A New Dimension of the Digital Revolution

He adds that today we can identify three key areas where the use of data and analytics is of fundamental importance. The first is insurance pricing, where AI helps propose the best possible premium for the customer relative to the risk assessment.

“In a nutshell, customers who have a claim-free insurance track record to date, for example in in motor insurance, can expect a lower price. To this end, risks are forecast on the basis of a number of attributes, not just those related to the frequency of events; as a result, the premium can be calculated for each customer individually, using predictive analytical models. Those whose exposure to certain mitigated risks is reduced, for example as a result of preventative actions, can likewise anticipate a better pricing. A good example is smart agriculture and our pilot solution iAgro, where we use a system for monitoring and signalling plant disease outbreaks based on data from meteorological stations and mathematical models. This allows us to react in time by, for example, applying sprays to reduce the risk of crop failure”, explains Marek Wilczewski.

Thanks to AI, telematics and machine learning, it is also becoming possible to monitor customers' needs and behaviour in near-real time and conduct deep analytics. This, in turn, paves the way for the development of more accurate predictive systems that not only alert users to potential dangers and unsafe actions, but also assist them in making choices that drastically cut down on potential dangers.

Furthermore, insurance sales development can benefit from AI by better understanding consumer wants and utilising CRM campaigns to this end. This helps with the improvement and expansion of digital channels and enables the customisation of offers.

“Analytics allows us to group insurance customers into segments, that is groups of people with similar expectations and preferences, based on the vast amount of data we obtain from multiple sources at different stages of our relationship with them. This allows us to offer a product that is relevant and tailored to their needs. Therefore, we don’t generate unnecessary and unappealing offers with a low probability of conversion to sale”, Marek Wilczewski explains.

When it comes to online sales, artificial intelligence can help figure out why a customer who wanted to buy a policy online ultimately chose not to finish the purchase and, on this basis, improve the digital channel or offer contact through a traditional channel (agent, employee). PZU reports that over the past two years, the number of users of the myPZU self-service platform has increased from 1 to 4 million. Moreover, a significant proportion of online policy purchase transactions may have been initiated by customer contact with intermediaries (e.g. agents).

Finally, artificial intelligence supports the area of benefit payments and policy claims, also helping to counter fraud. PZU was the first insurer in Poland, and one of the first in the world, to implement an AI solution in vehicle damage management in which customers send photos and a description of the damage the vehicle suffered, and the AI-based system analyses, in real time, the extent and type of damage, the list and availability of parts needed for repair, and generates a preliminary cost estimate on this basis. This significantly streamlines and speeds up the process of developing initial estimates and processing damage claims. This is accomplished in large part through application of a machine text recognition mechanism combined with automatic text analysis, using semantic OCR. This tool uses AI to automatically interpret content from various types of documents, search for relevant data and feed it into PZU's systems.

Artificial intelligence is also used to an increasing extent in healthcare. Under the “One minute for Skin” prevention programme, PZU offers employees of selected companies that purchased life insurance from PZU Życie a skin cancer identification service, using an innovative, medically certified “SkinVision” mobile app. Besides that, telemedicine using artificial intelligence is being continually developed within PZU Health. Recently, its portfolio of services was expanded to include two new packages for diagnosing and monitoring heart rhythm disorders. Earlier, the company introduced lifebands, which monitor patients' basic vital signs and alert staff in the event of irregularities.

“We are developing robustly in applying AI. This means implementing further improvements not only for the benefit of our customers, but also for our staff. We want to do away with repetitive tasks, such as entering or analysing data. Let me give you an example of damage assessment in agriculture – in the past, our appraiser had to travel to the site to carry out the appraisal in person. Today, we can use satellite data”, points out Marek Wilczewski, and notes that the company's ambition is to streamline internal processes by using robotics wherever possible.

“The immediate future appears to lie with generative artificial intelligence, of which ChatGPT is the best-known solution. This technology makes it possible to summarise content, analyse input material and then suggest conclusions and solutions in many areas of insurance”, says Marek Wilczewski, adding that although artificial intelligence is being applied to an ever increasing extent, human supervision is still necessary.

“Care must be taken to ensure that, in addition to having the appropriate consents when analysing data and constructing offers, there is no anti-selection of customers based on gender or views, for example”, he adds, while acknowledging that the use of AI is a huge opportunity to gain a competitive advantage. However, as he points out, effective analytics does not happen without good data quality and an understanding of analytics techniques.

“In the end, it comes down to the business idea: how to use data to create relevant insurance products or service delivery. After all, the most important thing is to maintain the best possible long-term relationship with the customer, by providing sustainable value. Therefore, whoever is able to gain knowledge from the data in their possession and then use this knowledge will be successful in business”, emphasises Marek Wilczewski.

Multiple areas and functions within PZU are working on developing solutions that employ AI components. In addition to the Information Management Office, there is also a team of experts called the Artificial Intelligence Factory, which develops new solutions based on AI and utilises the latest developments in AI and machine learning to analyse terabytes of data. The Innovation Lab, in turn, as part of the PZU Ready for Startups programme, focuses on facilitating PZU's collaboration with startups and other innovative technology companies.

In addition, the GPT Lab initiative has been in place at PZU for the past few months. Its main task is to collect, analyse and develop ideas for leveraging the potential of generative AI, prepare safe ways of testing it while ensuring the highest regulatory and ethical standards are followed, and implement the initial pilot implementations of solutions based on this technology. Pilot activities are currently underway in four areas selected from over 100 ideas collected from the entire organisation.

partner