Rethinking Risk Through Events That Never Result in Claims

by Adriana Zambon

For a long time—and to a large extent still today—the insurance industry has built its understanding of risk by observing events that result in a loss: accidents, claims, damages, and compensation requests.

But what if we tried to look at risk from a different perspective?

A claim represents the moment when risk materializes and becomes measurable: accident frequency, loss severity, geographical distribution, and the characteristics of the drivers and vehicles involved.

Today, however, the growing availability of telematics data is opening new perspectives on risk.

If a claim represents the final outcome of a sequence of events, is it enough to observe only the result? Or is there an even broader source of information hidden within all those situations where an accident could have happened but did not?

The question may seem paradoxical, but it raises an important consideration: the future of risk analysis may depend less on understanding what happened and more on understanding what was about to happen.

A Claim Tells Only Part of the Story

A road accident is a relatively rare event.

Every day, millions of people travel billions of kilometres without generating a single claim. Yet during these journeys, situations involving potential exposure to risk occur continuously: a sudden slowdown in traffic, emergency braking, an unexpected pedestrian crossing, an evasive manoeuvre, or the successful navigation of a particularly complex road segment.

These events do not necessarily result in an insurable loss. They do not appear in claims statistics. They do not generate compensation requests. Yet they reveal a great deal about how risk develops and is managed in everyday mobility.

Focusing exclusively on accidents means, in a sense, concentrating only on the final moments of a much more complex sequence of events.

The evolution of telematics has made it possible to observe mobility behaviour with a level of detail that would have been unimaginable just a few years ago. Today, millions of journeys can be analysed, providing insights not only into where and when accidents occur, but also into how risk emerges, evolves, and—in most cases—is managed before it turns into a claim.

The concept of a near miss is not new and has long been used in various fields of safety management and insurance research. What is changing today is the ability to observe these events at scale through telematics data and transform them into a new source of risk intelligence.

In this context, information emerges that has traditionally not been part of insurers’ analytical assets:

• high-risk driving events that do not result in a claim;
• exposure to complex traffic environments;
• interactions with particularly challenging infrastructure;
• behaviors that help avoid potentially dangerous situations;
• recurring patterns that precede accidents.

These signals make it possible to observe risk at a much earlier stage than when it ultimately materializes as a claim.

Perhaps the most interesting aspect is that many risk situations follow recurring patterns. Before an accident occurs, there are almost always several conditions that increase its likelihood: inappropriate speed for the context, high traffic density, distractions, aggressive driving behaviours, or simply a combination of unfavourable factors. The difference is that, in most cases, the chain of events is interrupted before damage occurs. A driver reacts correctly. Another road user changes course. Traffic conditions shift. The risk is neutralized.

Analysing these situations means studying not only system failures, represented by accidents, but also successful outcomes. In other words, it means understanding the factors that enable millions of people every day to avoid becoming involved in a claim.

From Avoided Accidents to Prevention

This perspective opens new opportunities for the insurance industry.

While traditional models are primarily based on the historical observation of claims, the availability of telematics data now makes it possible to integrate a preventive dimension. Understanding where latent risks are concentrated, which behaviours most effectively reduce accident probability, and which contexts generate greater exposure can help insurers develop more effective prevention and risk management strategies.

The objective is not to replace claims analysis, which remains a fundamental component of the insurance business, but to enrich it with a more comprehensive understanding of the dynamics that precede a loss.

In this scenario, the most valuable data may not be only the data describing accidents that have already occurred. It may be the data that tells the story of the thousands of avoided accidents each day.

In this context, OCTO observes millions of journeys and billions of kilometres travelled every day. The growing availability of telematics data creates opportunities to explore new dimensions of risk, including those associated with critical events that do not result in an accident.

Understanding avoided accidents means beginning to interpret risk differently: not merely as an event to be recorded, but as a phenomenon to be understood and, potentially, anticipated. It is precisely within this still largely unexplored dimension that one of the next frontiers of insurance intelligence may lie.

Why territory is no longer just a pricing variable, but a strategic source of insurance intelligence

by Tina Martino

The recent debate on regional disparities in Motor Third-Party Liability (MTPL) premiums confirms that geography remains a key variable for the insurance market. However, the focus is no longer merely on identifying where insurance premiums are higher. The real challenge today is identifying which factors make an area more exposed to risk, how these factors evolve over time, and how they can be transformed into actionable insights for insurers.

