actions are closing. Incidents are within normal range. Production may have had a rough week, but nothing looks too alarming. Maintenance is busy, but maintenance is always busy. Overtime is up, but there is an explanation.

There is always an explanation.

Still, someone around the table has that feeling that something is not quite right. Not enough to call a crisis. Not enough to stop the meeting. Not enough to point to one number and say, “There it is.” Just enough to know the business may be trying to say something the report is not quite saying.

That is where traditional reporting often reaches its limit. Reports are useful. Dashboards are useful. KPIs are useful. They help leaders see selected information in a structured way. When they are designed well, they create a common picture and support better conversations. But most reports answer questions we already knew how to ask.

How many incidents did we have? How many actions are overdue? How many inspections were completed? What was production output? How much downtime did we record? How much overtime did we use? Which sites are red, amber or green? Those questions matter, but they are not always enough. The more difficult question is what is happening between the numbers, because that is where the real operating picture often lives.

It might sit in repeated comments no one has time to read. It might sit in weak action close-out notes. It might sit in the gap between a maintenance request, a competency record, a production delay and a near miss. It might sit in the difference between what was reported, what was investigated, what action was taken and whether the same issue appeared again. That is not a simple reporting problem. It is a line-of-sight problem.

Most organisations are not short of data. They are short of usable visibility. They have incidents, hazards, inspections, audits, observations, actions, investigations, maintenance requests, production data, rosters, overtime records, competency records, contractor activity, downtime, defects, customer issues, rework, delays, comments, photos, attachments, close-out notes, classifications, dates, owners and overdue items.

They have people doing work every day that creates evidence about how the business is really operating. But having the information and being able to use it are not the same thing.

That is why I think there is now a more important question for leaders to ask.

Are you using the data you can see, searching for the data you think you have, or finding signals in the data you did not know was already there?

Most organisations are still working somewhere between the first two. They are using the data they can see: reports, dashboards, scorecards, KPIs, structured fields, monthly summaries, graphs and trend lines. That work is important because it gives leaders a starting point.

They are also searching for the data they think they have. That is the world of better reporting, better workflow design, better data quality, better system integration and better access to information that is already being captured somewhere in the organisation. That work matters too.

But the third question is where things start to change.

Can we find signals in the data we did not know was already there? That is the shift AI now makes possible. Not because AI is magic, and not because it removes the need for human judgement. It matters because AI can review far more operational information than humans can practically review on their own, at a speed and depth that traditional reporting has never really been able to achieve.

It can read across large volumes of information. It can look for repeated themes. It can compare records. It can surface contradictions. It can flag missing context. It can notice where actions do not appear to match causes. It can help connect signals across workflows that have traditionally been managed separately.

The important signals are not always sitting neatly in a KPI. They may be hiding in the relationship between different parts of the business.

Imagine production has dropped, so overtime increases. Some of the overtime sits on shifts where key competencies are thin. A machine is operated incorrectly, maintenance issues increase, downtime follows, and production drops again. More overtime is used to catch up. Fatigue builds, mistakes become more likely, and then incidents, defects or rework start to appear. Each part of that story may be visible somewhere. Production has its data. Maintenance has its records. Safety has its incidents. HR or training has competency information. Supervisors know overtime is increasing. Operators know which shifts are stretched. Finance can probably see some of the cost.

Each part may even be analysed on its own. Production had a tough week. Overtime was approved. Maintenance is busy. Competency gaps are being managed. Incidents happen. The action is closed. The dashboard is still mostly green. But together, the pattern may be telling a very different story.

The business may be building pressure.

That is the signal we are trying to see before the car is stuck in the carpark. Before the production issue becomes an overtime problem. Before the overtime problem becomes a fatigue issue. Before the competency gap becomes a quality problem. Before the maintenance delay becomes downtime. Before the weak signal becomes disruption, loss or harm.

The issue is not that a human leader could not understand that pattern. Of course they could. Good leaders and operators often sense that the answer is somewhere in the work. The problem is scale, speed and connection.

It is exceptionally hard for a human to find that pattern manually across multiple workflows, systems, comments, records and time periods while the business is still moving.

It is not hard for a human to know the analysis should be done.

That distinction matters.

AI does not replace the leader. It changes what the leader can reasonably see. It can bring forward the pattern, the supporting evidence and the questions worth asking next. It can show where the available information points in the same direction and where confidence is weak. It can help leaders see whether the business is looking but not seeing, seeing but not responding, or responding but not improving.

That is very different from simply making existing administration faster.

Faster administration may be useful. It may reduce effort, speed up review and remove some friction from the process. But if all AI does is help us process weak information faster, then we have not changed much. We have just accelerated the existing system. The real opportunity is not faster administration. It is finding the signals in the work before they become proof you acted too late. That is where the leadership conversation needs to move. Not “how do we use AI?”, because that question starts with the technology. The better question is, “What do we need to see earlier, understand faster and act on with more confidence?”

That starts with the work. It starts with the decisions leaders need to make, the disruptions they are trying to avoid, the outcomes they need to deliver, the risks they need to control and the people they need to protect.

AI becomes useful when it serves those things. It becomes useful when it helps the organisation see what was previously hidden in plain sight.

Because most businesses already have more data than they realise.

The question is whether they can hear what it is trying to tell them.

Paul