One of the key pieces of information in Property & Casualty insurance (P&C) is the location. This applies both to pricing and risk assessment, on the level of a single insurance and the entire insurer’s portfolio. Due to the reinsurance capacity, particularly the capital requirements, assuring the right quality and an ongoing monitoring of one’s exposition have also become extremely important. Modern technologies more often and to a greater extent enable insurers to collect greater amounts of information, automate processes and more efficiently manage their portfolio. Among them are technologies and systems, which use broadly-defined geospatial information (data with a feature in the form of geocoordinates – longitude and latitude).
The key condition which must be met before an insurer is able to start using this technology is ascribing geographical co-ordinates to the owned portfolio of insured locations. This could prove to be quite a challenge, as one has to ensure a proper quality of data for a given location. Individual elements of the address, i.e., the name of the street, the number of the building and apartment, postal code, locality, all should be separate features in the database. A factor significantly improving the data quality is also introducing vocabularies in the source system, where the address data is entered, which would ensure accuracy and cohesion of the data. Such solutions might seem natural, because we come across them almost every day, however it has not always been the case and it is not always easy to implement either. This is an unescapable challenge for the insurers on the markets, where long term agreements are signed, or have a tacit renewal option. There, if the policyholder does not terminate the contract, it is continued for the next, usually a year-long, insurance period. From the technical perspective, very often it is solved by copying the data from the previous period or extending the insurance period within the existing policy. As a result, the data and its structure remain unchanged for many years; the address data quality also might require an automatic or manual correction. In order to be able to fully benefit from the potential of the geospatial technologies, it is necessary to provide the right accuracy of geolocation, which is determined by the quality of the address data and the coverage of a geolocation services provider of a given area.
In line with the quality expectations, such a service should guarantee the largest possible amount of addresses located on the “rooftop” level of a particular building. With lower geolocation accuracy, e.g., up to the middle of the street, locality or area of a given postal code, the results of the analysis and the benefits from using geospatial data might be significantly affected.
Once you have the data, including geographical co-ordinates, there are many options of using it. The GIS (Geographical Information System) software enables you, most of all, to analyse and visualise the data, owned by the insurer, such as the premium, the amount of the sum insured, the number of insured items or the amount of claims, e.g., in the form of a heat map.
An example of using such analysis is monitoring, i.e., verification of what the total value of insured objects (the sum of sums insured) is, e.g., at some specific distance from each other or in some area. Each insurer has a reinsurance protection within a certain limit. When signing individual insurance agreements, the carrier verifies if their sums insured do not exceed that limit. It is very difficult and imprecise to verify the entire portfolio and dependencies between individual agreements without using the GIS technology. The lack of such analyses or a mistake might be very expensive for the insurer. In a situation, where the carrier insured several clients within one location, and the total sum of all these agreements exceeds the reinsurance capacity, it is likely, that the insurer will find out about once the clients report claims. In the worst-case scenario, in the case of total losses, it might turn out, that the reinsurance protection was not sufficient, and that the insurer will have to participate in the damages in the extent larger than planned. This might disrupt the insurer’s solvency
Another example of such use might be following trends in the number of reported claims within a given area. Detecting an increased number of claims might entail increasing the number of experts or establishing cooperation with a new contractor (a car repair shop or medical facility).
A reason for such a spike, say, in motor insurance, might suggest illegal activity in a given area, where fraudulent claims are reported. Crisis situations, like catastrophic events, are another area where geographical data can be applied. When you have the information about the location, where a given phenomenon – a flood, a hurricane or an earthquake – occurred, from the media reports, for example, you can calculate a number of clients exposed to potential damages and estimate their value. This, in turn, makes it possible to plan for a larger team servicing the helpline or more appraisers in that area. In other words, one can sustain operational capacity and avoid adverse repercussions of being unprepared for the unexpected.
Insufficient number of claims registrars might adversely impact the net promoter score (NPS) of the aggrieved parties, which in turn might impair the trust of the customers and future sales. Without enough experts, the process of claims handling will be longer, exposing the insurer to additional costs, connected to the damage increasing or potential fees and interest for failure of damage liquidation in the required period.
Apart from analysis using geographical data, GIS technology and tools make it possible to create own layers of data included in the maps. Such layers link the areas with a common feature or value. Those might be used in underwriting or pricing during the sales process. Most often insurers use the layers of natural hazards (floods, hurricanes, earthquakes, etc.), prepared by external partners. There can be one or many layers for a particular type of risk. In case of a flood, one layer can map out the areas exposed to a flood which happens once every 100 years, another layer those exposed to the flood which happens once every 50 years, with others using different recurrence intervals. In case of hurricanes, particular layers can be created, e.g., based on different wind velocities. When an insured location overlaps with a layer, a certain premium might be calculated, or such information might be used in business rules (the terms of insurance might change or an individual assessment by an insurer might be necessary). Layers can be used for a variety of information, e.g., when generating leads for new clients based on the layers of areas, where particular agents operate or when directing patients in health insurance to the closest medical facilities.