For decades, territory has been one of the fundamental elements in assessing MTPL risk. The rationale is straightforward: some areas experience a higher frequency of accidents, greater average claim costs, or a higher concentration of claims than others. As a result, the policyholder’s place of residence continues to have a significant impact on insurance premiums.

However, viewing territory solely as a pricing variable risk oversimplifying a much more complex phenomenon. Regional differences in premiums are merely the visible outcome of deeper dynamics, including infrastructure quality, mobility intensity, route characteristics, traffic density, observed accident rates, and driving behavior.

In a context where mobility is becoming increasingly dynamic and data is more readily available than ever, an important question arises: is it still sufficient to assess risk based on a policyholder’s place of residence, or is it necessary to understand the broader mobility environment in which the vehicle operates on a daily basis?

Many pricing models continue to interpret territory through traditional geographic categories, which often fail to capture the complexity of real-world mobility.

Two drivers living in the same area are frequently assigned a similar risk profile, even though they may travel on very different roads, move through different environments, and face entirely different traffic conditions. In other words, place of residence does not always reflect the actual mobility environment in which risk develops.

At the same time, the way people move has changed dramatically. Cities continue to expand, traffic flows are constantly reshaped, and daily mobility now spans increasingly broad and diverse territories. As a result, risk itself has become more dynamic.

For insurance companies, understanding this complexity is an increasingly important challenge.

What is needed is a new form of territorial risk intelligence—not simply a geographical snapshot of premiums or claims, but an analytical capability able to interpret territory as a living system, shaped by flows, behaviors, and varying exposure conditions.

From Geography to Real Mobility

The evolution of telematics and data analytics has introduced a new perspective: observing not only who is driving, but also the environment in which driving takes place. Today, multiple data sources can be integrated to build a more comprehensive representation of territorial risk, including:

• Infrastructure risk;
• Historical accident records within the area;
• Mobility exposure, traffic intensity, and weather conditions;
• Trip Score, a statistical assessment of the average risk level of trips that originate, end, or pass through a specific area.

Among these factors, the Trip Score introduces a behavioral dimension into territorial analysis by linking area risk to actual observed driving styles. Together, these insights make it possible to move beyond a purely geographical perspective and develop a more accurate understanding of the mobility environment.

This is where the perspective shifts: territory is no longer viewed merely as the place where the policyholder lives, but as the set of real environments—across both space and time—in which the vehicle moves, is exposed to traffic, travels through different infrastructures, and encounters varying levels of risk.

The true innovation, however, is not simply the use of more data. It lies in transforming territory into an information asset using telematics data.

By combining multiple layers of information—including road infrastructure, observed claims experience, mobility intensity, and driving behavior—it becomes possible to create a composite risk indicator capable of highlighting significant differences even within the same municipality or province.

This approach enables insurers to gain a more granular understanding of actual risk exposure. In this context, OCTO has developed a Risk Heatmap that integrates infrastructure data, claims information, mobility exposure, and behavioral indicators, transforming territory into genuine territorial risk intelligence.

This is not about adding another map to insurance processes. Rather, it is about enabling a deeper understanding of territorial differences—one that helps distinguish between areas that may appear similar from an administrative perspective but differ significantly in terms of exposure, actual mobility patterns, and risk profile.

A New Tool for Underwriting and Pricing

A more advanced territorial view can support a variety of insurance processes.

During quotation, it helps assess new customers even when individual telematics data is not yet available. In underwriting, it enables more precise risk segmentation. In portfolio management, it allows insurers to identify critical territorial concentrations and opportunities for optimization.

Prevention initiatives can also benefit from this approach through targeted coaching programs, contextual alerts, and more effective customer engagement strategies.

Furthermore, this methodology enables a shift from a predominantly reactive approach—based on historical claims observation—to a more predictive and operational model, capable of supporting insurance decisions that are more sustainable, targeted, and aligned with actual risk exposure.

The Next Evolution of Motor Insurance

Motor insurance is gradually evolving from a model focused on observing the past toward one increasingly capable of understanding and anticipating risk.

Following the introduction of telematics and behavioral analytics, the next step is to integrate a deeper understanding of the mobility environment in which customers are truly exposed to risk, moving beyond simple geographic residence data or administrative boundaries.

In this scenario, tools such as Risk Heatmaps represent a new generation of territorial models: not merely geographic maps, but intelligence platforms capable of correlating all risk variables, transforming complex data into operational insights, and supporting insurance companies as they become increasingly focused on personalization, technical sustainability, and prevention—the ultimate social objective.