Managing major losses digitally

Managing major losses digitally | Corporate Risk & Insurance

Managing major losses digitally

By Tobias Büttner and David Feghelm

Tobias Büttner, head of claims, and David Feghelm, senior solutions manager in claims at Munich Re discuss digital technology in managing losses

Tobias Büttner: Claims management in particular is an area that will benefit enormously from new technologies, most of all from digitalisation. InsurTech start-ups such as Lemonade in the USA have already demonstrated that fully automated claims handling of mass business using artificial intelligence is no longer just a vision. But also the management of very large losses will increasingly build on the use of new technology.

David Feghelm: Intelligent image recognition is becoming more and more relevant for (re)insurers in the management of major losses such as natural disasters. Automated evaluation of satellite images and aerial photographs using algorithms provides a rapid overview and accurate estimate of what losses have occurred where. Resources such as loss adjusters and on-site appraisals can then be deployed more efficiently, cutting both costs and claims processing times. In an ideal scenario, once you combine the information obtained in this way with portfolio data (insured risks and their location), you can quickly make a remote estimate of the loss amount in your own portfolio. The main challenge here is data quality, particularly in terms of the geocoding of risks and adjusting algorithms for claims classification purposes.

Büttner: With complex major losses, it is vital to have very large amounts of relevant data in sufficient quality. In this context, I can see major benefits for everyone involved, in that subjective assumptions and gut feelings will be replaced by decisions based on concrete data. However, loss data needs to be complemented by data on exposure and policies, with unrestricted linkability between the data sources wherever admissible.

Feghelm: Where legally permissible, a central data lake* enables enormous amounts of data to be stored and linked together. In this way, risk and claims data can be enriched with relevant aspects, knowledge gaps closed, and a holistic understanding established. Using modern data analytics, relevant information for claims management purposes will be made available quickly, easily and in a targeted manner. Claims trends will be recognised earlier and claims benchmarks generated. With the help of predictive analytics**, algorithms will be able to forecast claims amounts much more precisely. This will enable us to detect major losses with unusual run-off patterns more quickly. Providing our clients with this data will bring prompt agreement with the policyholder and increase customer satisfaction, while at the same time minimising the risk that a claim may turn out to be larger than expected. Underwriting will also benefit if the acquired data can be processed in such a way as to derive long-term findings on risk selection and pricing.

*A data lake is a method of storing data within a system or repository, in its natural format, that facilitates the collocation of data in various schemata and structural forms, usually object blobs or files.

**Predictive analytics encompasses a variety of statistical techniques from predictive modelling, machine learning, and data mining that analyse current and historical facts to make predictions about future or otherwise unknown events.

Büttner: Another major challenge will be making available processed and collated information from claims files, claims reports and other unstructured sources. As far as possible, unstructured information will be converted into structured data to make it usable for data analyses.

Feghelm: Progress is being made in this area with the help of text mining* and natural language processing – software-based analysis of text in an automated process. The consistency and high quality of the data are the main benefits of using such technologies for data collection. For example, to allow loss events to be promptly detected and trends identified, we are already systematically searching news sites, blogs and social media sources.

*Text mining, also referred to as text data mining, roughly equivalent to text analytics, is the process of deriving high-quality information from text.

With both mass business and very large claims, digitalisation can also be used for loss prevention purposes. Insurers systematically collect data from all legally accessible sources which can be used for this purpose, such as social media. As a global reinsurer, we have the benefit of being able to access primary data while safeguarding our clients’ interests. This data is then automatically assessed with the help of artificial intelligence. Data analytics will allow claims trends and unusual developments to be identified at an earlier stage than is possible today, and will also facilitate mitigation of the associated risks. In this way, insurers can begin active claims management much sooner.

Büttner: Alongside loss prevention, modern technology will also produce lasting improvements in claims management, both during and after the occurrence of a loss. Take the example of a large fire in an industrial plant: claims management in this instance is a complex process, and full automation of the process poses a major challenge.

Feghelm: Particularly with major and complex risks, effective communication with all stakeholders will play a key role, all the way from the policyholder to the reinsurer. Efficiency can already be greatly enhanced today by ensuring rapid access to all relevant information, and greater transparency between the different parties involved. Further improvements will be achieved in the future supported by new mobile and cloud solutions. Information can then be simultaneously shared among participants. For example, if a loss adjuster uses a laptop or tablet to make digital records of data on-site, including photos and videos, these can be made available directly and simultaneously to the insurer and the reinsurer via a data connection and cloud services. This eliminates any waiting times for reports and reduces processing times.

Büttner: Automation of claims management will greatly enhance efficiency – making the whole process faster at the same level of quality, and cutting payout times significantly. The enormous potential involved here can be illustrated by the example of a large-scale fire in a production facility: remotely controlled drones can produce image and video material at a time when it would be much too dangerous for people to enter the site.

Feghelm: At the same time, claims managers can get a good idea of the loss situation directly through a video transmission to their workstations, and advise policyholders on catastrophe management and what steps they should take next. The claims manager will be assisted by intelligent chatbots* that record information from communication with clients and assess it in real time. So policyholders receive their money sooner and enjoy better service as well, while processing costs are reduced on the insurance side. Eventually, we will be able to identify the cause of loss fully automatically using artificial intelligence and make initial estimates of the loss amount.

*A chatbot is a computer program which conducts a conversation via auditory or textual methods.

Büttner: Sensor technology and the Internet of Things* will also become much more important. Here I am thinking of devices like networked sensors that automatically send massive amounts of data to insurers. With the right evaluation, the data obtained in this way can provide information on losses, and also on early detection of losses and loss mitigation.

*The Internet of Things (IoT) is the network of physical devices, vehicles, and other items embedded with electronics, software, sensors, actuators, and network connectivity which enable these objects to collect and exchange data.

Feghelm: For example, sensors are already being installed on wind turbines today to act as early warning systems, drawing attention to problems in the gears, and triggering a system shutdown before any greater damage can occur. The production processes of the client (the manufacturer of the equipment) could also be improved at the same time through intelligent analysis of the data. There are virtually no limits to the possibilities in this area. To ensure smooth-functioning processes and allow them to develop their full potential, it is essential for information flows from smart sensors to be seamlessly integrated into the internal system landscape.

Büttner: As with claims management, the role of the claims manager will also undergo changes. The role of central contact providing services directly to the client will be added to the function. The internal claims technician will then be given the task of managing the process as a highly qualified manager.

Feghelm: Data scientists* will also become increasingly important, structuring the data in a practical way, filtering the essential information from the enormous mass of data and identifying future loss patterns. This works particularly well if claims managers, risk engineers and data scientists all work very closely together.

*Data science is an interdisciplinary field about scientific methods, processes, and systems to extract knowledge or insights from data in various forms, either structured or unstructured, similar to data mining.

Büttner: All in all, the vast array of options offered by digitalisation make it possible for insurers and reinsurers to get a much faster and clearer picture of a loss and to optimise their decision-making processes. For their part, policyholders benefit from faster claims processing and payment, and from the additional services that digital tools offer.

Originally published on Munich Re’s Topics Online.


Related stories:
Cyber concerns outrunning risk management
AI: a risk management game changer