In the world that surrounds us, more and more data resources can be treated as geospatial, which can constitute a source or extra element of additional analyses. They can also be applied in the processes typical to insurance companies. An example of such a database is cadastral data, which includes information about cadastral parcels, a number of buildings, their surface or structure. Such data can be used in the sales process of property insurance, as on the one hand it simplifies the process by reducing the amount of information which the client has to share and on the other, it enriches the risk analysis and, as a result, increasing the accuracy of premium calculation. An example of the data used during the claims handling process, might be the databases compiled by weather stations. Based on such information one can automatically verify, whether there was rain and wind, which could cause the damage in the area, where the claim was reported. With this data, an insurer might decide to qualify the cases verified in this manner for a simplified and quicker – fast track – process of claims handling or even go for a process of full automation, significantly reducing the processing time. However, there are some limitations, when it comes to utilising these types of external databases – access to particular scope of data might differ market to market. Often, the data available in Germany is not available in the UK or Poland. GIS technology might also be helpful in a situation, where the data you seek is not included in any available set – by helping us create it.
High resolution photos are widely available commercially from at least several providers. Also, the frequency, with which a satellite returns to any point above the Earth has significantly increased. This opens a possibility to develop algorithms of an advanced analysis of satellite images and retrieve the data which has not been available before. There might be a number of potential applications for such a photo analysis, which could help reduce costs, increase the sales, or improve the NPS.
An example of using this type of data in the claims processing, could be an occurrence of catastrophic events, such as a flood or a hurricane. With up-to-date photos of the area affected by the disaster, it would be possible to diagnose damaged buildings, and in some cases also the scale of the damage. A reported claim can be quickly confirmed, and the compensation estimated and paid without a person being involved. We can go a step further in these deliberations and consider an insurer proactively contacting the insured with the information about a diagnosed potential damage and offering to pay the compensation or additional assistance, like cleaning or transport services.
When it comes to more complex damages of corporate clients, it is also possible to undertake steps mitigating the damages with the use of data extraction from photographs. Assuming, that a production hall was flooded, the information about water depth obtained from the latest satellite pictures will make it possible to identify, which machines or their components were damaged. Obtaining such information early, before the water subsides and it is possible to verify the situation on-site (which might take even a few weeks), allows to order the delivery of replacement parts earlier, reduce the time of recovery to full operational capacity of the company and reduce the amount of compensation for the business interruption. In case of the corporate clients, this type of technology might be also used in risk assessment.
The amount of information analysed by underwriters in corporate insurance is vast and often unstructured if not outright inaccessible. The information includes the number of buildings, their structure, technical condition, etc. Such information can be easily generated through photo analysis. Besides, satellite photos can be a source of information about hazards present in the vicinity or the way in which flammable materials are stored close to the buildings. Accessing them might significantly simplify and automate the underwriting process, as well as minimise the necessity to involve loss control engineers, assessing the clients on-site, versus the assessment done remotely in the office. The time saved might be spent on evaluating a larger number of clients and a better selection of the insured ones, and therefore a higher profitability and quality of the portfolio.
Sollers Consulting had an opportunity to cooperate a Customer, who implemented this type of technology, using it to collect information about the sizes of insured agricultural buildings, their spatial location on the insured plot, which made it possible to identify them in the event of a damage. Before introducing this solution, the measurements and drawing up the plans of the farm were done by insurance agents. Because of that, the sales process was time-consuming and complicated. Automating the process enables the agent to focus on the sales activities, which bring additional value in the form of a premium. On the other hand, from the point of view of the insured customer, the process of signing an agreement got simpler and doesn’t require making measurements, which has a positive impact on his/her customer journey and a general perception of the insurer.
Geospatial technology in its broad definition, is developing dynamically, and the experience of Sollers Consulting demonstrates that they are implemented more and more frequently. The impact on the insurance sector will be visible and will mostly show within the scope of simplifying and automating the processes. The amount of information connected to location, provided by the clients or the aggravated parties, will be reduced to a minimum. More processes, supported by the data obtained using the geospatial technologies, will be STP (straight-through-processing) processes, where the contact between an insurer’s employee and a client will not be necessary. This might be particularly visible in the case of individual customers. A potential impact of the technologies, discussed above, might be illustrated with an example of insuring a house. The process of signing a policy might soon boil down to simply providing your personal data and the property’s address. The address will be geocoded, geographical co-ordinates will make it possible to collect information from external databases and photos. The surface area will be calculated, as well as its value – based on the real estate prices in the area. The technical condition and surroundings of the building, e.g., a brand of the vehicle on the driveway, will enable the insurer to estimate the insured material status and the value of the house furnishings. The offer will include an insurance for a dog, because an algorithm found a doghouse on the satellite image. A risk of an occurrence of natural losses will be estimated based on the map, which shows the recurrence intervals of such events. The premium for insurance against theft will be determined based on the number of burglaries in the area from the statistics of the local police office. If a hurricane blows away the roof, the insurer will inform the customer about a proposed amount of compensation, details of the local companies who might repair the damage and a few hotel suggestions to stay during the repairs…
Geospatial technologies were presented using the examples of the non-life insurance, on which they will likely have the greatest impact. Nevertheless, the technology applied in the analysis of satellite photos can also be used in the case of traditional photos, taken by a smartphone, for instance. This opens new possibilities of practical use, e.g., in motor claims handling.