Skip to the content

While there is still time to act

The knowledge needed for a good decision may exist somewhere in the organisation. That does not mean it will reach the person who needs it.

Caspian Ravn,

1 The world as it is

Insurance is very good at putting things back together.

After a loss, it helps a business recover. It also has a long history of prevention. Risk engineers study how operations work and advise on the protection they need.

A site visit or a policy renewal gives everyone a moment to assess the risk and agree what is needed. But that moment passes, and the week carries on. No set of requirements can account for every situation every day brings.

At Novo Nordisk, I worked alongside people who understood this. They knew what they were protecting and how their own work could affect the operation around them.

Even so, I once saw two routine tasks overlap in a way that left a critical safeguard exposed. Each job made sense on its own; together, at that moment, they created a gap in protection. That is how risk can move: not always through a dramatic failure, but through ordinary work meetings at the wrong time.

I only noticed because I happened to be there.

Since starting my company, I have heard versions of that problem in different countries, industries and organisations. Even businesses with dedicated risk teams describe it. Their specialists are brought in for the big changes. The everyday ones fall to the people on the spot, who know their work but cannot always see what else depends on the thing they are about to change.

The knowledge needed for a good decision may exist somewhere in the organisation. That does not mean it will reach the person who needs it.

We believe it should reach them while there is still time to act.

2 Insurance has been upstream before

Insurance has long helped organise prevention.

In the decades after the Great Fire of London, insurers employed their own firefighting crews. Limiting the damage protected their customers and reduced their losses.

In 1835, a Rhode Island mill owner named Zachariah Allen made his mill safer and asked his insurer for a lower premium. He was refused. So he joined with other mill owners to form a mutual insurer that accepted factories committed to preventing loss.

Insurance kept finding reasons to act earlier. Putting out a fire could spare the cost of rebuilding. Making a mill safer could prevent the fire altogether. We believe the next step is to bring that knowledge to the everyday changes where a loss can begin.

Before factories, every shirt was sewn by hand. Then machines came, one worker could make hundreds, and shirts became cheap. The economists William Baumol and William Bowen showed that some work never changed this way. A doctor still needs time with each patient, so as wages rise, care stays expensive.

Risk engineering faces the same limit. Standards and training help knowledge travel. Understanding what a change means for a particular operation still takes an experienced person’s attention. An hour spent on one operation is an hour they cannot spend on another.

Insurance did not forget how to prevent. It has always had to choose where to put its people.

A large organisation may be able to afford its own specialists. A smaller business still has people to protect and difficult decisions to make. We believe both deserve expert help when they need it.

3 What becomes possible

We believe AI can help specialist knowledge reach more of the everyday changes an engineer cannot personally review.

In the operation I described, the problem was not in either task. It was in the part they shared, at the moment they shared it. We want to make risks like that, the ones that are easy to miss, easier to catch.

Risk specialists know risk well. But catching a problem like that takes more than knowing risk. It takes knowing what else is happening in the organisation, and what is physically true on site at that moment.

New technology may make that the natural next step. Not as a tool that gives an answer, but as a collaboration. The people on site share what they see. The engineer brings what they know. In time, AI could do more than connect the two. It could join the conversation, raise the questions and keep the picture current. Together, they would work through it until the answer holds. Every day is different, and a tool cannot have that conversation.

People have always worked together to take on serious tasks. Managing risk, at any scale, is one of them. It deserves to be held to the highest standard.

If we can meet that standard, more everyday decisions can benefit from specialist knowledge. And that help should reach businesses that could never afford a risk team of their own.

4 The future we are building

We will build alongside the people doing the work. They see the operation change every day. They also know when advice is impractical, when a detail has been missed and when a proposed solution would create a problem somewhere else. Their experience has to shape what we build and how we judge it.

Requirements, inspections and protective systems remain essential. We want to help people recognise when the work has moved beyond what those arrangements assumed. A signed permit can establish what was authorised. It cannot, on its own, establish that the protection still fits what is happening now.

Better information also creates a choice. It can help a business become safer, or simply give an insurer more reasons to turn it away. Allen’s mutual rewarded prevention, but it also selected which mills to insure. Knowing more about risk does not settle how that knowledge should be used.

We want it to reward care. A business that takes steps to protect itself should have those steps recognised. And insurance must keep sharing the risks that remain. The value of better knowledge should be felt in the help a business receives and the losses it can avoid.

We will judge our work by whether it helps people make better protective decisions in time to act. That means examining where the advice helped and where it was wrong. Fewer good decisions should depend on someone happening to be in the room.

Whether you run an operation or help insure it, what changed this week that was difficult to judge? Bring us that example. We want to build this with you.